June 2, 2025 • 27 mins
e516 with Michael and Michael - #AI #prompts, #browsers, #ModelCollapse & #automation along with a teeny tiny #PicoMacNano and so much more.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
[Music]
(00:10):
This is GamesItWork.biz, your weekly podcast about gaming, technology, and play.
Your hosts are Michael Martine, Andy Piper, and Michael Rowe.
The thoughts and opinions on this podcast are those of the hosts and guests alone,
and are not the opinions of any organization which they have been or may be affiliated with.
(00:32):
This is episode 516.
[upbeat music]
Good morning, good evening. Good afternoon, Michael Rowe here. It's Friday. I get to talk tech with my second favorite Michael and my co-host and friend Michael Martin because you know, you got to be your own first favorite Michael, right? Oh, man. How are you doing?
(00:59):
You're my first favorite, Michael Michael.
Hang it in there and we're missing Andy this week, but that's again why we have three co-hosts.
We can carry on forge forward, craft our way into the future and you see what I've done there.
(01:19):
And we've got a great show as we always do curated references, listener links and all kinds of good fun stuff.
So if you've got a story, you want us to...
the touch on, you do pass it our way.
So all kinds of news this week around models, models, models, all over the place.
Yes
(01:40):
And the starting point we have here is an R's technica article about AI prompts,
about how, at least for this article says anthropic is controlling Claude 4.
And for those of us that have been playing in the space.
For a while they're usually very comprehensive sets of instructions that are typically all prompts that various organizations will use to say when a user asks a question like what is the square root of pie or tell me something interesting about the particular book.
(02:23):
That gets parsed by the prompts that are behind the scenes.
And this particular article is talking about the level of complexity that's there that is not usually visible because you ask a question like tell me everything about this new movie or this book or whatever it might be.
(02:46):
And then you get an output and maybe you get a chain of reasoning and maybe get some explanation and some further transparency, but you don't typically see how does the interaction happen between your question.
I like this article because it gets into a little bit more detail and we've had some examples in previous shows about how you can unpack what is in there.
(03:13):
So the example here that we've talked about on gosh a couple of occasions are how the responses from some of the models in the interactions are telling you how wonderful you are for considering such a thought.
the sycophant models. Yes. What I really enjoyed about this article was really the fact that for people who are not deep in the space or haven't spent the last couple years like you and I actually building prompts and doing work, whether it's a hackathon, day job, whatever,
A wonderful question right the relentlessly positive tone right.
(03:55):
And they download
the tool that allows you to locally install an LLM and then they wonder why it doesn't do very well.
So you get the model and you know, you can get a hugging face, you can get llama, you can get what a deep seek.
All these different models locally install them and then not get the results you want.
(04:19):
Well, that's because the model is just the way to stay.
And statistically present the next word.
The important part is in the prompt that explains, you know, how to behave.
What are the guardrails that you have in place?
What's the, as you say, the chain of reasoning that you should take place?
(04:43):
And I've been doing some interesting models in my day job lately around how can I generate my weekly.
This is everything I've done in report.
hmm
And it's working really well.
Oh, is it now?
Yes, yes.
(05:04):
It's pretty interesting.
When you have a entire workplace that is instrumented and sniffed, consumed, gobbled up by the model, you can do some really, really cool things with it.
(05:27):
And so I, I usually, before I shut off on the end of day on Friday, I will run my new agent that I created and it does 90% of the work that I need.
To be able to capture everything I need for, you know, quarterly reviews status updates, blah, blah, blah, and I think that's really cool.
(05:50):
But again, underneath the actual model it's probably using is probably one of the big ones that you know.
This is the magic sauce.
(06:11):
Right, and there once was a time when the concept of a prompt engineer was a real job.
And that was a bit of a flash in the pan, wasn't it, Michael, where, yeah, for a period of maybe about four or five, six months is like, wow, that, you know, you really need people who could do all that, right?
Oh Yeah
Well, as a general as a general position. Yeah, but but for the people who sell
(06:34):
solutions They need somebody to structure that unless you're building your own model
You need somebody who understands prompt engineering at a level to build that
approach to how to answer the question.
And I think it's a specialized skill for a very small group of people now, and that's the interesting part.
(06:56):
Yeah, it is not it's right. It's not a broad skill that you would expect you know a consumer package goods company would hire prompt engineers, right? The consumer package good company should buy us a
Right why would they people who are really good at consumer package goods
What right and if they're using our official intelligence to power their workloads or to design agents or or or that that level of prompt engineering is
(07:22):
is actually almost in the operating system.
