Reinventing Newsrooms: AI as Friend, Not Foe - with Pete Pachal - The Trending Communicator

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Daniel Nestle (00:00):
Welcome or welcome back to the trending Communicator. I'm your host, Dan Nestle. You know, people say marriage is hard. You know, like the other day my wife and I were just stressing out about, like. Well, maybe it was me stressing out, really. We were trying to figure out what to have for dinner. Always the cause of marital strife. So what do you want for dinner? She says, whatever you want. I say, oh, boy. Seriously, you have no ideas? She says, okay, how about we do something with chicken? I say, okay, what kind of chicken? She says, and this goes on and on, five, ten minutes until we got a plan. Then she opens the fridge in the freezer. There's no chicken. Sometimes it feels like we're just on both sides of a losing argument.

(00:54):
Which is exactly how I feel about this weird marriage between communications pros and the media. On the one hand, we've got the comms people pitching for coverage, doing what they can to get attention, and on the other, we have journalists and reporters cooking up the stories and content we want to consume. But audiences hate the media. Trust is at an all time low. They're bored by too much dull content. They're looking elsewhere for news. There's no chicken. So what do we do about it? How can we change? And how can this marriage be saved if it can be saved? Do we need a new relationship? Well, it may shock you that the answers can be found in the proliferation of AI and the ways we adapt and adopt to change our professions. And my guest today is right in the thick of it.

(01:39):
A journalist who's held senior roles in newsrooms at Coindesk and Mashable. A tech and AI thought leader who's made numerous appearances on cnn, Fox Business, and the Today Show. He's in demand these days as an AI trainer and consult, helping companies of all sizes incorporate Gen AI into their work. He's the founder and CEO of the popular Substack newsletter, a terrific newsletter, by the way, and you all must subscribe immediately. The Media Copilot. Please join me. Welcoming to the show, Pete Paschel. Pete, how are you, Dan?

Pete Pachal (02:09):
I'm great, thank you. Thanks for that intro.

You know, first time I talked about chicken in the introduction of a. Of a serious podcast, I think about, you know, about AI in our profession. And it just got me thinking that, you know, we are all circling around to grab a piece of this pie that has been sitting in the sun and rotting for too long. Maybe, I don't know, there's. There's a lot to it, but at the same time, there is a serious reward for good content, for delivering news that is, you know, real and good and truthful and, you know, for writing better than the competition, you know, for just delivering good stuff. And, you know, listeners of the show will know that I am absolutely obsessed and a good fan and a big fan of AI, and I think it is a suit.

(03:04):
It is a superior tool for boosting our capabilities to do the things we need to do. But you're right in the thick of it. And I wanted to ask you, really, your take and have a good discussion about this and pick up a couple of really good ideas about, you know, the state of the media and what you're seeing with regards to AI adoption and what it's going to bring to the industry. And I know that's just a small topic, you know.

Yeah, totally.

Why don't we start?

Yeah, yeah.

So I was going to say why don't we. Why don't we start, you know, by just, you know, what's your story, Pete? So how did you get to where you are? I mean, it's really interesting and, you know, I think more people should really get to know you.

Cool. Well, thank you. Yeah, thanks for the opportunity. So I've always been interested in tech. Actually, not many people know this, but my undergrad degree that I got, I don't even want to say how long ago, but it was a while back, is in engineering. And I always joke that my engineering degree is never used because I immediately went to journalism school after my undergrad, got my journalism degree, and then started writing. Of course, one of the first gigs was about tech, and it was at Sound Envision magazine, which people with long memories will remember. It used to be Stereo Review. It's kind of still around in some form online somewhere. I think they discontinued the print magazine just a little bit ago. It was sort of changed hands all over the place. But that was back when tech was like man caves, right?

(04:26):
It was like surround sound and DVD players. And, you know, pretty soon I migrated online where all tech and all tech enthusiasm was going. I did a tech blog when tech blogs were the thing to do for NBC Universal. It was called Device with no E because you had to misspell everything on the Internet. Yeah, I think that still kind of applies today. Oh.

Nowadays they're using a lot of Y's and I's, but, yeah, it seems.

Yeah, totally. And then spent a good chunk of the 2010s at Mashable. I was the tech editor there. It was a great place to be. Not just because tech was exploding and changing, but digital media was like all the rage. And Mashable had its, you know, 15 minutes of fame, is kind of a media darling that had figured out how to use social media to expand. So it was a very exciting time. I worked with a lot of super talented people. We tried everything with regard to media formats and strategy. If there was like a new thing, we would pivot to it.

(05:22):
We were, were basically the original site that pivoted to video and, you know, that's now kind of a derisive term in media and it sort of conjures up fear and rightly so, I will say, because a lot of those pivots kind of led to dead ends, right? It was all about scale, scale for scale's sake. And that ended up being kind of a bankrupt strategy in the long term. You know, I'm not, I don't think I'm telling any tales out of school here in terms of that and sort of migrated to a couple other places and spent a good chunk of time at Coindesk right when crypto was exploding, which was super fascinating. Learned a hell of a lot about crypto. Do you think tech is mental? I mean, just tried working in crypto for a couple years, it's a different level.

