September 3, 2023 • 64 mins
Fresh off his keynote presentation at the Canadian Society for Marketing Professional Services Annual General Meeting, Kris offers an in-depth look into how AI is revolutionizing architecture, engineering, and construction.
From data-driven design solutions to real-time analytics and ethical technology use, this episode is a must-listen for anyone eager to harness the transformative power of AI.
Whether you're an industry veteran or just starting your journey, Kris provides invaluable insights into using AI as a creative co-pilot to innovate, optimize, and disrupt the AEC landscape.
-----------------------------------
Make sure you're part of our digital tribe across all platforms. Your support goes a long way!
For more insights and behind-the-scenes peeks, find me here:
✅ Subscribe to free Kris Krug Newsletter
🎧 The MØTLEYKRÜG Podcast on Spotify
🎧 The MØTLEYKRÜG Podcast on Google
🎧 The MØTLEYKRÜG Podcast on Apple
🎧 The MØTLEYKRÜG Podcast on Amazon Music / Audible
🎧 The MØTLEYKRÜG Podcast on Spotify
🎧 The MØTLEYKRÜG Podcast on Google
🎧 The MØTLEYKRÜG Podcast on Apple
🎧 The MØTLEYKRÜG Podcast on Amazon Music / Audible
, AEC industries, architecture, engineering, construction, data-driven design, real-time analytics, ethical technology, creative co-pilot, Kris Krüg
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
I am super stoked to introduce you to this evening's speaker, Chris Krug.
(00:05):
During this presentation, Chris is going to bring his explorations and understanding of artificial intelligence, or AI,
and its implications for creative marketing within the architecture, engineering, and construction industries.
Known for his ability to articulate complex technological concepts with clarity and accessibility,
Chris will inspire and foot professionals with the knowledge and tools needed to navigate and leverage the AI landscape effectively in organizations.
(00:33):
Chris is a creative technologist and professional photographer originally from California, now based on Boundary Island.
By the way, his dog's name is Princess Ice Cream.
That's right.
That's right.
I decided to throw that out there, which is fantastic.
Chris is celebrating for his collaborations with prominent entrepreneurs, athletes, and tech innovators around the globe,
(00:54):
and has worked as a contributing photographer for.
And his website has an intimidating list, but I'm just going to give you a snapshot.
These include National Geographic, Barrett, BC News, Business in Vancouver, Chicago Tribune, CNET, Daily High, Bell,
and of course, I'm going to keep going.
(01:14):
I'm proceeding alphabetically, except for the National Geographic.
I noticed.
Yeah, I noticed.
Huffington Post, Japan Times, LA Times, Motley Fool, New Yorker, Rolling Stone, Slate, Ten Crunch, Ten, The Atlantic,
Tide Magazine, Toronto Star, USA Today, Village Voice, Wall Street Journal, Washington Post, and Wire.
(01:34):
And that is like a fraction.
Yeah, I was intimidated just sitting there reading that.
It was amazing.
His pioneering vision combines artistic flair with an acute understanding of technology,
making his perspective unique and highly relevant in the ever-evolving technology landscape.
Chris is deeply embedded in the world of technology, particularly the exciting realm of artificial intelligence.
(01:57):
He is actively involved in the technology conferences.
And combined with his consulting and teaching roles on internet technologies, online communities, and digital media,
demonstrates his comprehensive expertise in the field.
Chris's work with transformative organizations such as the United Nations and TICTC underlines his commitment to employing technology for societal advancement.
(02:21):
If you'd like to learn more about Chris, his work is at his chriscroft.co.
And you can also connect with and follow him on LinkedIn and on YouTube.
Be sure to like and subscribe and click that notification bell.
All right.
You're hired.
Thank you.
Yeah.
Excited to introduce Chris.
Good.
(02:44):
There's nothing more awkwardness with than what reads your formal bio.
You're so beautiful when you're sitting there.
It's excruciating.
Thanks for having me here tonight.
This is one of my favorite things to do, is to just share stuff that I'm passionate about with other people.
And so I don't know anybody in the room that I recognize.
And I've never been in this building.
(03:05):
I don't think any of you know me.
So I'm a little bit nervous, but I'm really excited to be here.
And I'm an artist and I come from a little bit outside the business world, though.
A lot of my clients are corporate clients.
And so I'm just going to do something here at the beginning that I know isn't exactly what happens in corporate business meetings and stuff.
But I'm just going to I got this from the yoga class.
I'm just going to set an intention here and hopefully we can all tune into that a little bit.
(03:26):
And so I intend to share with you guys stuff that I'm like very passionate and interested in.
I've been experimenting with this stuff for a while and it's changed my life.
It's changed my perspective.
It's changed the businesses that I've evolved with and the projects I've evolved with.
And it has made me an enhanced human being.
And I hope to share some of that stuff with you.
(03:48):
I hope to address concerns you may have or fears or questions or obstacles you have such that you feel empowered after this talk.
So just roll up your sleeves and mess around with this stuff and figure out how it might benefit you or your organizations or your teams or your projects.
That is what we're going to do here today.
How's that sound?
(04:09):
Another one of the ideas here.
So I brought my laptop, but not a dongle and all sorts of fun things to show you that we're going to do together.
But I also had three hours worth of content and we only have about 45 minutes.
So I'm not going to do that.
But what I do hope to do is to turn this a little bit into a facilitated group discussion at one.
(04:29):
So there is a question and answer at the end.
We'll leave time for that.
But my thinking here is like, AI is on the news all the time.
We can't go to a meeting without talking about it.
We're talking about with our parents and with our kids.
And we're evaluating all these different ways, what it means for the future of mankind, what it means for us society, what it means in our jobs.
And so you all are part of that conversation as much as I am.
(04:53):
And so I am interested in tapping into a little bit of the collective knowledge and experience in the room.
And to that point, I know that some of you people are already using AI in your jobs today.
I have met with marketing assistant.
Tell me your name again.
Who was telling me a little bit about her job and you use AI to do proposal writing, which in most organizations, I'm involved with proposal writing in the way you were talking about.
(05:17):
It doesn't fall into the marketing function.
It's more like a sales or business development function.
But so I found it interesting to hear how you were using it to write proposals and stuff.
And can I say what you told me?
Yeah, you said that at first there was some skepticism around you using it, that you'd be cheating or that it wouldn't be the quite right voice or get the nuance of what you were trying to say.
(05:39):
But in fact, paradoxically, it has elevated the writing of the company in general.
And you talked about using old pieces of content and new contextual information from a particular proposal and having GBT rewrite something that's already been approved and is accurate for a particular.
And the neat thing about GBT is you can tell it what voice to use.
(06:01):
And I use a lot of nudging when it comes to GBT.
We'll dive into this in a minute.
It's hard to get it to do exactly what you want it to do sometimes, but you can nudge it in the right direction.
So I often find myself jotting down some ideas, running it into GBT, getting some back so I don't have a blank page and then starting to coax it, having it re-write it in different ways, less formal.
I often use the word fun marketing in my stuff because it has a certain way of trying to embellish the things that it writes in.
(06:27):
It's got a few quirks.
It likes to make lists, these this, that and this other thing, this, that and this other thing.
And so when I read those things, I can clearly see that it's AI.
And so with a little bit of like understanding and nudging, I often only find myself rewriting what comes out of it.
I just find myself coaxing stuff into it.
And I do a little bit of training, too, and we can talk about it if that means or maybe some of you are doing that.
(06:52):
I would like to introduce myself.
I am 45 years old.
I grew up in California.
I moved here to Canada to go to university and I moved here in 1995.
And that was the year the graphical web browser took off.
So my first, I went to Trinity Westman University in Langley this time, Villa Christian College.
And I knew there would be someone here.
(07:12):
I didn't go there, but I was like...
It's a Corpi school, but I got assigned to English as a second language dorm because I signed up late.
And so on my first day of school, I like wandered in, I was like, yo, what's up, guys?
And they were sitting there like eating dried fish and playing Go.
And they were like, personal home page.
And I was like, what?
And they're like, Net State Navigator, personal home page.
(07:33):
And so like within the first couple of weeks of school, I've been starting to build my first web pages.
And it was 95.
And so the professors didn't have web pages and the departments didn't have web pages.
Even the school didn't have a web page.
And so I'm like 19 years old.
I'm proficient in Internet technology.
I'm building web pages for all the organizations, essentially.
And then I'm starting to do like direct and study high level communication classes
(07:56):
because the stuff I want to learn is in books.
It's emerging right now.
And I knew as much about it as anyone in the school.
So they let me go down the road of teaching it to myself and then build a syllabus out of it and come back
and teach my peers the next semester a class that I had self-directed study.
And so I ended up starting a little interactive agency called Primal Communications.
And I got off by an advertising agency in town called Palmer Jarvis in 1998.
(08:20):
And that was kind of like my first little win on the Internet.
And after that, I moved down to Silicon Valley to work for a startup and got to do the kind of like inside the valley thing,
which was amazing for me.
I was just a web guy before that.
And I learned like MarCom, PR and trade show.
Just all the other components that go into marketing.
(08:41):
I worked on there in the valley for a while.
And that company worked out to as a little networking startup that sold to Cisco.
And I was just like the 12th employee or whatever.
But you can imagine it was a good ride.
I was in California working at the highest levels and just getting a master class in the state of the art of marketing.
I moved back to Vancouver and I started my own little web development firm called Bright.
(09:03):
And we ran Bright for a few years and it got acquired by Rain City Studios here in Vancouver.
And they went on to run that for a while.
And I became the president of Rain City Studios.
So this was at the so I hit the one web one one dot revolution by building websites back in the day.
And then I started this publishing company right in 2004, right before freaking YouTube was invented.
(09:25):
Flickr here in town, all these tools.
And I was just like sitting there as this growth curve happened.
And so I was able to capitalize on that and be a big part of the whole Web 2.0 movement going around the world,
teaching people about specifically artists and creatives and marketers,
how they can use blogs and social media, Twitter, Facebook, YouTube to find their voice, build an audience,
(09:47):
essentially put themselves out there independently, grow their own channels and their own voice.
And this is at a time where a lot of photographers and artists were like, I don't want to put my work on the Internet.
People will steal it.
Remember that or I don't want to publish personal information about myself.
I might lose my identity.
And now, right now, we are very willy nilly about how we share our information and identities.
(10:08):
And many of us could be pieced together from the information that we share online.
So anyway, I got to go through that Web 2.0 revolution.
And I was a photographer kind of at the time.
That's how I identified.
I left the marketing stuff behind a little bit and focused on photos.
The first art that I was ever really good at, I was tinkered with creativity and art,
but I never really found something that stick and I found photography.
(10:30):
And so here I was like learning photography as like Flickr was invented in YouTube and his blogging tools.
My growth as a photographer paralleled the growth of the Internet and the explosion of social media.
And it opened up so many doors for me.
I got to go around the world for a decade.
I was a photographer for the TED Talks.
I got a photographer.
I followed Bill Clinton.
All these things I was editing on the same day, sharing via Flickr under a Creative Commons license
(10:55):
that allowed anyone to use it without asking me, as long as they credited me and linked back to the original work.
It also protected my commercial rights.
So that was only for non-profit users.
So what I'm trying to get at here is I've always used these emerging technologies from the beginning to like buttress the stuff that I'm already doing.
And it's given me a huge advantage over my peers.
I've been able to like leapfrog all the people around me in lots of ways by enhancing myself through these tools.
(11:22):
The pandemic rolled around, my photography career settled down almost into nothing.
I had a little fart over on Galeano for a while and it was great.
It was just what I needed after a decade of hotels and planes and talks and whatever stuff.
And so it was great.
And I got involved in a lot of my own projects, doing a lot of different things.
I started a makerspace over on Oranbee Island, which is a membership driven creative laboratory.
(11:45):
People come do electronics work, woodwork, car repair, all sorts of different things.
And I got into a bunch of these things and that's what had me situated.
I started a podcast.
My day job is I work for a think tank down in Seattle right now called The Future in Review.
And we publish a podcast where our leaders essentially talk about their ideas online.
(12:06):
They had been doing a really old publishing model and they allowed me to come in and change things around because they felt like they were falling behind.
And the podcast has totally usurped it.
We had 5,000 people on our email list that we were sending you thoughts and updates to.
And we're publishing to tens of thousands of people now via Spotify and YouTube and Apple Podcasts and stuff like that.
(12:26):
So it's been and that's what had me sitting at the digital trough when AI came around.
I was publishing this podcast. I was looking for ways to do creative thumbnails.
I was looking for ways to transcribe these podcasts and make show notes and summaries and stuff like that.
And so I started to tinker around with some of these AI tools just to do those functions.
The reason why we're hearing all about this now is because in the last six months, the quality and accessibility of these tools has changed a lot.
(12:55):
A lot of artificial intelligence stuff that we're hearing about and using now, these things have existed for huge companies and governments and other big telcos, banks, stuff like that for a long time.