Yeah. Which, or at least in the browser, yeah.
Which kind of takes us to the next article.
Really nicely. You saw where I was going to go.
And this article about Opera's new AI browser.
I'm sure you're probably going to go somewhere very similar with this than there where I was, which is.
(07:43):
Yeah, lovely. It's a browser and has an interface into a large language model.
Hooray.
Yeah. (laughs)
It's the user experience.
it happens to be a browser. Why isn't it an app? Why isn't it a hard way?
Why isn't it a plug-in into a browser? It's a differentiator in the sense that, "Oh, this browser is an AI browser. Nifty needo. Is that really that special?"
(08:11):
Well, yeah, it's interesting because, um, and we don't have a link for it for the show.
I had gotten an article, not an article, but an email from, uh, "Have you ever played with the ARC browser?"
So I played with that one a year, a year and a half ago, whenever it first came out for a few months, and it was yet another new take on browsers.
(08:32):
Um, built off of, I think it was built off the Chromium engine, it might have been built off of the, the, um, oh god, what's the Safari engine, the open source part under.
Wept it.
WebKit.
Thank you.
Is it WebKit?
I don't think, no, no, WebKit's the API stuff, uh, there's a, anyway, it doesn't really matter.
I, maybe, maybe may not, I don't know.
One of the, one of the models, um, and, uh, so the ARC browser has decided that they've,
(08:58):
they're scrapping what they were doing with the browser and coming up with their own AI thing, just like this, just like Opera.
Um, and yeah, it's, as you say, interesting.
Yeah.
But the real question is, is this a hype train that they're getting on or are they taking a browser experience and extending it into something new and exciting, right?
(09:23):
Which is like an app right or development environment, you know
Um, right, right.
And, well, or even better, you know, it could be like sky app, which is the next one we want to talk about, um, which, uh, I, I actually signed up for the, the wait list on this one because this looks interesting.
It's not a browser. It's by the guy's or the team I should say that created workflows which became shortcuts when they were purchased by Apple and this does something really interesting because it does download those LLMs of choice and then instrument your Mac. Getting into all the data you tell it can get into and and...
(10:11):
presents you with dynamic creation of... we'll call them agents to do things on your Mac.
Kind of like what Apple showed last year at WWDC, but never delivered.
And that's the interesting thing here.
There's an interesting long-form article by Federici...
(10:35):
Yeah, Vitici. Federico Vitici, I can never pronounce his first name,
together.
And going through his couple of weeks of experimentation with it, as it came out,
to limited alpha users and beta users.
And I was listening to him on a podcast this week, and he was saying,
(10:56):
"He's wondering who will release those features first?
Will Apple break all of this with whatever comes out in the next week and a half?
Or will they be bought by Apple in order to make it so that they actually
promised last year, or will they deliver the functionality that Apple promised,
(11:19):
you know, months, if not a year ahead of Apple delivering it whenever they finally deliver it?"
This looks cool.
And my question becomes, does it truly stay all on your Mac?
If it does, I'd be willing to do it.
If it's sending stuff to the cloud, this is into your personal data.
this is into your email, into your calendar.
(11:41):
into your to do is into all the documents on your machine, kind of like co-pilot, but with scripting capabilities to kind of generate agents to do things.
Yeah, in the olden days, we called these sort of things intelligent workflows.
And part of the promise of shortcuts were to connect the dots, if you will, between certain things that you do on a regular basis, that will make your life easier.
(12:09):
And Automator was a predecessor in a lot of ways to the shortcuts. And, you know,
just being a little vulnerable on a particular front, I tried to create an automator,
And I didn't try to do it via shortcuts.
I tried to create an automation that would allow me to launch a Teams meeting, copy the URL for that launched Teams meeting, and put it in my clipboard so that I could just pop it into a Slack message for somebody to say, "Hey, come and join me on this Teams meeting, I just launched."
(12:41):
And unfortunately, it's probably my mad coding skills weren't quite sufficient to get all the way to the end.
You didn't vibe correctly.
Well, but that's what this is trying to do.
OK, look, there's some things that would be a nice little speed up, you know, to let you do something faster.
And before everyone writes in, yes, I do know about the Slack automation to launch a team's meeting.
(13:05):
I use that all the time and it's really, really good.
But I was trying to go this route, you know, to learn a little bit more, try it out and see where we go.
And the notion of being very granular about what elements of data on your device on your Mac.
that you're going to allow your agent to see and can you contain it in a way that it doesn't.
(13:34):
Sift it, train it, use it, get it out into the wide world.