(06:08):
And now I'm doing my own thing. Like you said, I was super fascinated by AI. One of the things I did towards the end of my run at Coindesk was figure out their AI strategy because ChatGPT had just arrived and it was clearly going to change everything in media. It was just super apparent, if not chat GPT, specifically generative AI. And so that was really cool to do. And when I left Coindesk, I was like, well, this is what, this is really what I want to lean into AI and media and what does it all mean? And that's basically everything I focus on now.

That is just an amazing path because I could just see myself as on the consumer side of that, right? Like, I, I never really got super into crypto, but I've been following and I'm, you know, familiar with Coindesk and you know, with, certainly with the whole crypto world as a dabbler. But Mashable, man, that brings me back because, you know, I, I was always an early adopter on things and I was, I definitely subscribed to Mashable as soon as I saw it the very, like, the very first time I, is it, this is the future of media? Like, we talk about alternate media or alternative media these days. And I think Mashable was the granddaddy of what might be, eventually become alternative media, although they were not.

(07:22):
They were trying to be almost like a mainstream media in an alternative media suit, if that kind of makes sense in a way. Right. Like a little too early, maybe for what they wanted to do and what you want to do. But, you know, all I remember now, I mean, of course I remember a lot of stuff from Mashable, but all I remember now that sticks in my mind is that Pete Cashmore never answered any of my direct messages.

Well, let me know if you need to get a hold of him. He and I still occasionally DM here and there. Pete's a good guy, but, yeah, Mashable was a great time. I think you're. You're sort of dead on in that the Mashable was trying to be kind of like the millennial version of CNN at some point. So really, you know, struck gold with its strategy, which very few people were doing at the time, in the late 2000s, early 2010s of going all in on social media. So right when social media was just taking off it had sort of created the strategy of distributing on there and then reflecting back a lot of what social media networks, which were full of early adopters and enthusiasts about social media, wanted to read. Right. And so that heavily influenced sort of what Mashable ultimately become, became.

(08:40):
And we had a time again, were trying to be CNN for millennials, but that was also a time when a lot of other brands were trying to do it. And, you know, our innovation of going social first, you know, kind of had a half life on it. Eventually, you know, everyone sort of figured out similar strategies and then it came down to, you know, the talent and your focus and your audience. And I don't think, honestly, don't think there were any winners in the long term. I mean, if you just look at probably what a lot of people think of as the most successful of those, buzzfeed now, it's kind of lost and in a bit of shambles. Right.

(09:17):
BuzzFeed News is gone and BuzzFeed itself is rapidly downsized and is sort of just trying to figure out how it's going to survive with all this debt and stuff like that. Again, I don't mean to pick on BuzzFeed, but it is kind of like. But it's like, you know, there were a lot of brands that sort of tried to become the next generation of the News for the digital generation. And largely what happened was the old generation of news just sort of adopted those strategies, hired all the people who worked at Mashable and Buzzfeed and Mike and Vox and now, you know, the what's old is new again. And so you have the most successful digital media companies now are like the New York Times and various others that are more old school brands.

Well, it's the power of reach and the customer base and exactly like you said, hiring it's cash. Right. They're hiring all the people and adapting the technologies when they have the money to do it. And I think that was the time though, that whole call it 20 years. Right. Like the first two decades of the 2000s where you know, you saw the dot.com. Bomb era and the first news sites started to pop up, the first search engines, all of that were also dealing with news as well. And you started to get aggregators and eventually got upstarts like Mashable. It took a while, but the big lumbering organizations dug into their pockets and were able to just buy and build and get and catch up and then surpass and.

Right. It's.

Yeah, it's because they're all playing the same game though. Like it's, they are.

And, and I would say those old school brands, the Times, the Post, CNN even, all had brands they had established pre digital. Right. So ultimately there was a generation of people and even, you know, even younger people like knew what to expect from of these digital brands, were kind of still figuring themselves out and brand ended up sort of counting for a lot in the long term. Yeah, I would say. But there's also, you know, an interesting reflection on that period and what sort of worked had a lot to do with tech. Right. And had a lot to do with social media and search and the incentives that those distribution platforms ended up giving to the news media and particularly digital news. So like, you know, Mashables wasn't alone in this.

(11:54):
But you know, the quick hit, the idea of essentially the reblog, the Twitter react post, which I often deride as probably the lowest form of journalism, if it even is that it's kind of like a virus, you know, is a virus a life form? I don't know, I just know they, it spreads, it's easy and we don't really want it, you know, so that's how I sort of require think about those old Twitter react posts everyone used to do. But the thing that's changing now, right obviously is AI and AI is affecting everything with regard to media at every layer, like the distribution, the production, just the story ideas even. And that idea of doing something quick and easy and employing whether it's interns or junior level people to just fire out those posts, that's a job for a robot now. Yeah.