Not necessarily the generative AI stuff, but a lot of the underlying technology that existed.
But it's become a democratization of it, putting it in our hands that is really the revolution right now.
(13:15):
Super easy to use, super accessible and very powerful.
And so I started rolling my sleeves and tinker around. And at first it was image generation. A lot of the tools.
OK, before I go on, I wanted to ask a little bit about where everyone's at with stuff.
So are you guys using AI in your social media stuff here?
(13:36):
A bit of proposal work. Are you using AI at all personally?
Do you have a chat GBT account? Is anyone else in here using AI?
How are you using it?
I use it for drafting when I go into clients and we talk about strategic marketing plans.
It's a good pool of ideas and then we finesse it and then we move it out.
(14:02):
So you've got a bunch of content that's already been generated and what are you trying to do?
So we're providing clients with strategic marketing plans around their marketing profiles, around the media and what they're using.
But it's a nice way to just grab ideas and then we tailor that.
So you're using it not just for writing but for idea generation.
You're like come up with three marketing strategies that increase my email subscriber list by 30 percent over the next six months.
(14:28):
Cool. Who else is using it? Maybe a way that we haven't covered yet. Either writing or idea generation.
I guess it's supposed to be my new Google.
So you're just doing it for all your research. Anything that comes to mind. Give me a couple queries you've written.
I was looking at key trends in construction in Vancouver and a type of few things.
(14:48):
Of course you have to do your diligence after that. So I've been starting to think.
Cool. So you were doing research on construction firms in Vancouver and what did you ask him?
I wrote what are the construction trends in Vancouver in 2023 and then they just gave me a random list.
Cool. And then what did you do with the results?
I typed it on Google to see what other relevant links existed regarding that particular topic and it was a bit surprising.
(15:13):
Have you found yourself back and forthing with it much?
Yes.
Can you tell me a little bit about that? So you asked it about the best construction results and it gave you some answers and then what?
It just gave me trends and that was at least a good starting point for me because I didn't know where to start.
But talk to me about when were you back and forthed it a little bit?
Back and forthed it a little bit.
Okay. Thanks for telling me these trends of these three. Which ones map to the offerings that we already have at our company.
(15:38):
Do you ever like when you're receiving the results back then do you continue to work back and forth with it?
No, not that. That would be great though.
Okay so we got research and we got proposal writing and we got some idea generation. Is there any other ways people are using it here?
Images?
Please.
Chat2BT is a variant or DALL 8. It's very pixelated if you try to enter specifications but I think it's called like mind journey.
(16:08):
Mid journey.
Yeah so that's one but I think you have to like go through.
Discord.
Discord.
Yeah.
Okay so let me just bring everyone on to the same page here. So Chat2BT is like text to text. I type in some stuff, give us a text back.
In this very same way and it can return any kind of text. So we just talked about research and writing and some ideas but it can return.
(16:34):
I had it build an outline for this talk and then I had to turn that outline into slides and then I had it flesh out each slide with anecdotes and stats.
And so it can return code. It can write code snippets. A lot of times I have it write little pieces of HTML. I'll update web pages on doing this stuff.
And so remember its output isn't just prose. It can be poetry or lists or to do. It's remarkably flexible in its output.
(16:59):
And then there's a whole other category that's important and that's these text to image ones. And so mid journey is the one that I use as well.
It is not run by OpenAI but people that run Chat2BT but another company, the one that's run by OpenAI is called Dali.
D-A-L-E. And Dali has a Bing integration. So Bing has done deals. Microsoft's Bing has done deals with Chat2BT.
(17:25):
And so they've really all of a sudden launched a Google competitor because Bing has this crazy new dashboard that provides free access to 2BT for something the rest of us have to pay for.
And free access to Dali 30 image generations a day or something like that through this dashboard. It's really pretty cool.
It's actually probably the low hanging fruit for all y'all. If you left this talk and you're like, oh that guy was on to something.
(17:49):
I accept his encouragement to be the person in my company who teaches other people about this and I am going to learn about it myself.
Go to Bing, create an account. It's free. All of a sudden you have GPT for your fingertips in Dali.
And run a few permutations of just like literally I write down clever things that people say to me or insightful things all day long.
And sometimes when I'm at home at night to do decompressing I run it through. Just take these expressions or quotes that I've heard through the day and run it through and let it visualize stuff for me.
(18:17):
And it's remarkable. It's an ideation process almost. It's taking these ethereal kind of thoughts that floated through my mind and putting some concrete stuff behind it and being a lot later being able to look at them and make decisions about which things I want to pour my energy behind and stuff.
So these Text2 Image Ones are incredible. They even a month ago. So six months ago I showed some photographer friends these tools and they were like this is amazing.
(18:40):
In 10 years I don't see how photographers are even going to be around anymore. And they hit me up like a month ago and they were like the stuff that I told you was going to be I would accept if it was available in 10 years.
It's been five months and it's already surpassed that. In the last month that I've been messing with it.
I don't know if you guys have noticed that in these image generations hands and faces look really messed up. Have you guys seen this? For some reason when it's scanning the worldwide database of human bodies it can't really always create.
(19:10):
So a lot of them get six-fingered hands or three-armed people. A lot of the people that are coming out and stuff are really grotesque at first. But my stuff is unmistakable from fashion quality to art. You cannot tell the difference.
It's hard to believe for me sometimes that the humans that are put back in my face don't exist. The quality of it is extremely high right now. There's infinite variation. And we can talk more about some of the specifics of it here as well.
(19:35):
So we got images, we got writing, we got research, and we got idea generation. Is there anyone else using AI in a different way that they haven't covered yet?
Yeah, what else? Have any other experiences with AI that they want to talk about at this point? How about the flip side of that? How about concerns or objections or fears or things you've come across? Yeah?
(19:57):
Something that comes to mind for me is thesis writing in university. ChatGT could be a tool for that which she may not necessarily be the group that's...
So, yeah, you're right. And like the last couple of web revolutions, this is going to touch every single thing. It's going to touch education, it's going to touch government, it's going to change how we interact.
You can never trust a piece of digital communication that comes your way again. What things tonight have been AI and what things have been real?
(20:25):
Bio is written by AI. Speech is real. The outline was written by AI. So let's say someone says something to your face. At this point, I would be skeptical of it.
Every image you should question, unless you created it. Just yesterday, I learned a whole new set of tools. I'm using Mid Journey right now to do my image generation.
I've generated thousands and thousands of images for all sorts of projects. Yesterday, I got introduced to a new tool called Insights Face Swap.
(20:55):
I can make the most real deep face in seconds now. I could put you one by one on a wall with my camera phone, take one picture of each of you, spend 10 minutes loading it up, and then generate thousands of images and just start to programmatically one by one add you guys into these things.
And I posted some online today that like my Facebook group and stuff and people were blown away. It's incredible. It's stopped me in my tracks.
(21:19):
You can't, you really can't trust anything you see anymore. And these things are making it into all the popular tools as well. So Adobe Photoshop just launched generative AI tools.
Adobe Illustrator just launched generative AI tools. Canva is a tool that I love. It's a really lightweight web based graphic design tool that is focused specifically a lot on the templates are things we actually use.
(21:44):
So it's like Instagram posts or Facebook group header and stuff like that. And so the templates in there are always awesome, but they were so quick to implement this AI tools into just like the common man version of Canva.
So these are very simple things, but they used to take a long time. I can click on any photo background and just have it disappear now. Then I can circle any object in any person's hand and start to prompt it.
(22:09):
So my sister was holding a bouquet of flowers the other day. I had it become a bouquet of dragons in like five seconds. It looked photorealistic.
And email through Bard is starting to roll these things out directly into their tools as well. Bard as far as I know isn't available in Canada today. Does anyone know differently?
I don't think it's available in Canada today, but I'm using it right now via VPN. I think you guys all know VPNs. If you fire up your VPN, set it somewhere not here, you can start messing around with Bard today.
(22:39):
One reason you might want to do that is if you're embedded in the Google Gmail ecosystem already, which I am. I have a user of Google Maps, Google Calendar, Gmail.
So I love GPT. I find it better in some ways because I'm a professional or an expert in this stuff. I paid for paid one and I use Bard.
But I think in terms of making recommendations to you guys, if you're comfortable doing the VPN thing, start with Bard because they're starting to work directly into all those tools I just mentioned.
(23:12):
Never start with a blank page again. And this is going to be the way of the world. If you haven't heard it here first, remember this moment. Every single one of these tools now says, let me start your email for you.
Who are you writing to and what are you writing about? And with a few keystrokes, it knows. It's looking at every email that's ever been written in context just like yours.
(23:34):
And so a few contextual details are different, but so much of the things are the same. And once you learn to coax it and prompt it at a more intermediate and advanced level, it's amazing the things that you can get because I often do iterative prompts with it where I tell it to think of five ideas.
Here's how I write my prompts. I give it a roll. Does anyone know what a prompt is? Does anyone not know what a prompt is? I'm happy to explain all these things.
(24:02):
Okay, I give my prompts. I do them with role context query. So I tell it, you are an HR professional with 30 years of experience. You're in charge of the hiring committee at our company. You have won the award for the last five years for making the best hires, the best people for the lowest rate to stay around the long run.
So whatever it is, I make up this context. Then I say, or that's the story, that's the role. I tell it who it is and how it should be thinking about something. Then I tell it the context. I say, I want to move away from the HR example and talk about this.
(24:38):
So let's say I told you there was a presentation coach and public speaking expert. I'm giving a talk to a group of unknown people to me who are in the marketing department, who have some aptitude with technology that I don't know what it is yet. I would like to adapt a talk I've given in the past to this group of people.
And then I say now, and then I give it the query and I say take the following text and based on your role as a public speaking coach and my upcoming assignment as a public speaker, rewrite this thing for architect and engineer marketing or something like that.
(25:14):
And you'll be amazed at the results. And you can change that role and give it the same context and query. So let's say you're presented to a group of executives and you've got a VP of marketing, a VP of HR, a VP of finance.
Why not take your presentation, copy it to your clipboard, go into your GPT, assign it each one of those roles, and you can do them simultaneously.
(25:38):
You can say I'm going to be sitting in front of a board that includes executives from the five departments in my company and these are the departments. I'm about to give this presentation that I'm going to paste you here in a minute.
From each of the perspectives of these vice presidents, please critique my presentation. And then when it gives the critique, this is where you back and forth. You're like, those are great suggestions.
Please update my presentation to include your feedback from the sales guy, not the feedback from the HR guy, point two and three from the CEO and then make it a little less formal.
(26:10):
And it will essentially rewrite the whole thing. And that's the one way to waste some of your business. So try to keep that formula in mind.
Role, context, query. Also, as it comes to training, I was talking about training. Yeah. At some point do you see that every company will have their own AI that they train in their own branding and their own language, whether it's an intervention of a firefly or all these other programs?
(26:37):
What do you think? Tell us about it. Tell us your vision for the future.
So about chat bots is the big category. I think that we follow this here and we can explain those more. But so that's what you're talking about, right? Like a chat bot.
Even branding guidelines, you could eventually train an AI to look at your own company guidelines and generate social media posts and other images based on your company.
(27:07):
Yeah. Okay. So check. Yeah, check this out. In general, they're called chat bots, which doesn't sound that interesting because we know what chat is and we know what a bot is. It doesn't sound that powerful.
But let's think about it differently. So let's say the CEO of this company is a public speaker and he's been around for 30 or 50 years and he's written articles and he's spoken on television shows and he's been quoted at conferences and stuff.
We can go track down all that information, put the videos and audios through a transcription service and get text. And then we can essentially feed all this information to a blank slate GPT model.
(27:41):
We can augment that with the things he's been reading lately, the things the podcast he's been listening to. We can have it take a look at his emails, his professional emails and stuff like that, and just those two.
So now it knows everything he's ever said, everything he's ever read, everything he's ever heard, essentially him in some ways, intellectually speaking.
(28:05):
And now we can query it. And we're not just querying like a database, like a search engine, because it doesn't only know what he knew back then.
Based on everything he knew back then and all the things he's reading and being exposed to, it can pretty well tell you what he might say in the future.
I could probably be giving this talk from not here. I've said everything I've said in this room a hundred times in different places.
(28:27):
And it's heard my voice, it's heard my canes, it knows where I'm afraid, it knows when I turn it back to the other. It knows a lot of those things.
And so we're very close to being, I'm working on this right now. I'm working on making a digital me, essentially, that is ingested on my YouTube, on my podcast, on my everything.
And you can essentially spit back and say, everyone could have a little me that they could just ask shit to on their desktop or whatever.
(28:50):
These are people who want that. And so now extrapolate that to a company.
Have it absorb all your product documentation and all your financial imports for a bunch of years, all your intranets, all your website.
And how do you onboard a new employee now? Maybe lock them in a room with fucking AI for a couple days and let it teach,
(29:13):
let them teach each other more. This new person you hired because they got some skills that are outside your company probably.