It's going to be really tricky because even if you do read all the terms and conditions and hold that thought because we're going to get there soon.
Well, well, that's that's that's the interesting thing here because remember last year at WDC WDC one of the things that Apple really really pushed in order to enable this
You may not be in a position where you say, well, I trust it or could those terms and conditions change.
(14:06):
coming soon ecosystem was for apps.
And they've been pushing it for about three years, but they really, really went hard last year, and those intents are not consumable by anybody but Apple or Apple tools like shortcuts.
(14:35):
And so if these guys, the guys who wrote shortcuts have figured out how to
intercept the intents in order to allow them to build these dynamic agents based off of natural language, that is taking us down the path, and that could possibly solve your challenge, too. Yeah. Yes.
Oh, excellent point.
Well, yeah, I mean, I found another work around that works just fine for me.
(14:57):
And it actually is in the Slack environment, which is even better, right?
But, but there, there are going to be situations where, you know, like our post show production, right?
Wouldn't it be awesome that after our show completes and of the recordings are available, they can automatically drop into our shared folder where everybody has
I can get close, but not all the way, and that's because what triggers an event,
(15:19):
Mm-hmm. Yeah.
what's registered, what's recognized, there are going to be a ton of those sort of things.
Now, let, let, let's continue the, the theory here on models and, and how models behave one way or another.
Very cool Oh, yeah
And there is a, another version of the general, hey, models are going down a path where they're going to be less and less beneficial.
(15:44):
that we've been talking about for years.
Yeah, so all the ideas, synthetic data, recursive model development,
GAN models trying to learn from one another, showing up again here in the register.
And while, while search is now really being a consumer, and I think also potentially a
(16:07):
feedback loop into models as well, all you need to do is get some searches that are not quite exactly where they need to be.
Don't have a human intervention around any sort of supervised learning and you're right, you're going to drift into, you know, hey, Michael, you're your new novel, you know, that plot summary for it's pretty cool, you know, so, you know, thanks, thanks for, thanks for writing that new novel, you didn't know you even wrote yet, right.
(16:35):
Yeah, yeah. Anyway, so I just thought this article was interesting of kind of restating what we've been talking about for a couple of years now. And it came across my my mastodon feed from Mike Elgin, who does a whole thing on intelligent machines. But how about something fun? How about something tiny and fun? I want a new Mac, a nice little... Yeah, yeah.
(16:59):
really? And you want to go, you want to go to a new Mac before we talk about Louise?
Yeah. Okay. Let's, let's, all right.
Yeah, because this will be just a quick hit, kind of a pallet cleanser. So another thing that came across the feed this week that I just thought was hilarious was the... What do they call this thing?
(17:19):
I don't know what they call it, but it's... I've actually played with a image of this software,
A Pico Macnano is what it is, a Pico Macnano.
right? You can download this software. That's it. Thank you. The... the...
software images is a Mac operating system from 1990 maybe. And there's a... there's a... a company that is actually printing little tiny Macs, running a Raspberry Pi Pico Zero, and have installed Mac OS running on that in the little tiny form factor that I just...
(18:05):
think is cute. Yes. Yeah. And... and I was thinking... I... I'm somewhere I still have a couple of old three and a half inch floppies of software that I had on the Mac back then, including I think the Star Trek game. If you remember that, there was a network Star Trek game. And I was wondering if it's got a U.S.
It comes in a pea caso Macintosh shipping box. That's funny
(18:27):
Hmm. Yeah. Yeah. We had a way to version of that on
Yeah, yeah, I was wondering if they had a U.S.B.A. plug on it to where I can plug a floppy drive in and read it off and load it up on the SD card on the Pico Mac Nano.
A USB-A plug on this device would be 25% of the size of the device.
(18:47):
Pico Mac Nano. That's a mouthful.
Yeah, I know. It's huge.
The plug, not the Mac.
Yeah, now it's it's cute and I'm sure Andy when he saw this too was smiling quietly to himself because you know This is very much one of those great, you know hack kind of things of
(19:14):
Yeah, running on a three-volt CR2 battery. I mean ridiculous. So very very very
Yeah, yeah, so we do have a listener link and and and I had to interject before before we got to it from from Lewis
Suarez on the internet of consent
(19:34):
What did you think of this article? I thought it was spot-on
I don't agree to anything in the...