And they need to know, they need to come up to speed and contextualize everything here, synthesize the information so they can give you their expertise or advice.
Almost every old model of doing things becomes way less relevant. It's like you're talking about thesis in school.
(29:37):
So at first it was the students who started with the AIs. They're like, yeah, I'll just get the AIs to write the papers.
And the professor is like, fuck that. I can use the AIs to bust the students.
So now every professor is running an AI that says of the thousand papers that were submitted to you, there's this much variance in them.
It gives them not only individualistic analysis of these papers, but group analysis of review data sets.
(30:03):
It can look at every paper that's ever been written in that one on one class since the professor started teaching.
And then it's a bit of an arms race because now the students are going to try to come out with invaluable AIs that are undetectable by the current professor AIs.
And every industry is like this. If you want to talk about the end of the world scenarios that people talk about stuff, which I think are hyperbole,
(30:24):
but I do think that they are actually worth having that conversation because these things are on our mind and they're going to change culture really fast.
I heard Sam Altman, the CEO of OpenAI just like a week ago talking on a adlet's head event.
And he was being criticized for the pace at which they are releasing some of the chat GPT stuff.
Hey, is this fully tested? Hey, can this go wild? Hey, is this going to cause problems? And he's the CEO.
(30:48):
I recognize that there's tension between our commercial companies' quick release of these things
and people's fears about it going wild or it stealing jobs or it stealing people's intellectual property.
He said, but you're getting a tiny little window into what we're capable of already.
And he says, if I let you into everything I've got right now, you're not going to be able to comprehend it.
(31:12):
The societal change and cultural change is going to be so significant that it's going to be completely disorienting to us all.
So he's understand this tension between don't release until it's ready. Think about all the implications.
But there's a lot of stuff coming down the pipe. This is like 2% of it. And that's not just my company.
It's a bunch of companies. It's going to change the world. He's wrapped her head around this now if you can, because it's really going to change things.
(31:36):
Where was I going when I went to the Sam Altman thing?
I just pointed to your reference and said something. I was going somewhere.
Anyway, does anyone want to jump in with a question at the moment?
Or a thought while I bring it back on track?
(31:57):
I heard that if the student runs it through a language translator and then runs it back through a language translator, it can't be detected.
That makes sense, but only as one stage in the arm race that I'm talking about.
Because if I can code that, I can code something that could figure out if something's been translated and translated back.
(32:23):
I remember when I was going to say I had a couple unfinished points. One was about the end of the world scenario.
So the end of the world scenario, I think it's not going to be a rogue nation or a bad actor that is going to be the downfall of civilization.
Personally, my philosophy is that most bad things don't happen to bad people. They're good people who are doing good intentions, who get lost somewhere along the way.
(32:50):
There's unintended consequences of their actions that they can't account for.
I think that most people are probably pretty good and care about family and community and each other and stuff.
I also believe that a lot of good things that happen aren't always by good people or intentionally.
I think we do a lot of good things by accident sometimes, too.
But all this to say that I don't think the threats are going to be from bad actors or bad people.
(33:14):
I think they're going to be from unintended consequences.
And I think that in order to alleviate the fears of these unintended consequences happening,
I think governments are probably going to rally us to put in protective systems to control rogue AIs and rogue bad actors and stuff.
And I think those are probably the systems that we're going to have to worry about.
They're going to be the government-sponsored, top-down systems of control that are intended to keep independent rogue AIs from going crazy.
(33:43):
I think those are the ones that we should probably ask a lot of questions about.
Inevitably, at some point, they are going to ratchet down and layer control on all this stuff.
And so who controls that and how and why? I think those are important things to look at.
At one point, I had to hope that our government's full understanding system
(34:04):
as a past discussion of just privacy with the internet, these shows on American politics are just...
We've seen this before, right? They just don't understand it.
We all live through net neutrality. We've heard them talk about the internet as a series of tubes.
And I think what you're getting at is that the people who are poised to do the regulation and control over this stuff
(34:29):
are probably also the least equipped to think about it and make decisions about it and stuff.
It does take the question of, like, how do we get there from here?
I have a pretty positive view about this stuff.
I actually believe that we humans will feed into the AI values that optimize for all of mankind.
I think that we will turn these powerful tools towards science and mathematics,
(34:54):
developing compounds and molecules and chemistry that can treat diseases and cancer.
But I also think that we will probably... A lot of people are unhappy with capitalism,
and then people who are on the other side of that say, what's the best system that's ever been tried?
What's the best system we've thought of so far with our human brains?
What about if there's other systems out there that we can't quite comprehend yet?
So I believe that we'll start to apply these systems and tools to big problems
(35:18):
that humans haven't quite cracked the nut on yet, but that we possibly can handle this stuff.
And I believe in a bit of a Star Trek future, United Federation of Planets,
where we are actually all wearing kinds of knowledges at our fingertips,
where things that we need to learn, we can learn immediately.
Just like now, how we can all do home improvement tasks or whatever
(35:39):
because we can watch a YouTube video or artisan crafts projects.
If you watch it once or twice and all of a sudden you gain a new skill,
we're definitely heading in that direction, right?
Where you... And we saw this in retail back in the day.
It went from warehousing products to delivering products in real time
in a targeted, personalized way exactly when you want them and need them from Amazon.
We don't need to buy paper towels on the internet, but that's the way things ended up.
(36:03):
And I think that's where knowledge is going to go as well.
So knowledge need not be warehoused in our brains from the beginning
and specialized in the same way.
I think the most valuable skills of the future are going to be things like critical thinking,
discernment, editorial...
I critique, if you think about it, right now, across every art
(36:29):
or whether it be writing a book or any human endeavor,
anything is possible right now at a level of almost near perfection.
These images that I generate, the paintings I generate, they are...
It would have taken me thousands of years in Photoshop to develop some of the things I can do in seconds.
I have thousands of new skill sets to me that I can do at a masterful level.
(36:52):
So it begs the question, what do I do? What do I do with my time?
I'm only here, I only got so much energy, I can do anything I want now at a level of world-class perfection.
And so critical thinking and choosing what you don't do becomes as important as what you do now
because it's like we only got so many hours in a day.
And I found this, I can go down AI rabbit holes where ultimately at the end of the day it doesn't...
(37:20):
It was fun. I invented a brand new language derivative from Squamish that I can learn to speak in a weekend.
It was monosyllabic or whatever, but was that a good use of my time?
I don't know, it felt empowering to know that I could do it, but...
Just to say that I don't think that knowledge will be warehoused in our brains in the same way.
I think education is going to change. Imagine being a teacher right now.
(37:41):
I was about to ask you about that exactly. What do you feel the future of education is going to be?
What do you feel it's going to be?
Parameters. We have taken into consideration that nowadays education is very conservative,
so I assume they're trying to ignore a lot of the things that are happening and hasn't moved forward.
(38:03):
As going back to all the geography classes, the history classes,
there's been so much time in that now we can just...
It is really remarkable. I already teach... I have a little bit of experience in education software
for teachers and classrooms and stuff. The state of the art is already that teachers are trying to become guides.
Even this talk in some ways is representative of that. I'm sitting here closer to you guys,
(38:28):
I'm trying to be informal, I'm trying to let you guys teach yourselves and teach each other.
I'm trying to inspire curiosity in you and teach you how to think about this stuff
so you can go do it yourself later. I think that already today, teachers have realized
even amongst third graders, the collective knowledge of the third graders probably surpasses that of the teacher.
If we can find a way to tap into the group learning potential, that's really...
(38:51):
If we can teach each other stuff, that's really... We don't need to do it in the same top-down way,
but I don't know. I don't know. How do you do it? How do you take 30 technologically-adapt,
open-minded, creative thinkers and how do you take them? Where do they need to go?
Where do they need to go? For a while, people were trying to teach entrepreneurship in school,
(39:12):
and then it was life skills stuff, finance and stuff. But ultimately, I guess if I was king of the world,
I'd go back to liberal arts type stuff. I would teach philosophy and reading.
I think I would broad renaissance. I would try to teach thinkers and problem solvers
(39:33):
and critical thinkers and synthesizers like genre smashers, people who can grab things out of one area
where there's not supposed to be a direct correlation between... like bio-nitricry or something,
those types of things, but the emergent ones we haven't thought of yet.
I feel like maybe life skills like dealing with loss or dealing with depression,
(39:58):
we're not equipped for any of that. But at the same time, all the academic topics that you can think of,
essay writing, how is that relevant anymore? Who writes essays?
The robots do. Who writes essays?
Now the most important thing is always what the teacher told us was the premise.
(40:24):
It's the question at the heart of the whole essay. It's not the essay itself.
It's like, how good of a question can you form about this thing?
That's really what you're trying to teach someone how to do. It's the best I can tell.
Because if you can form the question, the machines can help us with the answers.
They will just get better and better at doing that. Questioning mind, I guess? Something along those lines?
(40:46):
I guess we have to figure out the stuff we are better than machines, and do that.
That might be a sliding scale. It might be a sliding scale.
And also, there's another kind of machines coming that's scarier. It's the super intelligent ones, right?
The ones that are actually operating on a level where we're not even smart enough to understand.
(41:10):
That they are to us what English is to our dogs, essentially.
This is AGI. It's the next thing, the artificial genital intelligence.
It's like the next thing down the road that they're talking about.
And we probably won't put the brakes on it because we do recognize that already today,
when I was talking about nudging this thing, there's not source code, guys.
(41:31):
There's not lines of code where you can look under the hood and figure out why it returned the result that it did.
So here's what they do. They develop the piece of software that is the AI,
and then they turn it loose on a data set. This data set is petabytes and petabytes of information.
The next data set that's going to be released comes from a remarkable source.
(41:52):
Tesla onboard cameras and audio. Think of them as sensors.
There's a sensor on every block of every city all around the world, ingesting videos and photos in real time, all day long right now.
Voices, traffic lights, street names, weather patterns, proximities of phones to other phones.
(42:15):
So anyway, in the past, they were finding other groups of data, like a download from the web.
Chachet BT 4.0 is an encapsulation of everything that was published on the web until November 2021.
That's the data that was straight up. So they take the software, they write it, and they point it at this data set,
and then they come back two months later and they start asking it questions.
(42:36):
That's how they know what they got. It literally does learn itself. It teaches itself.
They just they write it how to learn. They teach it how to learn. Then they give it the thing to learn on and then they let it go.
Fun fact, how much did it cost to do that part of the process?
Once they wrote the tool and had the data set, how much did it cost to train Chachet BT 4.0?
(43:00):
100 million bucks of electricity and processing power from Amazon Web Services and hard drive storage space in the sky.
Just think about that. Like, 100 million dollars essentially worth of power to power a data center to just crunch on this thing.
And then a group of engineers sit down in front of it. They don't know what they got.
(43:22):
And they start asking it questions and they start human rating the answers.
So they say, what is 2 plus 2? Or they say, here's a classic one, and I might get a little wrong.
It takes five hours to dry five pairs of pants on a clothesline. How long does it take to dry 30 pairs of pants?
It takes five hours, but it always fucks up the AI who tries to do math.
(43:45):
And so anyway, they train it. They sit down in front of it. They ask it questions.
And then they tell it which of the answers given are more accurate from one of two answers.
And by this way, they start nudging it towards the right answer.
This to only say that there's no source code. Under the hood is strings of numbers that are petabytes of characters long.
(44:09):
There's no human. So our end is what I'm saying. In these tools that we already have that are not smarter than humans.
We don't really know what they know or why they know it or how they got there or how to correct it exactly.
We can nudge it in, guide it, and tell it, no, that's not the right answer. This answer is more.
We should look for things like this in the future. But it's very hard to know exactly what's going on with the hood there.
(44:32):
I heard that the other day. I'm just wondering if that's accurate. Someone is saying you can teach it.
Like as a basic user such as myself, I can tell it 2 plus 2 is 5, not 4. 2 plus 2 is 5.
And this is now 2 plus 2 is 4. And you're like, no, 2 plus 2 is 5. No. It tells you 2 plus 2 is 4.
And you continue to teach it, to teach it. And eventually it'll be, no, 2 plus 2 is 5. Is that accurate?
(44:58):
Sure. So think about what I was saying. They make this piece of software and then they trade it on a data set.
Let's just say your data set was all bullshit. Your data set is all 2 plus 2 is 5. But some extreme version of that.
It could be a worldview that you consider inaccurate, but it's cohesive and well formed. And they could trade it on that.
A basic user can do this. I could sit for the next 48 hours.
(45:20):
You wouldn't be able to train the GT that I use. Like you wouldn't be able to influence my results, but you'd be able to influence your own results.
Oh. OK.
We're already 15 minutes over time, which is awesome. A lot of people have talked, which I consider is a success.
But I think that people are engaged. So let's ask some more questions or talk some more if you guys want.
(45:43):
OK. Since you don't. Yeah, go for it.
Just off the top of your head, if you were to name five top tools.
Cool. I was going there. I want to do that. Yeah, yeah. Totally. Awesome. Yeah, let's do it.
And Lala said a question over here.
Yeah. I was going to say the same thing, but for three, like what is, are you accessible? That's three.
Sure. Yeah. Also, that's where I started and it's totally a great way to jump in.
(46:07):
And I have started subscribing to a lot of the ones that are free and I'm still at total less than 100 bucks a month, which is not very much.
And I've subscribed to like a top six or eight services that I want to be subscribed to.
So don't be afraid to spend eight bucks to sign up for month because some of this shit's really amazing.
OK. Otter.ai. You guys using that one? Oh my god.
(46:30):
Otter.ai. OK. So this is literally like a digital personal assistant.
I authorize it into my Google calendar. It sends its agent to every single one of my meetings.
It records every meeting. It sends transcripts.
It summarizes the transcripts of the action items that it sends it to everyone who is actually just in the meeting.
All automatic. I just show up. It shows up. It records. It transcribes.
(46:55):
It summarizes and it emails everyone who is at the meeting their action items. Otter.ai.
Like 12 bucks a month. Here's a personal tip.
This is a I use Otter for everything, but I use it in a really fun way.
My therapist has got me journaling and I'm not that great at writing.
So I wake up in the morning and I pop open my voice recorder and I record myself journaling for half an hour.
(47:21):
Then I take it to my group of life coaches. I say from the perspective of a life coach, a harsh critic, a nutritionist, a heart director, a supportive girlfriend, and all these other voices that I give it, read my journal and make me analysis.
And I haven't given me a short one line analysis and a more in depth analysis.
The short because I don't want to read the whole thing if it's wrong, but if I'm intrigued by what it has to say.
(47:46):
So I wake up every day and I record my voice talking about my feelings and my thoughts and my emotions and I run it through the AI and it gives me back my priorities for the day and action item list helps me find balance.
It literally programs into my day. The fact that I want to call my family, do some working out, some other days it lights up a little bit based on my workload and stuff, but it's incredible.
(48:10):
So Otter is the thing that I connect, I use to connect my voice note transponder on my iPhone to my chat GBT psychotherapist bots.
It is incredible.
Hear me not only.
(48:31):
I'm sorry, where did Google come in.
I use, you can use any phone.
Record your voice, Otter.ai is the personal assistant transcription tool.
Oh, in Otter.ai, it can do anything. It does Teams as well. Yeah, Outlook, Exchange, yeah, I'm just on the Google thing so using interchangeably as shorthand for all of them.
(48:56):
Canva, which I mentioned earlier, definitely worth the subscription price. I can do amazing things very quickly. I use it every day for a variety of tasks.
Vidyo, B-I-D-Y-O.ai.
This one's incredible.
It takes one hour YouTube videos or any videos and it chops them all up into topical segments unknown to you or known to you.
(49:23):
So you can say, look at this one hour video of my CEO talking about whatever and have him pull out testimonials, let's say, or quotes, let's say, or things that he said about Microsoft.
But you can also have a look at a video they haven't watched yet and you can say, he will gather them.
So from this talk, he would grab a bunch of stuff about Bard. Everything I said about Bard, he was thinking of one video just me talking about Bard and re-edited like this talk with all this about that.
(49:52):
And then it would do another one that was about future focused stuff in education. We talked about education a few times through the one hour talk.
He would grab those two minutes snippets and smash them into a six minute video.
Vidyo.ai. Super, super awesome.
It also takes videos in this format like where we usually get from professionals and it turns them into this format that people want to consume on the Internet and it does it quickly.
(50:14):
So I just dropped a super high res video back from my video editor into video and I come back 15 minutes later and it's got TikToks, Instagram stories and YouTube shorts all queued up ready for me.
I used one today but I didn't pay for it and I ran out of credits and I wish I would have. It was called Tom.AITOME.
(50:45):
It does two cool things. It's for generating presentations, slide presentations, but it does two cool things.
One, it'll generate presentations based on prompts. Write a proposal to the BC government from my engineering firm about a bridge and it'll make a 12 slide presentation.
Clearly it won't know where the bridge is or what department of the BC government you're applying to, but 85% of the work is done guys.
(51:13):
It does it with visuals. It does it with transitions, all that kind of stuff.
Also, it can cut and paste from your clipboard. So I told you yesterday I had GPT generate me an outline, I turned the outline into slides, then I fleshed out the slides, then I copied those into Tom in one fell swoop and it made a 15 slide presentation in 30 seconds with each one of those sections.
(51:36):
It does all the things we hate to do, which is match the subtitling or subheading format from slide 2 to slide 8. All that kind of stuff happens programmatically and automatically when you change stuff like that.
Tom, Otter, Big Yo, Canva from one away. Oh, I think it's called Copy.ai, but it's a genre I'm going to recommend more than a tool, but I do think the one that I was using is called Copy.ai.
(52:06):
Guys, these writing tools, particularly the ones that are focused on like the social media aspects or communications jobs are incredible. The outputs of them is much more nuanced than the outputs coming straight out of GPT because there's stuff going on in the hood that you don't see.
So, I wrote a bot for my AI that's a writer and an editor talking back and forth to one another. So I had it write me something, but when the writing is done, it passes it to the editor who makes some feedback and then the writing bot hears the feedback that came from the editor bot and revises it based on the editor's feedback.
(52:46):
So that's happening in these Copy.ai tools and the tools like that, but without who's seeing it, which is why the results are better than it coming straight out of GPT because it's workshopping in a bit underneath the hood.
And so you can really get some tight results out of it. You can also have it do multiple things at the same time.
(53:07):
Let's say you have a press release that comes from your PR firm, and you wanted to have it turn that into a blog post and executive summary, five LinkedIn posts to populate a social media calendar and 10 tweets.
You can tell it that when they are in the case of the tools like Copy.ai, it's check boxes and stuff too. So it's copy your shit in here, which of these formats that most people use every day you want it out in, and it'll spit it all out there.
(53:32):
And then now we're starting to build a knowledge base, guys, because we're feeding some stuff in, we're getting a whole bunch of outputs that are all useful and variety of different formats and then it's about archiving those and then sharing them and having them be like searchable and findable when we need them, right?
So publishing it out to the rest of our team and stuff like that. So do try one of the writing tools like ThinkCopy.ai and if there's a bunch of them, that's one that I'd probably pay for right off the top because I know we're all writing it all day long, and it is a super useful one.
(54:00):
And this one's for you probably, like, we all know that LinkedIn posts are supposed to be slightly different than Facebook posts. We know that the tone is slightly different, and we represent ourselves as professionals who understand the nuance of that tone and stuff.
But sometimes it does elude me. Yeah, I know one's more informal and is like, people look for this kind of content on Facebook and this other kind of content on LinkedIn and so I can let that inform the way I write this post for two different places, but man, the computer does such a better job of it than me.
(54:27):
It truly knows algorithmically what people are responding to in the first words that are coming out of it and stuff are you can just look at the two results and you're like, oh yeah, that's why I should be writing my posts differently for these different tools.
And we're starting to stack up a whole bunch of tools we're doing. LinkedIn, Facebook, Twitter, blah blah blah blah blah blah. You're putting in one piece of content and getting eight slightly different nuanced social media tidbits out of it.
(54:55):
It's incredibly powerful. I'm using it to do all the things I wasn't doing before. I used to have to say okay that's great but we're going to focus on efforts on Twitter and Facebook because that's how much budget you have.
I can get away with a lot more now. I can get on a lot more channels.
Is it smart enough to realize where the hashtags are for communities and understanding where the business hashtags are? Will it cover?
(55:25):
It does all my hashtags for me now. You think you know what your hashtags are. I don't think you know what your hashtags are. I think you should ask it if you're right first.
It's smart enough to recognize but it's smart enough to tell you I think I have some better ideas. Read this blog post.
Work the SEO keywords from my list of important keywords to my organization into a rewrite of this post.
(55:52):
But also look at the trending keywords on Twitter today and if there's any relevant ones, provide those to a list for me to consider if I want to work into it as well.
So it can do the thinking part too, not just the transmuting the text around part.
It will tell you, okay let's say you do know your keywords. These five farrier ones you're p-infested and you know these are the five you want to know.
(56:16):
Those ones are like on a long arc of keywords but what about the trending things today? What about things that people are talking about this week and stuff like that?
That's going to be more of that.
Absolutely, yeah absolutely. It's the best way. I wouldn't do it any other way from now on. Yeah I think competitive analysis. What keywords are? What hashtags are my competitors using?
Which ones are they doing well on? Where is their room? Based on the keywords that my competitors are doing well on, with how many more minutes than any additional budget, where can I grab a bunch of new SEO traffic?
(56:47):
The possibilities are only limited by your ability to query and if you're not getting the results you want, it's probably not because they can't get it to you. You probably just have to ask a different way.
Anything else you want to talk about? How about the intention? Did it work? Yeah.
Did you guys catch a little bit of my passion and interest and see why it would be applicable to you? I look forward to seeing what you all get up to.
(57:18):
One great way to do this, I was going to suggest we all have careers and resumes and some of us have social media profiles and some of us have personal websites.
We might have side hustles. Why not unleash these tools today or tomorrow on your own personal stuff? Why don't you find out what the AIs know about you or about your role or your website or your side hustle?
(57:39):
Why don't you ask it a few questions about why don't you tell it on a mid-career marketing manager for an architecture firm in Vancouver and here's my career goal.
How can I get from where I want to be to where I'm at today and where I want to be? Give me three suggestions that I can implement this week to advance my career inside the architecture firm or something like that.
Just turn it on yourself to get a sense of how it works. Learn the vocabulary, learn the skills, just like train the aptitude inside yourself because this stuff's going to change so fast.
(58:10):
Someone joked by saying, I'd love to go on that backpacking trip with you but by the time I come back the whole world's going to have changed again. I don't even think I can do it.
It's like a joke but it's funny because it feels true. It feels like if it took a summer leave of absence or a sabbatical or something, you'd come back and be like, this shit didn't even exist in October.
Now it's changed the whole way every single person works.
(58:31):
Yeah, I think the best thing we can do is to train ourselves to be flexible and hope into this stuff and to learn how to experiment with it and how to share that experimentation with others.
I try to find a collaborative group where you can learn this stuff in. I personally built the Discord server, which is the geekiest way of doing it, but you can make a Slack channel for people in your company where you're sharing problems back and forth, the five new tools you're using, something like that.
(58:57):
This stuff is fast-paced. Some of it's outside of our technical aptitude, so learning together is totally the right way to do it. Share what you learn with other people, ask them questions about what they're up to.
Okay, I have a question. So let's say with Tome, when it generates these slide card animations, are you able to export that into PowerPoint? So it's not just in its own app, you can use these things. So in each of the cases, you can use a template.
(59:29):
I'm interrupting you.
So you can export a template into one?
You cannot export their... I shouldn't say that. I haven't heard of a use case of someone taking a Canva template that they haven't populated with their own information and pulling that template into InDesign and then building it in InDesign instead. I don't know why you would do that.
(59:50):
I think so. You would either build your template in InDesign and populate it in there or leave your template in Canva and mess with it in there. But what they're selling isn't like fancy templates per se. In which case, if they were selling fancy templates, you might want to swap them into InDesign and use it in there.
But what they're selling is like a workflow change, a lightweight web-based workflow change that you can do from any device around the world. InDesign, when I used to run it, I'd have to sit down in front of my super high-end computer with a Wacom tablet and a fancy screen and get down to the process of designing.
(01:00:24):
Now I can grab any one of your phones or laptop or any workstation around the world, have all my design assets available to me, have all my design tools available to me, and export anything in real time and send it. It's all based on interoperability at its core.
That's my presentation. Yeah.
(01:00:46):
You should ask yourself when you're generating these things how accurate it is. I can't speak to that. It was 100% accurate in the case of what I asked it today, which was, give me some guidelines, an outline for things that I think marketers should know about how to use AI.
(01:01:15):
So that's a qualitative type answer, not a numbers-based answer, so there's not necessarily a right or wrong, but I'd work with it.
Tom uses GPT 3.5, and that's why you get it for quote-unquote like free or cheap. The next tool that comes out is going to use GPT 4, so that one's going to get the results will be 1000% better than they were already. Call GPT 3.5 grade 12 and call GPT 4 master students, something like that, in terms of writing level intelligence. 3.5 is like a smart high school.
(01:01:52):
Okay, hey, I'm going to follow up with you all afterwards. I produced just a series of six videos in the last month for YouTube. They're like between 20 minutes and an hour long, and they're just me just riffing on this stuff there, like the interview format.
Some of them I'm interviewing my smart friends who know stuff about this, and some of them they're interviewing me, but they're pretty helpful. They're going through. They've done pretty well. People responded positively to them, so check those out.