So so at least long-time listener friend of ours
from the wow doggier nation days and and that
Linked out on massed on he's starting to finally get into massed on to an article by an ill dash
(20:02):
Called the internet a consent where he puts together kind of a
basic flow of
The quote good old days on the internet to now and how do we reclaim
consent and the reason I think this article is so
important how many terms of services updates have you gotten in the last two days okay I've got like four or five services who have all sent terms of service updates in the last two days
(20:35):
so something either regulatory in the US or something else has changed that they all feel a need to send out turn and these are media ones for like paramount paramount plus
app ones website ones etc so something has changed and what this article goes through and you know we all remember this is the continual and gradual shift that many of your favorite internet service providers from years ago went through very quickly that completely changed the relationship of the data and the consent to the data between the user and the platform
(21:16):
Well, I know you actually read those documents very carefully, right?
And that's something you've shared on the podcast here on many occasions.
And it's a right, why rise and right thing to do?
Because when you're confronted with, hey, we want to collect your data.
Well, what do you want to collect it for?
And in some cases, it's really clear.
(21:36):
It's like, well, if we don't know your name, it's really hard for us to properly greet you and provide you with the service that you're agreeing to use.
You know your data is your data and if you're not knowing how you're paying for a service and it's free you're paying for it with your data, the thing that I've seen most frequently here in the last couple of days is all of the, and I'm going to call them tricks because that's how I've kind of interpreted them for how cookies are going to be placed on your device as a part of your browsing a website.
(22:13):
So that I've seen lots of recently because
you know, there are different aggregators of that response, so you can say, look, I don't want tracking cookies, and yes, I know that there have to be, you know, certain functional ones that always have to be there, but I don't want marketing, I don't want this, I don't want that.
And in some cases, it becomes easier and easier for you to click on the wrong button because you're like, it isn't like, except every cookie, it is, do you want to use the site or click here for more information?
(22:44):
or the ones that just say here, you can click to our cookie policy.
Yeah, yeah.
Right, there's no choice involved. We've made changes and here you can click to our cookie policy.
off you go. You're not holding a cookie in your hand right now, are you? No. Okay. Okay. It looked like it looked like a cookie there for a minute. It was round. It was brown. It's like,
Okay, what if I don't agree? Right. No, it's a it's a coaster for a drink from from a customer from many, many years ago in Little Rock, Arkansas.
(23:14):
So, ah.
oh, it could be a cookie. So yeah, so I pay attention, people. If you're getting these updates coming through, you know, and if you can discern, you know, maybe this is a great you know, drop them all into a rag, you know, chat with documents thing and say, hey, what changed here? Are there similarities, right? Maybe a new law went into effect somewhere that
(23:37):
My my favorite is is all of these ones. I've got they've sent emails that say hey, we've changed our policies Here's a quick high-level summary. We've made changes to this
We've improved that and you're like what were the changes define improvement and then you look at them It's like oh so so basically you've increased your ability to track me across multiple sites and to sell my data to your third party
(24:07):
And you've reduced my ability to opt out gotcha
hooray. hooray. oh my goodness. yes. yeah. well let's do the listener link story since we're we're paying paying attention to listener elements. so for those of you out in games at work land and our Pokemon card players we've got a lovely listener link here that focuses around two North Carolinians not named Michael and Michael.
Well, we are almost out of time, but we have two two quick stories
We might only do one of them
(24:42):
Just in case you're wondering that yes, who are some winners in the Pokemon card game contest that happened.
- No, Nathan and Caleb.
Where's it in Georgia? Not that long ago the the Pokemon Atlanta regional championships.
So pretty pretty cool.
Yeah, what I really like about this story is it's more than just one tournament, right?
(25:06):
It's telling you how these kids got in at, you know, 10, 14 years old. They're 18 and a little bit less now.
How they came at it with very little knowledge of the TV show or the electronic games. It's literally the card game, how they, you know, were mentored by other players. And now, I mean,
(25:29):
One of the kids made
You know Forty something thousand dollars. They're going to regional and national championships and stuff
Really cool. It shows you that
That a game can lead you
To improving things and they talk about you know, it looks like a kid's game
(25:51):
But when you look at the game itself, it's all about strategy and statistics and and analysis of stuff and
The Pokemon card game player
But I thought this was really really cool and and thank you listener who sent
And with that, we're going to go and close out for the day.
(26:11):
So hope you will send us some of your interesting things too.
And you can find us again right here, right next time on games at work.biz.
See everybody.
[upbeat music]
You've been listening to gamesatwork.biz, the podcast about gaming technology and play.
We are part of the blueberry podcasting network.
[upbeat music]
(26:32):
And we'd like to thank the band, Random Encounters for their song Big Blue.