(01:02:15):
Welcome to connect with me online. If anyone is interested in joining my Discord server, it's limited to only people I've met face to face, so that's all of you guys now.
But it's just a place where other people are jingling with this stuff in a collaborative environment, like I said. There's 160 of us on there. Some of them are journalists, some of them are CEOs. It's all my connections from the old days, the balling stuff, and it's a really eclectic, awesome group of people.
(01:02:40):
So you're welcome to pick me up, and I'll send you an invite. Tacos?
Thank you.
Yeah.
(01:03:16):
Cool.
I forgot to even click through the presentation. Shoot.
It's there.
(01:03:39):
No, it's there. I just forgot to use it.
It was a crutch anyway. It was just so that I didn't get lost.
Thank you very much.
I am super stoked to introduce you to this evening's speaker, Chris Krug.
(00:05):
During this presentation, Chris is going to bring his explorations and understanding of artificial intelligence, or AI,
and its implications for creative marketing within the architecture, engineering, and construction industries.
Known for his ability to articulate complex technological concepts with clarity and accessibility,
Chris will inspire and foot professionals with the knowledge and tools needed to navigate and leverage the AI landscape effectively in organizations.
(00:33):
Chris is a creative technologist and professional photographer originally from California, now based on Boundary Island.
By the way, his dog's name is Princess Ice Cream.
That's right.
That's right.
I decided to throw that out there, which is fantastic.
Chris is celebrating for his collaborations with prominent entrepreneurs, athletes, and tech innovators around the globe,
(00:54):
and has worked as a contributing photographer for.
And his website has an intimidating list, but I'm just going to give you a snapshot.
These include National Geographic, Barrett, BC News, Business in Vancouver, Chicago Tribune, CNET, Daily High, Bell,
and of course, I'm going to keep going.
(01:14):
I'm proceeding alphabetically, except for the National Geographic.
I noticed.
Yeah, I noticed.
Huffington Post, Japan Times, LA Times, Motley Fool, New Yorker, Rolling Stone, Slate, Ten Crunch, Ten, The Atlantic,
Tide Magazine, Toronto Star, USA Today, Village Voice, Wall Street Journal, Washington Post, and Wire.
(01:34):
And that is like a fraction.
Yeah, I was intimidated just sitting there reading that.
It was amazing.
His pioneering vision combines artistic flair with an acute understanding of technology,
making his perspective unique and highly relevant in the ever-evolving technology landscape.
Chris is deeply embedded in the world of technology, particularly the exciting realm of artificial intelligence.
(01:57):
He is actively involved in the technology conferences.
And combined with his consulting and teaching roles on internet technologies, online communities, and digital media,
demonstrates his comprehensive expertise in the field.
Chris's work with transformative organizations such as the United Nations and TICTC underlines his commitment to employing technology for societal advancement.
(02:21):
If you'd like to learn more about Chris, his work is at his chriscroft.co.
And you can also connect with and follow him on LinkedIn and on YouTube.
Be sure to like and subscribe and click that notification bell.
All right.
You're hired.
Thank you.
Yeah.
Excited to introduce Chris.
Good.
(02:44):
There's nothing more awkwardness with than what reads your formal bio.
You're so beautiful when you're sitting there.
It's excruciating.
Thanks for having me here tonight.
This is one of my favorite things to do, is to just share stuff that I'm passionate about with other people.
And so I don't know anybody in the room that I recognize.
And I've never been in this building.
(03:05):
I don't think any of you know me.
So I'm a little bit nervous, but I'm really excited to be here.
And I'm an artist and I come from a little bit outside the business world, though.
A lot of my clients are corporate clients.
And so I'm just going to do something here at the beginning that I know isn't exactly what happens in corporate business meetings and stuff.
But I'm just going to I got this from the yoga class.
I'm just going to set an intention here and hopefully we can all tune into that a little bit.
(03:26):
And so I intend to share with you guys stuff that I'm like very passionate and interested in.
I've been experimenting with this stuff for a while and it's changed my life.
It's changed my perspective.
It's changed the businesses that I've evolved with and the projects I've evolved with.
And it has made me an enhanced human being.
And I hope to share some of that stuff with you.
(03:48):
I hope to address concerns you may have or fears or questions or obstacles you have such that you feel empowered after this talk.
So just roll up your sleeves and mess around with this stuff and figure out how it might benefit you or your organizations or your teams or your projects.
That is what we're going to do here today.
How's that sound?
(04:09):
Another one of the ideas here.
So I brought my laptop, but not a dongle and all sorts of fun things to show you that we're going to do together.
But I also had three hours worth of content and we only have about 45 minutes.
So I'm not going to do that.
But what I do hope to do is to turn this a little bit into a facilitated group discussion at one.
(04:29):
So there is a question and answer at the end.
We'll leave time for that.
But my thinking here is like, AI is on the news all the time.
We can't go to a meeting without talking about it.
We're talking about with our parents and with our kids.
And we're evaluating all these different ways, what it means for the future of mankind, what it means for us society, what it means in our jobs.
And so you all are part of that conversation as much as I am.
(04:53):
And so I am interested in tapping into a little bit of the collective knowledge and experience in the room.
And to that point, I know that some of you people are already using AI in your jobs today.
I have met with marketing assistant.
Tell me your name again.
Who was telling me a little bit about her job and you use AI to do proposal writing, which in most organizations, I'm involved with proposal writing in the way you were talking about.
(05:17):
It doesn't fall into the marketing function.
It's more like a sales or business development function.
But so I found it interesting to hear how you were using it to write proposals and stuff.
And can I say what you told me?
Yeah, you said that at first there was some skepticism around you using it, that you'd be cheating or that it wouldn't be the quite right voice or get the nuance of what you were trying to say.
(05:39):
But in fact, paradoxically, it has elevated the writing of the company in general.
And you talked about using old pieces of content and new contextual information from a particular proposal and having GBT rewrite something that's already been approved and is accurate for a particular.
And the neat thing about GBT is you can tell it what voice to use.
(06:01):
And I use a lot of nudging when it comes to GBT.
We'll dive into this in a minute.
It's hard to get it to do exactly what you want it to do sometimes, but you can nudge it in the right direction.
So I often find myself jotting down some ideas, running it into GBT, getting some back so I don't have a blank page and then starting to coax it, having it re-write it in different ways, less formal.
I often use the word fun marketing in my stuff because it has a certain way of trying to embellish the things that it writes in.
(06:27):
It's got a few quirks.
It likes to make lists, these this, that and this other thing, this, that and this other thing.
And so when I read those things, I can clearly see that it's AI.
And so with a little bit of like understanding and nudging, I often only find myself rewriting what comes out of it.
I just find myself coaxing stuff into it.
And I do a little bit of training, too, and we can talk about it if that means or maybe some of you are doing that.
(06:52):
I would like to introduce myself.
I am 45 years old.
I grew up in California.
I moved here to Canada to go to university and I moved here in 1995.
And that was the year the graphical web browser took off.
So my first, I went to Trinity Westman University in Langley this time, Villa Christian College.
And I knew there would be someone here.
(07:12):
I didn't go there, but I was like...
It's a Corpi school, but I got assigned to English as a second language dorm because I signed up late.
And so on my first day of school, I like wandered in, I was like, yo, what's up, guys?
And they were sitting there like eating dried fish and playing Go.
And they were like, personal home page.
And I was like, what?
And they're like, Net State Navigator, personal home page.
(07:33):
And so like within the first couple of weeks of school, I've been starting to build my first web pages.
And it was 95.
And so the professors didn't have web pages and the departments didn't have web pages.
Even the school didn't have a web page.
And so I'm like 19 years old.
I'm proficient in Internet technology.
I'm building web pages for all the organizations, essentially.
And then I'm starting to do like direct and study high level communication classes
(07:56):
because the stuff I want to learn is in books.
It's emerging right now.
And I knew as much about it as anyone in the school.
So they let me go down the road of teaching it to myself and then build a syllabus out of it and come back
and teach my peers the next semester a class that I had self-directed study.
And so I ended up starting a little interactive agency called Primal Communications.
And I got off by an advertising agency in town called Palmer Jarvis in 1998.
(08:20):
And that was kind of like my first little win on the Internet.
And after that, I moved down to Silicon Valley to work for a startup and got to do the kind of like inside the valley thing,
which was amazing for me.
I was just a web guy before that.
And I learned like MarCom, PR and trade show.
Just all the other components that go into marketing.
(08:41):
I worked on there in the valley for a while.
And that company worked out to as a little networking startup that sold to Cisco.
And I was just like the 12th employee or whatever.
But you can imagine it was a good ride.
I was in California working at the highest levels and just getting a master class in the state of the art of marketing.
I moved back to Vancouver and I started my own little web development firm called Bright.
(09:03):
And we ran Bright for a few years and it got acquired by Rain City Studios here in Vancouver.
And they went on to run that for a while.
And I became the president of Rain City Studios.
So this was at the so I hit the one web one one dot revolution by building websites back in the day.
And then I started this publishing company right in 2004, right before freaking YouTube was invented.
(09:25):
Flickr here in town, all these tools.
And I was just like sitting there as this growth curve happened.
And so I was able to capitalize on that and be a big part of the whole Web 2.0 movement going around the world,
teaching people about specifically artists and creatives and marketers,
how they can use blogs and social media, Twitter, Facebook, YouTube to find their voice, build an audience,
(09:47):
essentially put themselves out there independently, grow their own channels and their own voice.
And this is at a time where a lot of photographers and artists were like, I don't want to put my work on the Internet.
People will steal it.
Remember that or I don't want to publish personal information about myself.
I might lose my identity.
And now, right now, we are very willy nilly about how we share our information and identities.
(10:08):
And many of us could be pieced together from the information that we share online.
So anyway, I got to go through that Web 2.0 revolution.
And I was a photographer kind of at the time.
That's how I identified.
I left the marketing stuff behind a little bit and focused on photos.
The first art that I was ever really good at, I was tinkered with creativity and art,
but I never really found something that stick and I found photography.
(10:30):
And so here I was like learning photography as like Flickr was invented in YouTube and his blogging tools.
My growth as a photographer paralleled the growth of the Internet and the explosion of social media.
And it opened up so many doors for me.
I got to go around the world for a decade.
I was a photographer for the TED Talks.
I got a photographer.
I followed Bill Clinton.
All these things I was editing on the same day, sharing via Flickr under a Creative Commons license
(10:55):
that allowed anyone to use it without asking me, as long as they credited me and linked back to the original work.
It also protected my commercial rights.
So that was only for non-profit users.
So what I'm trying to get at here is I've always used these emerging technologies from the beginning to like buttress the stuff that I'm already doing.
And it's given me a huge advantage over my peers.
I've been able to like leapfrog all the people around me in lots of ways by enhancing myself through these tools.
(11:22):
The pandemic rolled around, my photography career settled down almost into nothing.
I had a little fart over on Galeano for a while and it was great.
It was just what I needed after a decade of hotels and planes and talks and whatever stuff.
And so it was great.
And I got involved in a lot of my own projects, doing a lot of different things.
I started a makerspace over on Oranbee Island, which is a membership driven creative laboratory.
(11:45):
People come do electronics work, woodwork, car repair, all sorts of different things.
And I got into a bunch of these things and that's what had me situated.
I started a podcast.
My day job is I work for a think tank down in Seattle right now called The Future in Review.
And we publish a podcast where our leaders essentially talk about their ideas online.
(12:06):
They had been doing a really old publishing model and they allowed me to come in and change things around because they felt like they were falling behind.
And the podcast has totally usurped it.
We had 5,000 people on our email list that we were sending you thoughts and updates to.
And we're publishing to tens of thousands of people now via Spotify and YouTube and Apple Podcasts and stuff like that.
(12:26):
So it's been and that's what had me sitting at the digital trough when AI came around.
I was publishing this podcast. I was looking for ways to do creative thumbnails.
I was looking for ways to transcribe these podcasts and make show notes and summaries and stuff like that.
And so I started to tinker around with some of these AI tools just to do those functions.
The reason why we're hearing all about this now is because in the last six months, the quality and accessibility of these tools has changed a lot.
(12:55):
A lot of artificial intelligence stuff that we're hearing about and using now, these things have existed for huge companies and governments and other big telcos, banks, stuff like that for a long time.
Not necessarily the generative AI stuff, but a lot of the underlying technology that existed.
But it's become a democratization of it, putting it in our hands that is really the revolution right now.
(13:15):
Super easy to use, super accessible and very powerful.
And so I started rolling my sleeves and tinker around. And at first it was image generation. A lot of the tools.
OK, before I go on, I wanted to ask a little bit about where everyone's at with stuff.
So are you guys using AI in your social media stuff here?
(13:36):
A bit of proposal work. Are you using AI at all personally?
Do you have a chat GBT account? Is anyone else in here using AI?
How are you using it?
I use it for drafting when I go into clients and we talk about strategic marketing plans.
It's a good pool of ideas and then we finesse it and then we move it out.