You can follow us at our website at gamesatwork.biz.
[upbeat music]
(27:02):
[music]
[Music]
(00:10):
This is GamesItWork.biz, your weekly podcast about gaming, technology, and play.
Your hosts are Michael Martine, Andy Piper, and Michael Rowe.
The thoughts and opinions on this podcast are those of the hosts and guests alone,
and are not the opinions of any organization which they have been or may be affiliated with.
(00:32):
This is episode 516.
[upbeat music]
Good morning, good evening. Good afternoon, Michael Rowe here. It's Friday. I get to talk tech with my second favorite Michael and my co-host and friend Michael Martin because you know, you got to be your own first favorite Michael, right? Oh, man. How are you doing?
(00:59):
You're my first favorite, Michael Michael.
Hang it in there and we're missing Andy this week, but that's again why we have three co-hosts.
We can carry on forge forward, craft our way into the future and you see what I've done there.
(01:19):
And we've got a great show as we always do curated references, listener links and all kinds of good fun stuff.
So if you've got a story, you want us to...
the touch on, you do pass it our way.
So all kinds of news this week around models, models, models, all over the place.
Yes
(01:40):
And the starting point we have here is an R's technica article about AI prompts,
about how, at least for this article says anthropic is controlling Claude 4.
And for those of us that have been playing in the space.
For a while they're usually very comprehensive sets of instructions that are typically all prompts that various organizations will use to say when a user asks a question like what is the square root of pie or tell me something interesting about the particular book.
(02:23):
That gets parsed by the prompts that are behind the scenes.
And this particular article is talking about the level of complexity that's there that is not usually visible because you ask a question like tell me everything about this new movie or this book or whatever it might be.
(02:46):
And then you get an output and maybe you get a chain of reasoning and maybe get some explanation and some further transparency, but you don't typically see how does the interaction happen between your question.
I like this article because it gets into a little bit more detail and we've had some examples in previous shows about how you can unpack what is in there.
(03:13):
So the example here that we've talked about on gosh a couple of occasions are how the responses from some of the models in the interactions are telling you how wonderful you are for considering such a thought.
the sycophant models. Yes. What I really enjoyed about this article was really the fact that for people who are not deep in the space or haven't spent the last couple years like you and I actually building prompts and doing work, whether it's a hackathon, day job, whatever,
A wonderful question right the relentlessly positive tone right.
(03:55):
And they download
the tool that allows you to locally install an LLM and then they wonder why it doesn't do very well.
So you get the model and you know, you can get a hugging face, you can get llama, you can get what a deep seek.
All these different models locally install them and then not get the results you want.
(04:19):
Well, that's because the model is just the way to stay.
And statistically present the next word.
The important part is in the prompt that explains, you know, how to behave.
What are the guardrails that you have in place?
What's the, as you say, the chain of reasoning that you should take place?
(04:43):
And I've been doing some interesting models in my day job lately around how can I generate my weekly.
This is everything I've done in report.
hmm
And it's working really well.
Oh, is it now?
Yes, yes.
(05:04):
It's pretty interesting.
When you have a entire workplace that is instrumented and sniffed, consumed, gobbled up by the model, you can do some really, really cool things with it.
(05:27):
And so I, I usually, before I shut off on the end of day on Friday, I will run my new agent that I created and it does 90% of the work that I need.
To be able to capture everything I need for, you know, quarterly reviews status updates, blah, blah, blah, and I think that's really cool.
(05:50):
But again, underneath the actual model it's probably using is probably one of the big ones that you know.
This is the magic sauce.
(06:11):
Right, and there once was a time when the concept of a prompt engineer was a real job.
And that was a bit of a flash in the pan, wasn't it, Michael, where, yeah, for a period of maybe about four or five, six months is like, wow, that, you know, you really need people who could do all that, right?
Oh Yeah
Well, as a general as a general position. Yeah, but but for the people who sell
(06:34):
solutions They need somebody to structure that unless you're building your own model
You need somebody who understands prompt engineering at a level to build that
approach to how to answer the question.
And I think it's a specialized skill for a very small group of people now, and that's the interesting part.
(06:56):
Yeah, it is not it's right. It's not a broad skill that you would expect you know a consumer package goods company would hire prompt engineers, right? The consumer package good company should buy us a
Right why would they people who are really good at consumer package goods
What right and if they're using our official intelligence to power their workloads or to design agents or or or that that level of prompt engineering is
(07:22):
is actually almost in the operating system.
Yeah. Which, or at least in the browser, yeah.
Which kind of takes us to the next article.