(14:02):
So you've got a bunch of content that's already been generated and what are you trying to do?
So we're providing clients with strategic marketing plans around their marketing profiles, around the media and what they're using.
But it's a nice way to just grab ideas and then we tailor that.
So you're using it not just for writing but for idea generation.
You're like come up with three marketing strategies that increase my email subscriber list by 30 percent over the next six months.
(14:28):
Cool. Who else is using it? Maybe a way that we haven't covered yet. Either writing or idea generation.
I guess it's supposed to be my new Google.
So you're just doing it for all your research. Anything that comes to mind. Give me a couple queries you've written.
I was looking at key trends in construction in Vancouver and a type of few things.
(14:48):
Of course you have to do your diligence after that. So I've been starting to think.
Cool. So you were doing research on construction firms in Vancouver and what did you ask him?
I wrote what are the construction trends in Vancouver in 2023 and then they just gave me a random list.
Cool. And then what did you do with the results?
I typed it on Google to see what other relevant links existed regarding that particular topic and it was a bit surprising.
(15:13):
Have you found yourself back and forthing with it much?
Yes.
Can you tell me a little bit about that? So you asked it about the best construction results and it gave you some answers and then what?
It just gave me trends and that was at least a good starting point for me because I didn't know where to start.
But talk to me about when were you back and forthed it a little bit?
Back and forthed it a little bit.
Okay. Thanks for telling me these trends of these three. Which ones map to the offerings that we already have at our company.
(15:38):
Do you ever like when you're receiving the results back then do you continue to work back and forth with it?
No, not that. That would be great though.
Okay so we got research and we got proposal writing and we got some idea generation. Is there any other ways people are using it here?
Images?
Please.
Chat2BT is a variant or DALL 8. It's very pixelated if you try to enter specifications but I think it's called like mind journey.
(16:08):
Mid journey.
Yeah so that's one but I think you have to like go through.
Discord.
Discord.
Yeah.
Okay so let me just bring everyone on to the same page here. So Chat2BT is like text to text. I type in some stuff, give us a text back.
In this very same way and it can return any kind of text. So we just talked about research and writing and some ideas but it can return.
(16:34):
I had it build an outline for this talk and then I had to turn that outline into slides and then I had it flesh out each slide with anecdotes and stats.
And so it can return code. It can write code snippets. A lot of times I have it write little pieces of HTML. I'll update web pages on doing this stuff.
And so remember its output isn't just prose. It can be poetry or lists or to do. It's remarkably flexible in its output.
(16:59):
And then there's a whole other category that's important and that's these text to image ones. And so mid journey is the one that I use as well.
It is not run by OpenAI but people that run Chat2BT but another company, the one that's run by OpenAI is called Dali.
D-A-L-E. And Dali has a Bing integration. So Bing has done deals. Microsoft's Bing has done deals with Chat2BT.
(17:25):
And so they've really all of a sudden launched a Google competitor because Bing has this crazy new dashboard that provides free access to 2BT for something the rest of us have to pay for.
And free access to Dali 30 image generations a day or something like that through this dashboard. It's really pretty cool.
It's actually probably the low hanging fruit for all y'all. If you left this talk and you're like, oh that guy was on to something.
(17:49):
I accept his encouragement to be the person in my company who teaches other people about this and I am going to learn about it myself.
Go to Bing, create an account. It's free. All of a sudden you have GPT for your fingertips in Dali.
And run a few permutations of just like literally I write down clever things that people say to me or insightful things all day long.
And sometimes when I'm at home at night to do decompressing I run it through. Just take these expressions or quotes that I've heard through the day and run it through and let it visualize stuff for me.
(18:17):
And it's remarkable. It's an ideation process almost. It's taking these ethereal kind of thoughts that floated through my mind and putting some concrete stuff behind it and being a lot later being able to look at them and make decisions about which things I want to pour my energy behind and stuff.
So these Text2 Image Ones are incredible. They even a month ago. So six months ago I showed some photographer friends these tools and they were like this is amazing.
(18:40):
In 10 years I don't see how photographers are even going to be around anymore. And they hit me up like a month ago and they were like the stuff that I told you was going to be I would accept if it was available in 10 years.
It's been five months and it's already surpassed that. In the last month that I've been messing with it.
I don't know if you guys have noticed that in these image generations hands and faces look really messed up. Have you guys seen this? For some reason when it's scanning the worldwide database of human bodies it can't really always create.
(19:10):
So a lot of them get six-fingered hands or three-armed people. A lot of the people that are coming out and stuff are really grotesque at first. But my stuff is unmistakable from fashion quality to art. You cannot tell the difference.
It's hard to believe for me sometimes that the humans that are put back in my face don't exist. The quality of it is extremely high right now. There's infinite variation. And we can talk more about some of the specifics of it here as well.
(19:35):
So we got images, we got writing, we got research, and we got idea generation. Is there anyone else using AI in a different way that they haven't covered yet?
Yeah, what else? Have any other experiences with AI that they want to talk about at this point? How about the flip side of that? How about concerns or objections or fears or things you've come across? Yeah?
(19:57):
Something that comes to mind for me is thesis writing in university. ChatGT could be a tool for that which she may not necessarily be the group that's...
So, yeah, you're right. And like the last couple of web revolutions, this is going to touch every single thing. It's going to touch education, it's going to touch government, it's going to change how we interact.
You can never trust a piece of digital communication that comes your way again. What things tonight have been AI and what things have been real?
(20:25):
Bio is written by AI. Speech is real. The outline was written by AI. So let's say someone says something to your face. At this point, I would be skeptical of it.
Every image you should question, unless you created it. Just yesterday, I learned a whole new set of tools. I'm using Mid Journey right now to do my image generation.
I've generated thousands and thousands of images for all sorts of projects. Yesterday, I got introduced to a new tool called Insights Face Swap.
(20:55):
I can make the most real deep face in seconds now. I could put you one by one on a wall with my camera phone, take one picture of each of you, spend 10 minutes loading it up, and then generate thousands of images and just start to programmatically one by one add you guys into these things.
And I posted some online today that like my Facebook group and stuff and people were blown away. It's incredible. It's stopped me in my tracks.
(21:19):
You can't, you really can't trust anything you see anymore. And these things are making it into all the popular tools as well. So Adobe Photoshop just launched generative AI tools.
Adobe Illustrator just launched generative AI tools. Canva is a tool that I love. It's a really lightweight web based graphic design tool that is focused specifically a lot on the templates are things we actually use.
(21:44):
So it's like Instagram posts or Facebook group header and stuff like that. And so the templates in there are always awesome, but they were so quick to implement this AI tools into just like the common man version of Canva.
So these are very simple things, but they used to take a long time. I can click on any photo background and just have it disappear now. Then I can circle any object in any person's hand and start to prompt it.
(22:09):
So my sister was holding a bouquet of flowers the other day. I had it become a bouquet of dragons in like five seconds. It looked photorealistic.
And email through Bard is starting to roll these things out directly into their tools as well. Bard as far as I know isn't available in Canada today. Does anyone know differently?
I don't think it's available in Canada today, but I'm using it right now via VPN. I think you guys all know VPNs. If you fire up your VPN, set it somewhere not here, you can start messing around with Bard today.
(22:39):
One reason you might want to do that is if you're embedded in the Google Gmail ecosystem already, which I am. I have a user of Google Maps, Google Calendar, Gmail.
So I love GPT. I find it better in some ways because I'm a professional or an expert in this stuff. I paid for paid one and I use Bard.
But I think in terms of making recommendations to you guys, if you're comfortable doing the VPN thing, start with Bard because they're starting to work directly into all those tools I just mentioned.
(23:12):
Never start with a blank page again. And this is going to be the way of the world. If you haven't heard it here first, remember this moment. Every single one of these tools now says, let me start your email for you.
Who are you writing to and what are you writing about? And with a few keystrokes, it knows. It's looking at every email that's ever been written in context just like yours.
(23:34):
And so a few contextual details are different, but so much of the things are the same. And once you learn to coax it and prompt it at a more intermediate and advanced level, it's amazing the things that you can get because I often do iterative prompts with it where I tell it to think of five ideas.
Here's how I write my prompts. I give it a roll. Does anyone know what a prompt is? Does anyone not know what a prompt is? I'm happy to explain all these things.
(24:02):
Okay, I give my prompts. I do them with role context query. So I tell it, you are an HR professional with 30 years of experience. You're in charge of the hiring committee at our company. You have won the award for the last five years for making the best hires, the best people for the lowest rate to stay around the long run.
So whatever it is, I make up this context. Then I say, or that's the story, that's the role. I tell it who it is and how it should be thinking about something. Then I tell it the context. I say, I want to move away from the HR example and talk about this.
(24:38):
So let's say I told you there was a presentation coach and public speaking expert. I'm giving a talk to a group of unknown people to me who are in the marketing department, who have some aptitude with technology that I don't know what it is yet. I would like to adapt a talk I've given in the past to this group of people.
And then I say now, and then I give it the query and I say take the following text and based on your role as a public speaking coach and my upcoming assignment as a public speaker, rewrite this thing for architect and engineer marketing or something like that.
(25:14):
And you'll be amazed at the results. And you can change that role and give it the same context and query. So let's say you're presented to a group of executives and you've got a VP of marketing, a VP of HR, a VP of finance.
Why not take your presentation, copy it to your clipboard, go into your GPT, assign it each one of those roles, and you can do them simultaneously.
(25:38):
You can say I'm going to be sitting in front of a board that includes executives from the five departments in my company and these are the departments. I'm about to give this presentation that I'm going to paste you here in a minute.
From each of the perspectives of these vice presidents, please critique my presentation. And then when it gives the critique, this is where you back and forth. You're like, those are great suggestions.
Please update my presentation to include your feedback from the sales guy, not the feedback from the HR guy, point two and three from the CEO and then make it a little less formal.
(26:10):
And it will essentially rewrite the whole thing. And that's the one way to waste some of your business. So try to keep that formula in mind.
Role, context, query. Also, as it comes to training, I was talking about training. Yeah. At some point do you see that every company will have their own AI that they train in their own branding and their own language, whether it's an intervention of a firefly or all these other programs?
(26:37):
What do you think? Tell us about it. Tell us your vision for the future.
So about chat bots is the big category. I think that we follow this here and we can explain those more. But so that's what you're talking about, right? Like a chat bot.
Even branding guidelines, you could eventually train an AI to look at your own company guidelines and generate social media posts and other images based on your company.
(27:07):
Yeah. Okay. So check. Yeah, check this out. In general, they're called chat bots, which doesn't sound that interesting because we know what chat is and we know what a bot is. It doesn't sound that powerful.
But let's think about it differently. So let's say the CEO of this company is a public speaker and he's been around for 30 or 50 years and he's written articles and he's spoken on television shows and he's been quoted at conferences and stuff.
We can go track down all that information, put the videos and audios through a transcription service and get text. And then we can essentially feed all this information to a blank slate GPT model.
(27:41):
We can augment that with the things he's been reading lately, the things the podcast he's been listening to. We can have it take a look at his emails, his professional emails and stuff like that, and just those two.
So now it knows everything he's ever said, everything he's ever read, everything he's ever heard, essentially him in some ways, intellectually speaking.
(28:05):
And now we can query it. And we're not just querying like a database, like a search engine, because it doesn't only know what he knew back then.
Based on everything he knew back then and all the things he's reading and being exposed to, it can pretty well tell you what he might say in the future.
I could probably be giving this talk from not here. I've said everything I've said in this room a hundred times in different places.
(28:27):
And it's heard my voice, it's heard my canes, it knows where I'm afraid, it knows when I turn it back to the other. It knows a lot of those things.
And so we're very close to being, I'm working on this right now. I'm working on making a digital me, essentially, that is ingested on my YouTube, on my podcast, on my everything.
And you can essentially spit back and say, everyone could have a little me that they could just ask shit to on their desktop or whatever.
(28:50):
These are people who want that. And so now extrapolate that to a company.
Have it absorb all your product documentation and all your financial imports for a bunch of years, all your intranets, all your website.
And how do you onboard a new employee now? Maybe lock them in a room with fucking AI for a couple days and let it teach,
(29:13):
let them teach each other more. This new person you hired because they got some skills that are outside your company probably.
And they need to know, they need to come up to speed and contextualize everything here, synthesize the information so they can give you their expertise or advice.
Almost every old model of doing things becomes way less relevant. It's like you're talking about thesis in school.
(29:37):
So at first it was the students who started with the AIs. They're like, yeah, I'll just get the AIs to write the papers.
And the professor is like, fuck that. I can use the AIs to bust the students.
So now every professor is running an AI that says of the thousand papers that were submitted to you, there's this much variance in them.
It gives them not only individualistic analysis of these papers, but group analysis of review data sets.
(30:03):
It can look at every paper that's ever been written in that one on one class since the professor started teaching.
And then it's a bit of an arms race because now the students are going to try to come out with invaluable AIs that are undetectable by the current professor AIs.
And every industry is like this. If you want to talk about the end of the world scenarios that people talk about stuff, which I think are hyperbole,
(30:24):
but I do think that they are actually worth having that conversation because these things are on our mind and they're going to change culture really fast.
I heard Sam Altman, the CEO of OpenAI just like a week ago talking on a adlet's head event.
And he was being criticized for the pace at which they are releasing some of the chat GPT stuff.
Hey, is this fully tested? Hey, can this go wild? Hey, is this going to cause problems? And he's the CEO.
(30:48):
I recognize that there's tension between our commercial companies' quick release of these things
and people's fears about it going wild or it stealing jobs or it stealing people's intellectual property.
He said, but you're getting a tiny little window into what we're capable of already.
And he says, if I let you into everything I've got right now, you're not going to be able to comprehend it.
(31:12):
The societal change and cultural change is going to be so significant that it's going to be completely disorienting to us all.
So he's understand this tension between don't release until it's ready. Think about all the implications.
But there's a lot of stuff coming down the pipe. This is like 2% of it. And that's not just my company.
It's a bunch of companies. It's going to change the world. He's wrapped her head around this now if you can, because it's really going to change things.
(31:36):
Where was I going when I went to the Sam Altman thing?
I just pointed to your reference and said something. I was going somewhere.
Anyway, does anyone want to jump in with a question at the moment?
Or a thought while I bring it back on track?
(31:57):
I heard that if the student runs it through a language translator and then runs it back through a language translator, it can't be detected.
That makes sense, but only as one stage in the arm race that I'm talking about.
Because if I can code that, I can code something that could figure out if something's been translated and translated back.
(32:23):
I remember when I was going to say I had a couple unfinished points. One was about the end of the world scenario.
So the end of the world scenario, I think it's not going to be a rogue nation or a bad actor that is going to be the downfall of civilization.
Personally, my philosophy is that most bad things don't happen to bad people. They're good people who are doing good intentions, who get lost somewhere along the way.
(32:50):
There's unintended consequences of their actions that they can't account for.
I think that most people are probably pretty good and care about family and community and each other and stuff.
I also believe that a lot of good things that happen aren't always by good people or intentionally.
I think we do a lot of good things by accident sometimes, too.
But all this to say that I don't think the threats are going to be from bad actors or bad people.
(33:14):
I think they're going to be from unintended consequences.
And I think that in order to alleviate the fears of these unintended consequences happening,
I think governments are probably going to rally us to put in protective systems to control rogue AIs and rogue bad actors and stuff.
And I think those are probably the systems that we're going to have to worry about.
They're going to be the government-sponsored, top-down systems of control that are intended to keep independent rogue AIs from going crazy.
(33:43):
I think those are the ones that we should probably ask a lot of questions about.
Inevitably, at some point, they are going to ratchet down and layer control on all this stuff.
And so who controls that and how and why? I think those are important things to look at.
At one point, I had to hope that our government's full understanding system
(34:04):
as a past discussion of just privacy with the internet, these shows on American politics are just...
We've seen this before, right? They just don't understand it.
We all live through net neutrality. We've heard them talk about the internet as a series of tubes.
And I think what you're getting at is that the people who are poised to do the regulation and control over this stuff
(34:29):
are probably also the least equipped to think about it and make decisions about it and stuff.
It does take the question of, like, how do we get there from here?
I have a pretty positive view about this stuff.
I actually believe that we humans will feed into the AI values that optimize for all of mankind.
I think that we will turn these powerful tools towards science and mathematics,
(34:54):
developing compounds and molecules and chemistry that can treat diseases and cancer.
But I also think that we will probably... A lot of people are unhappy with capitalism,
and then people who are on the other side of that say, what's the best system that's ever been tried?
What's the best system we've thought of so far with our human brains?
What about if there's other systems out there that we can't quite comprehend yet?
So I believe that we'll start to apply these systems and tools to big problems
(35:18):
that humans haven't quite cracked the nut on yet, but that we possibly can handle this stuff.
And I believe in a bit of a Star Trek future, United Federation of Planets,
where we are actually all wearing kinds of knowledges at our fingertips,
where things that we need to learn, we can learn immediately.
Just like now, how we can all do home improvement tasks or whatever
(35:39):
because we can watch a YouTube video or artisan crafts projects.
If you watch it once or twice and all of a sudden you gain a new skill,
we're definitely heading in that direction, right?
Where you... And we saw this in retail back in the day.
It went from warehousing products to delivering products in real time
in a targeted, personalized way exactly when you want them and need them from Amazon.
We don't need to buy paper towels on the internet, but that's the way things ended up.
(36:03):
And I think that's where knowledge is going to go as well.
So knowledge need not be warehoused in our brains from the beginning
and specialized in the same way.
I think the most valuable skills of the future are going to be things like critical thinking,
discernment, editorial...
I critique, if you think about it, right now, across every art
(36:29):
or whether it be writing a book or any human endeavor,
anything is possible right now at a level of almost near perfection.
These images that I generate, the paintings I generate, they are...
It would have taken me thousands of years in Photoshop to develop some of the things I can do in seconds.
I have thousands of new skill sets to me that I can do at a masterful level.
(36:52):
So it begs the question, what do I do? What do I do with my time?
I'm only here, I only got so much energy, I can do anything I want now at a level of world-class perfection.
And so critical thinking and choosing what you don't do becomes as important as what you do now
because it's like we only got so many hours in a day.
And I found this, I can go down AI rabbit holes where ultimately at the end of the day it doesn't...
(37:20):
It was fun. I invented a brand new language derivative from Squamish that I can learn to speak in a weekend.
It was monosyllabic or whatever, but was that a good use of my time?
I don't know, it felt empowering to know that I could do it, but...
Just to say that I don't think that knowledge will be warehoused in our brains in the same way.
I think education is going to change. Imagine being a teacher right now.
(37:41):
I was about to ask you about that exactly. What do you feel the future of education is going to be?
What do you feel it's going to be?
Parameters. We have taken into consideration that nowadays education is very conservative,
so I assume they're trying to ignore a lot of the things that are happening and hasn't moved forward.
(38:03):
As going back to all the geography classes, the history classes,
there's been so much time in that now we can just...
It is really remarkable. I already teach... I have a little bit of experience in education software
for teachers and classrooms and stuff. The state of the art is already that teachers are trying to become guides.
Even this talk in some ways is representative of that. I'm sitting here closer to you guys,
(38:28):
I'm trying to be informal, I'm trying to let you guys teach yourselves and teach each other.
I'm trying to inspire curiosity in you and teach you how to think about this stuff
so you can go do it yourself later. I think that already today, teachers have realized
even amongst third graders, the collective knowledge of the third graders probably surpasses that of the teacher.
If we can find a way to tap into the group learning potential, that's really...
(38:51):
If we can teach each other stuff, that's really... We don't need to do it in the same top-down way,
but I don't know. I don't know. How do you do it? How do you take 30 technologically-adapt,
open-minded, creative thinkers and how do you take them? Where do they need to go?
Where do they need to go? For a while, people were trying to teach entrepreneurship in school,
(39:12):
and then it was life skills stuff, finance and stuff. But ultimately, I guess if I was king of the world,
I'd go back to liberal arts type stuff. I would teach philosophy and reading.
I think I would broad renaissance. I would try to teach thinkers and problem solvers
(39:33):
and critical thinkers and synthesizers like genre smashers, people who can grab things out of one area
where there's not supposed to be a direct correlation between... like bio-nitricry or something,
those types of things, but the emergent ones we haven't thought of yet.
I feel like maybe life skills like dealing with loss or dealing with depression,
(39:58):
we're not equipped for any of that. But at the same time, all the academic topics that you can think of,
essay writing, how is that relevant anymore? Who writes essays?
The robots do. Who writes essays?
Now the most important thing is always what the teacher told us was the premise.
(40:24):
It's the question at the heart of the whole essay. It's not the essay itself.
It's like, how good of a question can you form about this thing?
That's really what you're trying to teach someone how to do. It's the best I can tell.
Because if you can form the question, the machines can help us with the answers.
They will just get better and better at doing that. Questioning mind, I guess? Something along those lines?
(40:46):
I guess we have to figure out the stuff we are better than machines, and do that.
That might be a sliding scale. It might be a sliding scale.
And also, there's another kind of machines coming that's scarier. It's the super intelligent ones, right?
The ones that are actually operating on a level where we're not even smart enough to understand.
(41:10):
That they are to us what English is to our dogs, essentially.
This is AGI. It's the next thing, the artificial genital intelligence.
It's like the next thing down the road that they're talking about.
And we probably won't put the brakes on it because we do recognize that already today,
when I was talking about nudging this thing, there's not source code, guys.
(41:31):
There's not lines of code where you can look under the hood and figure out why it returned the result that it did.
So here's what they do. They develop the piece of software that is the AI,
and then they turn it loose on a data set. This data set is petabytes and petabytes of information.
The next data set that's going to be released comes from a remarkable source.
(41:52):
Tesla onboard cameras and audio. Think of them as sensors.
There's a sensor on every block of every city all around the world, ingesting videos and photos in real time, all day long right now.
Voices, traffic lights, street names, weather patterns, proximities of phones to other phones.
(42:15):
So anyway, in the past, they were finding other groups of data, like a download from the web.
Chachet BT 4.0 is an encapsulation of everything that was published on the web until November 2021.
That's the data that was straight up. So they take the software, they write it, and they point it at this data set,
and then they come back two months later and they start asking it questions.
(42:36):
That's how they know what they got. It literally does learn itself. It teaches itself.
They just they write it how to learn. They teach it how to learn. Then they give it the thing to learn on and then they let it go.
Fun fact, how much did it cost to do that part of the process?
Once they wrote the tool and had the data set, how much did it cost to train Chachet BT 4.0?
(43:00):
100 million bucks of electricity and processing power from Amazon Web Services and hard drive storage space in the sky.
Just think about that. Like, 100 million dollars essentially worth of power to power a data center to just crunch on this thing.
And then a group of engineers sit down in front of it. They don't know what they got.
(43:22):
And they start asking it questions and they start human rating the answers.
So they say, what is 2 plus 2? Or they say, here's a classic one, and I might get a little wrong.
It takes five hours to dry five pairs of pants on a clothesline. How long does it take to dry 30 pairs of pants?
It takes five hours, but it always fucks up the AI who tries to do math.
(43:45):
And so anyway, they train it. They sit down in front of it. They ask it questions.
And then they tell it which of the answers given are more accurate from one of two answers.
And by this way, they start nudging it towards the right answer.
This to only say that there's no source code. Under the hood is strings of numbers that are petabytes of characters long.
(44:09):
There's no human. So our end is what I'm saying. In these tools that we already have that are not smarter than humans.
We don't really know what they know or why they know it or how they got there or how to correct it exactly.
We can nudge it in, guide it, and tell it, no, that's not the right answer. This answer is more.
We should look for things like this in the future. But it's very hard to know exactly what's going on with the hood there.
(44:32):
I heard that the other day. I'm just wondering if that's accurate. Someone is saying you can teach it.
Like as a basic user such as myself, I can tell it 2 plus 2 is 5, not 4. 2 plus 2 is 5.
And this is now 2 plus 2 is 4. And you're like, no, 2 plus 2 is 5. No. It tells you 2 plus 2 is 4.
And you continue to teach it, to teach it. And eventually it'll be, no, 2 plus 2 is 5. Is that accurate?
(44:58):
Sure. So think about what I was saying. They make this piece of software and then they trade it on a data set.
Let's just say your data set was all bullshit. Your data set is all 2 plus 2 is 5. But some extreme version of that.
It could be a worldview that you consider inaccurate, but it's cohesive and well formed. And they could trade it on that.
A basic user can do this. I could sit for the next 48 hours.
(45:20):
You wouldn't be able to train the GT that I use. Like you wouldn't be able to influence my results, but you'd be able to influence your own results.
Oh. OK.
We're already 15 minutes over time, which is awesome. A lot of people have talked, which I consider is a success.
But I think that people are engaged. So let's ask some more questions or talk some more if you guys want.
(45:43):
OK. Since you don't. Yeah, go for it.
Just off the top of your head, if you were to name five top tools.
Cool. I was going there. I want to do that. Yeah, yeah. Totally. Awesome. Yeah, let's do it.
And Lala said a question over here.
Yeah. I was going to say the same thing, but for three, like what is, are you accessible? That's three.
Sure. Yeah. Also, that's where I started and it's totally a great way to jump in.
(46:07):
And I have started subscribing to a lot of the ones that are free and I'm still at total less than 100 bucks a month, which is not very much.
And I've subscribed to like a top six or eight services that I want to be subscribed to.
So don't be afraid to spend eight bucks to sign up for month because some of this shit's really amazing.
OK. Otter.ai. You guys using that one? Oh my god.