Really nicely. You saw where I was going to go.
And this article about Opera's new AI browser.
I'm sure you're probably going to go somewhere very similar with this than there where I was, which is.
(07:43):
Yeah, lovely. It's a browser and has an interface into a large language model.
Hooray.
Yeah. (laughs)
It's the user experience.
it happens to be a browser. Why isn't it an app? Why isn't it a hard way?
Why isn't it a plug-in into a browser? It's a differentiator in the sense that, "Oh, this browser is an AI browser. Nifty needo. Is that really that special?"
(08:11):
Well, yeah, it's interesting because, um, and we don't have a link for it for the show.
I had gotten an article, not an article, but an email from, uh, "Have you ever played with the ARC browser?"
So I played with that one a year, a year and a half ago, whenever it first came out for a few months, and it was yet another new take on browsers.
(08:32):
Um, built off of, I think it was built off the Chromium engine, it might have been built off of the, the, um, oh god, what's the Safari engine, the open source part under.
Wept it.
WebKit.
Thank you.
Is it WebKit?
I don't think, no, no, WebKit's the API stuff, uh, there's a, anyway, it doesn't really matter.
I, maybe, maybe may not, I don't know.
One of the, one of the models, um, and, uh, so the ARC browser has decided that they've,
(08:58):
they're scrapping what they were doing with the browser and coming up with their own AI thing, just like this, just like Opera.
Um, and yeah, it's, as you say, interesting.
Yeah.
But the real question is, is this a hype train that they're getting on or are they taking a browser experience and extending it into something new and exciting, right?
(09:23):
Which is like an app right or development environment, you know
Um, right, right.
And, well, or even better, you know, it could be like sky app, which is the next one we want to talk about, um, which, uh, I, I actually signed up for the, the wait list on this one because this looks interesting.
It's not a browser. It's by the guy's or the team I should say that created workflows which became shortcuts when they were purchased by Apple and this does something really interesting because it does download those LLMs of choice and then instrument your Mac. Getting into all the data you tell it can get into and and...
(10:11):
presents you with dynamic creation of... we'll call them agents to do things on your Mac.
Kind of like what Apple showed last year at WWDC, but never delivered.
And that's the interesting thing here.
There's an interesting long-form article by Federici...
(10:35):
Yeah, Vitici. Federico Vitici, I can never pronounce his first name,
together.
And going through his couple of weeks of experimentation with it, as it came out,
to limited alpha users and beta users.
And I was listening to him on a podcast this week, and he was saying,
(10:56):
"He's wondering who will release those features first?
Will Apple break all of this with whatever comes out in the next week and a half?
Or will they be bought by Apple in order to make it so that they actually
promised last year, or will they deliver the functionality that Apple promised,
(11:19):
you know, months, if not a year ahead of Apple delivering it whenever they finally deliver it?"
This looks cool.
And my question becomes, does it truly stay all on your Mac?
If it does, I'd be willing to do it.
If it's sending stuff to the cloud, this is into your personal data.
this is into your email, into your calendar.
(11:41):
into your to do is into all the documents on your machine, kind of like co-pilot, but with scripting capabilities to kind of generate agents to do things.
Yeah, in the olden days, we called these sort of things intelligent workflows.
And part of the promise of shortcuts were to connect the dots, if you will, between certain things that you do on a regular basis, that will make your life easier.
(12:09):
And Automator was a predecessor in a lot of ways to the shortcuts. And, you know,
just being a little vulnerable on a particular front, I tried to create an automator,
And I didn't try to do it via shortcuts.
I tried to create an automation that would allow me to launch a Teams meeting, copy the URL for that launched Teams meeting, and put it in my clipboard so that I could just pop it into a Slack message for somebody to say, "Hey, come and join me on this Teams meeting, I just launched."
(12:41):
And unfortunately, it's probably my mad coding skills weren't quite sufficient to get all the way to the end.
You didn't vibe correctly.
Well, but that's what this is trying to do.
OK, look, there's some things that would be a nice little speed up, you know, to let you do something faster.
And before everyone writes in, yes, I do know about the Slack automation to launch a team's meeting.
(13:05):
I use that all the time and it's really, really good.
But I was trying to go this route, you know, to learn a little bit more, try it out and see where we go.
And the notion of being very granular about what elements of data on your device on your Mac.
that you're going to allow your agent to see and can you contain it in a way that it doesn't.
(13:34):
Sift it, train it, use it, get it out into the wide world.
It's going to be really tricky because even if you do read all the terms and conditions and hold that thought because we're going to get there soon.