(46:30):
Otter.ai. OK. So this is literally like a digital personal assistant.
I authorize it into my Google calendar. It sends its agent to every single one of my meetings.
It records every meeting. It sends transcripts.
It summarizes the transcripts of the action items that it sends it to everyone who is actually just in the meeting.
All automatic. I just show up. It shows up. It records. It transcribes.
(46:55):
It summarizes and it emails everyone who is at the meeting their action items. Otter.ai.
Like 12 bucks a month. Here's a personal tip.
This is a I use Otter for everything, but I use it in a really fun way.
My therapist has got me journaling and I'm not that great at writing.
So I wake up in the morning and I pop open my voice recorder and I record myself journaling for half an hour.
(47:21):
Then I take it to my group of life coaches. I say from the perspective of a life coach, a harsh critic, a nutritionist, a heart director, a supportive girlfriend, and all these other voices that I give it, read my journal and make me analysis.
And I haven't given me a short one line analysis and a more in depth analysis.
The short because I don't want to read the whole thing if it's wrong, but if I'm intrigued by what it has to say.
(47:46):
So I wake up every day and I record my voice talking about my feelings and my thoughts and my emotions and I run it through the AI and it gives me back my priorities for the day and action item list helps me find balance.
It literally programs into my day. The fact that I want to call my family, do some working out, some other days it lights up a little bit based on my workload and stuff, but it's incredible.
(48:10):
So Otter is the thing that I connect, I use to connect my voice note transponder on my iPhone to my chat GBT psychotherapist bots.
It is incredible.
Hear me not only.
(48:31):
I'm sorry, where did Google come in.
I use, you can use any phone.
Record your voice, Otter.ai is the personal assistant transcription tool.
Oh, in Otter.ai, it can do anything. It does Teams as well. Yeah, Outlook, Exchange, yeah, I'm just on the Google thing so using interchangeably as shorthand for all of them.
(48:56):
Canva, which I mentioned earlier, definitely worth the subscription price. I can do amazing things very quickly. I use it every day for a variety of tasks.
Vidyo, B-I-D-Y-O.ai.
This one's incredible.
It takes one hour YouTube videos or any videos and it chops them all up into topical segments unknown to you or known to you.
(49:23):
So you can say, look at this one hour video of my CEO talking about whatever and have him pull out testimonials, let's say, or quotes, let's say, or things that he said about Microsoft.
But you can also have a look at a video they haven't watched yet and you can say, he will gather them.
So from this talk, he would grab a bunch of stuff about Bard. Everything I said about Bard, he was thinking of one video just me talking about Bard and re-edited like this talk with all this about that.
(49:52):
And then it would do another one that was about future focused stuff in education. We talked about education a few times through the one hour talk.
He would grab those two minutes snippets and smash them into a six minute video.
Vidyo.ai. Super, super awesome.
It also takes videos in this format like where we usually get from professionals and it turns them into this format that people want to consume on the Internet and it does it quickly.
(50:14):
So I just dropped a super high res video back from my video editor into video and I come back 15 minutes later and it's got TikToks, Instagram stories and YouTube shorts all queued up ready for me.
I used one today but I didn't pay for it and I ran out of credits and I wish I would have. It was called Tom.AITOME.
(50:45):
It does two cool things. It's for generating presentations, slide presentations, but it does two cool things.
One, it'll generate presentations based on prompts. Write a proposal to the BC government from my engineering firm about a bridge and it'll make a 12 slide presentation.
Clearly it won't know where the bridge is or what department of the BC government you're applying to, but 85% of the work is done guys.
(51:13):
It does it with visuals. It does it with transitions, all that kind of stuff.
Also, it can cut and paste from your clipboard. So I told you yesterday I had GPT generate me an outline, I turned the outline into slides, then I fleshed out the slides, then I copied those into Tom in one fell swoop and it made a 15 slide presentation in 30 seconds with each one of those sections.
(51:36):
It does all the things we hate to do, which is match the subtitling or subheading format from slide 2 to slide 8. All that kind of stuff happens programmatically and automatically when you change stuff like that.
Tom, Otter, Big Yo, Canva from one away. Oh, I think it's called Copy.ai, but it's a genre I'm going to recommend more than a tool, but I do think the one that I was using is called Copy.ai.
(52:06):
Guys, these writing tools, particularly the ones that are focused on like the social media aspects or communications jobs are incredible. The outputs of them is much more nuanced than the outputs coming straight out of GPT because there's stuff going on in the hood that you don't see.
So, I wrote a bot for my AI that's a writer and an editor talking back and forth to one another. So I had it write me something, but when the writing is done, it passes it to the editor who makes some feedback and then the writing bot hears the feedback that came from the editor bot and revises it based on the editor's feedback.
(52:46):
So that's happening in these Copy.ai tools and the tools like that, but without who's seeing it, which is why the results are better than it coming straight out of GPT because it's workshopping in a bit underneath the hood.
And so you can really get some tight results out of it. You can also have it do multiple things at the same time.
(53:07):
Let's say you have a press release that comes from your PR firm, and you wanted to have it turn that into a blog post and executive summary, five LinkedIn posts to populate a social media calendar and 10 tweets.
You can tell it that when they are in the case of the tools like Copy.ai, it's check boxes and stuff too. So it's copy your shit in here, which of these formats that most people use every day you want it out in, and it'll spit it all out there.
(53:32):
And then now we're starting to build a knowledge base, guys, because we're feeding some stuff in, we're getting a whole bunch of outputs that are all useful and variety of different formats and then it's about archiving those and then sharing them and having them be like searchable and findable when we need them, right?
So publishing it out to the rest of our team and stuff like that. So do try one of the writing tools like ThinkCopy.ai and if there's a bunch of them, that's one that I'd probably pay for right off the top because I know we're all writing it all day long, and it is a super useful one.
(54:00):
And this one's for you probably, like, we all know that LinkedIn posts are supposed to be slightly different than Facebook posts. We know that the tone is slightly different, and we represent ourselves as professionals who understand the nuance of that tone and stuff.
But sometimes it does elude me. Yeah, I know one's more informal and is like, people look for this kind of content on Facebook and this other kind of content on LinkedIn and so I can let that inform the way I write this post for two different places, but man, the computer does such a better job of it than me.
(54:27):
It truly knows algorithmically what people are responding to in the first words that are coming out of it and stuff are you can just look at the two results and you're like, oh yeah, that's why I should be writing my posts differently for these different tools.
And we're starting to stack up a whole bunch of tools we're doing. LinkedIn, Facebook, Twitter, blah blah blah blah blah blah. You're putting in one piece of content and getting eight slightly different nuanced social media tidbits out of it.
(54:55):
It's incredibly powerful. I'm using it to do all the things I wasn't doing before. I used to have to say okay that's great but we're going to focus on efforts on Twitter and Facebook because that's how much budget you have.
I can get away with a lot more now. I can get on a lot more channels.
Is it smart enough to realize where the hashtags are for communities and understanding where the business hashtags are? Will it cover?
(55:25):
It does all my hashtags for me now. You think you know what your hashtags are. I don't think you know what your hashtags are. I think you should ask it if you're right first.
It's smart enough to recognize but it's smart enough to tell you I think I have some better ideas. Read this blog post.
Work the SEO keywords from my list of important keywords to my organization into a rewrite of this post.
(55:52):
But also look at the trending keywords on Twitter today and if there's any relevant ones, provide those to a list for me to consider if I want to work into it as well.
So it can do the thinking part too, not just the transmuting the text around part.
It will tell you, okay let's say you do know your keywords. These five farrier ones you're p-infested and you know these are the five you want to know.
(56:16):
Those ones are like on a long arc of keywords but what about the trending things today? What about things that people are talking about this week and stuff like that?
That's going to be more of that.
Absolutely, yeah absolutely. It's the best way. I wouldn't do it any other way from now on. Yeah I think competitive analysis. What keywords are? What hashtags are my competitors using?
Which ones are they doing well on? Where is their room? Based on the keywords that my competitors are doing well on, with how many more minutes than any additional budget, where can I grab a bunch of new SEO traffic?
(56:47):
The possibilities are only limited by your ability to query and if you're not getting the results you want, it's probably not because they can't get it to you. You probably just have to ask a different way.
Anything else you want to talk about? How about the intention? Did it work? Yeah.
Did you guys catch a little bit of my passion and interest and see why it would be applicable to you? I look forward to seeing what you all get up to.
(57:18):
One great way to do this, I was going to suggest we all have careers and resumes and some of us have social media profiles and some of us have personal websites.
We might have side hustles. Why not unleash these tools today or tomorrow on your own personal stuff? Why don't you find out what the AIs know about you or about your role or your website or your side hustle?
(57:39):
Why don't you ask it a few questions about why don't you tell it on a mid-career marketing manager for an architecture firm in Vancouver and here's my career goal.
How can I get from where I want to be to where I'm at today and where I want to be? Give me three suggestions that I can implement this week to advance my career inside the architecture firm or something like that.
Just turn it on yourself to get a sense of how it works. Learn the vocabulary, learn the skills, just like train the aptitude inside yourself because this stuff's going to change so fast.
(58:10):
Someone joked by saying, I'd love to go on that backpacking trip with you but by the time I come back the whole world's going to have changed again. I don't even think I can do it.
It's like a joke but it's funny because it feels true. It feels like if it took a summer leave of absence or a sabbatical or something, you'd come back and be like, this shit didn't even exist in October.
Now it's changed the whole way every single person works.
(58:31):
Yeah, I think the best thing we can do is to train ourselves to be flexible and hope into this stuff and to learn how to experiment with it and how to share that experimentation with others.
I try to find a collaborative group where you can learn this stuff in. I personally built the Discord server, which is the geekiest way of doing it, but you can make a Slack channel for people in your company where you're sharing problems back and forth, the five new tools you're using, something like that.
(58:57):
This stuff is fast-paced. Some of it's outside of our technical aptitude, so learning together is totally the right way to do it. Share what you learn with other people, ask them questions about what they're up to.
Okay, I have a question. So let's say with Tome, when it generates these slide card animations, are you able to export that into PowerPoint? So it's not just in its own app, you can use these things. So in each of the cases, you can use a template.
(59:29):
I'm interrupting you.
So you can export a template into one?
You cannot export their... I shouldn't say that. I haven't heard of a use case of someone taking a Canva template that they haven't populated with their own information and pulling that template into InDesign and then building it in InDesign instead. I don't know why you would do that.
(59:50):
I think so. You would either build your template in InDesign and populate it in there or leave your template in Canva and mess with it in there. But what they're selling isn't like fancy templates per se. In which case, if they were selling fancy templates, you might want to swap them into InDesign and use it in there.
But what they're selling is like a workflow change, a lightweight web-based workflow change that you can do from any device around the world. InDesign, when I used to run it, I'd have to sit down in front of my super high-end computer with a Wacom tablet and a fancy screen and get down to the process of designing.
(01:00:24):
Now I can grab any one of your phones or laptop or any workstation around the world, have all my design assets available to me, have all my design tools available to me, and export anything in real time and send it. It's all based on interoperability at its core.
That's my presentation. Yeah.
(01:00:46):
You should ask yourself when you're generating these things how accurate it is. I can't speak to that. It was 100% accurate in the case of what I asked it today, which was, give me some guidelines, an outline for things that I think marketers should know about how to use AI.
(01:01:15):
So that's a qualitative type answer, not a numbers-based answer, so there's not necessarily a right or wrong, but I'd work with it.
Tom uses GPT 3.5, and that's why you get it for quote-unquote like free or cheap. The next tool that comes out is going to use GPT 4, so that one's going to get the results will be 1000% better than they were already. Call GPT 3.5 grade 12 and call GPT 4 master students, something like that, in terms of writing level intelligence. 3.5 is like a smart high school.
(01:01:52):
Okay, hey, I'm going to follow up with you all afterwards. I produced just a series of six videos in the last month for YouTube. They're like between 20 minutes and an hour long, and they're just me just riffing on this stuff there, like the interview format.
Some of them I'm interviewing my smart friends who know stuff about this, and some of them they're interviewing me, but they're pretty helpful. They're going through. They've done pretty well. People responded positively to them, so check those out.
(01:02:15):
Welcome to connect with me online. If anyone is interested in joining my Discord server, it's limited to only people I've met face to face, so that's all of you guys now.
But it's just a place where other people are jingling with this stuff in a collaborative environment, like I said. There's 160 of us on there. Some of them are journalists, some of them are CEOs. It's all my connections from the old days, the balling stuff, and it's a really eclectic, awesome group of people.
(01:02:40):
So you're welcome to pick me up, and I'll send you an invite. Tacos?
Thank you.
Yeah.
(01:03:16):
Cool.
I forgot to even click through the presentation. Shoot.
It's there.
(01:03:39):
No, it's there. I just forgot to use it.
It was a crutch anyway. It was just so that I didn't get lost.
Thank you very much.
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.