Well, well, that's that's that's the interesting thing here because remember last year at WDC WDC one of the things that Apple really really pushed in order to enable this
You may not be in a position where you say, well, I trust it or could those terms and conditions change.
(14:06):
coming soon ecosystem was for apps.
And they've been pushing it for about three years, but they really, really went hard last year, and those intents are not consumable by anybody but Apple or Apple tools like shortcuts.
(14:35):
And so if these guys, the guys who wrote shortcuts have figured out how to
intercept the intents in order to allow them to build these dynamic agents based off of natural language, that is taking us down the path, and that could possibly solve your challenge, too. Yeah. Yes.
Oh, excellent point.
Well, yeah, I mean, I found another work around that works just fine for me.
(14:57):
And it actually is in the Slack environment, which is even better, right?
But, but there, there are going to be situations where, you know, like our post show production, right?
Wouldn't it be awesome that after our show completes and of the recordings are available, they can automatically drop into our shared folder where everybody has
I can get close, but not all the way, and that's because what triggers an event,
(15:19):
Mm-hmm. Yeah.
what's registered, what's recognized, there are going to be a ton of those sort of things.
Now, let, let, let's continue the, the theory here on models and, and how models behave one way or another.
Very cool Oh, yeah
And there is a, another version of the general, hey, models are going down a path where they're going to be less and less beneficial.
(15:44):
that we've been talking about for years.
Yeah, so all the ideas, synthetic data, recursive model development,
GAN models trying to learn from one another, showing up again here in the register.
And while, while search is now really being a consumer, and I think also potentially a
(16:07):
feedback loop into models as well, all you need to do is get some searches that are not quite exactly where they need to be.
Don't have a human intervention around any sort of supervised learning and you're right, you're going to drift into, you know, hey, Michael, you're your new novel, you know, that plot summary for it's pretty cool, you know, so, you know, thanks, thanks for, thanks for writing that new novel, you didn't know you even wrote yet, right.
(16:35):
Yeah, yeah. Anyway, so I just thought this article was interesting of kind of restating what we've been talking about for a couple of years now. And it came across my my mastodon feed from Mike Elgin, who does a whole thing on intelligent machines. But how about something fun? How about something tiny and fun? I want a new Mac, a nice little... Yeah, yeah.
(16:59):
really? And you want to go, you want to go to a new Mac before we talk about Louise?
Yeah. Okay. Let's, let's, all right.
Yeah, because this will be just a quick hit, kind of a pallet cleanser. So another thing that came across the feed this week that I just thought was hilarious was the... What do they call this thing?
(17:19):
I don't know what they call it, but it's... I've actually played with a image of this software,
A Pico Macnano is what it is, a Pico Macnano.
right? You can download this software. That's it. Thank you. The... the...
software images is a Mac operating system from 1990 maybe. And there's a... there's a... a company that is actually printing little tiny Macs, running a Raspberry Pi Pico Zero, and have installed Mac OS running on that in the little tiny form factor that I just...
(18:05):
think is cute. Yes. Yeah. And... and I was thinking... I... I'm somewhere I still have a couple of old three and a half inch floppies of software that I had on the Mac back then, including I think the Star Trek game. If you remember that, there was a network Star Trek game. And I was wondering if it's got a U.S.
It comes in a pea caso Macintosh shipping box. That's funny
(18:27):
Hmm. Yeah. Yeah. We had a way to version of that on
Yeah, yeah, I was wondering if they had a U.S.B.A. plug on it to where I can plug a floppy drive in and read it off and load it up on the SD card on the Pico Mac Nano.
A USB-A plug on this device would be 25% of the size of the device.
(18:47):
Pico Mac Nano. That's a mouthful.
Yeah, I know. It's huge.
The plug, not the Mac.
Yeah, now it's it's cute and I'm sure Andy when he saw this too was smiling quietly to himself because you know This is very much one of those great, you know hack kind of things of
(19:14):
Yeah, running on a three-volt CR2 battery. I mean ridiculous. So very very very
Yeah, yeah, so we do have a listener link and and and I had to interject before before we got to it from from Lewis
Suarez on the internet of consent
(19:34):
What did you think of this article? I thought it was spot-on
I don't agree to anything in the...
So so at least long-time listener friend of ours
from the wow doggier nation days and and that
Linked out on massed on he's starting to finally get into massed on to an article by an ill dash
(20:02):
Called the internet a consent where he puts together kind of a
basic flow of
The quote good old days on the internet to now and how do we reclaim
consent and the reason I think this article is so
important how many terms of services updates have you gotten in the last two days okay I've got like four or five services who have all sent terms of service updates in the last two days
(20:35):
so something either regulatory in the US or something else has changed that they all feel a need to send out turn and these are media ones for like paramount paramount plus
app ones website ones etc so something has changed and what this article goes through and you know we all remember this is the continual and gradual shift that many of your favorite internet service providers from years ago went through very quickly that completely changed the relationship of the data and the consent to the data between the user and the platform
(21:16):
Well, I know you actually read those documents very carefully, right?
And that's something you've shared on the podcast here on many occasions.
And it's a right, why rise and right thing to do?
Because when you're confronted with, hey, we want to collect your data.
Well, what do you want to collect it for?
And in some cases, it's really clear.
(21:36):
It's like, well, if we don't know your name, it's really hard for us to properly greet you and provide you with the service that you're agreeing to use.
You know your data is your data and if you're not knowing how you're paying for a service and it's free you're paying for it with your data, the thing that I've seen most frequently here in the last couple of days is all of the, and I'm going to call them tricks because that's how I've kind of interpreted them for how cookies are going to be placed on your device as a part of your browsing a website.
(22:13):
So that I've seen lots of recently because
you know, there are different aggregators of that response, so you can say, look, I don't want tracking cookies, and yes, I know that there have to be, you know, certain functional ones that always have to be there, but I don't want marketing, I don't want this, I don't want that.
And in some cases, it becomes easier and easier for you to click on the wrong button because you're like, it isn't like, except every cookie, it is, do you want to use the site or click here for more information?
(22:44):
or the ones that just say here, you can click to our cookie policy.
Yeah, yeah.
Right, there's no choice involved. We've made changes and here you can click to our cookie policy.
off you go. You're not holding a cookie in your hand right now, are you? No. Okay. Okay. It looked like it looked like a cookie there for a minute. It was round. It was brown. It's like,
Okay, what if I don't agree? Right. No, it's a it's a coaster for a drink from from a customer from many, many years ago in Little Rock, Arkansas.
(23:14):
So, ah.
oh, it could be a cookie. So yeah, so I pay attention, people. If you're getting these updates coming through, you know, and if you can discern, you know, maybe this is a great you know, drop them all into a rag, you know, chat with documents thing and say, hey, what changed here? Are there similarities, right? Maybe a new law went into effect somewhere that
(23:37):
My my favorite is is all of these ones. I've got they've sent emails that say hey, we've changed our policies Here's a quick high-level summary. We've made changes to this
We've improved that and you're like what were the changes define improvement and then you look at them It's like oh so so basically you've increased your ability to track me across multiple sites and to sell my data to your third party
(24:07):
And you've reduced my ability to opt out gotcha
hooray. hooray. oh my goodness. yes. yeah. well let's do the listener link story since we're we're paying paying attention to listener elements. so for those of you out in games at work land and our Pokemon card players we've got a lovely listener link here that focuses around two North Carolinians not named Michael and Michael.
Well, we are almost out of time, but we have two two quick stories
We might only do one of them
(24:42):
Just in case you're wondering that yes, who are some winners in the Pokemon card game contest that happened.
- No, Nathan and Caleb.
Where's it in Georgia? Not that long ago the the Pokemon Atlanta regional championships.
So pretty pretty cool.
Yeah, what I really like about this story is it's more than just one tournament, right?
(25:06):
It's telling you how these kids got in at, you know, 10, 14 years old. They're 18 and a little bit less now.
How they came at it with very little knowledge of the TV show or the electronic games. It's literally the card game, how they, you know, were mentored by other players. And now, I mean,
(25:29):
One of the kids made
You know Forty something thousand dollars. They're going to regional and national championships and stuff
Really cool. It shows you that
That a game can lead you
To improving things and they talk about you know, it looks like a kid's game
(25:51):
But when you look at the game itself, it's all about strategy and statistics and and analysis of stuff and
The Pokemon card game player
But I thought this was really really cool and and thank you listener who sent
And with that, we're going to go and close out for the day.
(26:11):
So hope you will send us some of your interesting things too.
And you can find us again right here, right next time on games at work.biz.
See everybody.
[upbeat music]
You've been listening to gamesatwork.biz, the podcast about gaming technology and play.
We are part of the blueberry podcasting network.
[upbeat music]
(26:32):
And we'd like to thank the band, Random Encounters for their song Big Blue.
You can follow us at our website at gamesatwork.biz.
[upbeat music]
(27:02):
[music]
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy And Charlamagne Tha God!
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.