(12:46):
The cost of that content and the value of that content is now so low. That's honestly something I think that is going to be healthy for journalism. Again, part of it's not all AI, it's that search and social largely got, well, social anyway, got out of the business of distributing news. Search is declining too for myriad reasons, one of them being AI, but this whole sort of dependence on a nameless, faceless audience that's just going to give you a lot of clicks and then you know, responding to that incentive with volume is largely going away, you know. Yeah, like, because even if you can do it, even if you know whether it is humans or robots doing it for you, it's just so easy, it's not going to give you any long term value.

(13:35):
So I think ultimately these trends are healthy for media because it really forces all of whatever publication you are thinking about who is your audience, who are you trying to reach and how are you making sure they stay. You know, you've got to give them real value and just quick hits are not valuable.

Yeah, bottom line, it's like media itself is on its own hero's journey and now they're going through these trials and tribulations and they need to pass through them to get to the other side where they will be welcomed with open arms and redeemed if they follow the cycle. The by my sort of skepticism about this whole thing though is of course that. And by the way, I'm not skeptical about what you're saying. I just, I think that media itself is just in its death now in the way that we've, in the way that we've always looked at media like as media as a, as a, an institution I suppose. And, and you know there's numerous, there's myriad reasons for this.

(14:41):
I, I've written about it, you've spoken about, we talked about, you know, the fracturing of the media landscape, but really I think it's all about the consumer landscape and what you were just talking about where you know, you can now have essentially bots do the kind of low hanging fruit that, that boring, dull content that nobody really cares about still has to go out there because it's putting stuff. It's, it's filling out somebody's Twitter feed. I don't know.

But, but it's actually an AI reason to do it. But I can get to that later.

Yeah, yeah, and I do want you to get to that. In fact, like, I, I really. The idea of this all being healthy, you know, for journalism, I, I'm 100 on the same page. I think that, you know, you need to kind of on the one hand, you've got to, you know, build up the muscle. You've got to. You. The old, the old. We, we used to think that eating bread four times a day was a good thing. Now eating bread for. Now we know that's terrible. So, you know, you have to redo the metabolism, rebuild the body, but it's still there like you're still alive. Right. So find the right diet and nurse back to health. And I think that, you know, cause all these different health problems. That, that AIs, that journalism is having is ultimately healthy for the profession.

(16:03):
I agree with you. And then when we talk about, you know, you talk about how, you know, how AI is going to help, you know, this is an area like, this is really where I think the rubber hits the road. It's, it's something that is hard to. I think people are having a hard to. It's hard to conceptualize because they see AI as the competitor rather than as their enabler. So I've given you a lot, I think I've just given you a lot of stuff there to work with. But you were going to go into that a little bit. And, and how is AI going to help? You know, where are we going?

Yeah, so the quick application, I guess a statement I would say about applying AI to journalism is it's like super easy to abuse and harder to use. So out of the gate, we saw a lot of folks, not a lot of folks, but several notable folks try to get AI to do the most obvious use case kind of write articles.

Yeah.

You know, and we all know how that turned out. Not so great. That said, that use case isn't going away, and I'll return to it, but the idea of using AI in a constructive way in the story writing, building, distributing process is much more compelling. Of course, it requires both sort of an understanding of how to even approach AI first, then figuring out the right places to apply it into a newsroom or a storytelling process, and then sort of trying to think about the next level beyond just you know, efficiencies, because that's where everyone starts. It's like, okay, here's my story process. And I always think about this as like a three chunk pipeline. It's like the story conception, story writing, story distribution. And AI in its current form generally has more to offer the two ends of that pipeline.

(18:00):
So if you're, you know, AI is a brainstorm buddy, AI is a research buddy. I mean, this is widely used now. I don't know a single journalist now who doesn't at least use perplexity or something like it every now and then, at least. Right. And there's other ways to improve that. Chief among them is essentially using AI with your own data, whether it's your stories or in the case of, say, a legal reporter, using it on all these sort of legal documents around a case, which is a relatively easy thing to set up. If you think about, say, Google Notebook LM as a good example. Right.

If you, My love it, love.

We all love it. Right. The whole podcast thing is amazing. But if, you know, like that idea of let's put all the documents of this court case and maybe even all the stories I've written about it into a folder and then essentially have a conversation with that content. What are some story ideas, what are some things that I haven't been able to put together here within this corpus. And, and, or if you have an idea, bounce it off with that whole AI that's managing that super useful. Right. And not hard to set up anymore. Yeah, used to live about a year ago. It's a little harder. But you know, with these new tools, it's great. So, so there's a lot of good ways to sort of improve that process.

(19:19):
Certainly on the other end of the pipeline where you get your stories done needs to be edited, it needs to be proofread, it needs to be distributed on social media, it needs to be optimized for search. All these other things that you want to do with it, AI, that's a great job for AI. I don't know a single like journalist from the past 10 or 15 years who would love a bot to take over that. What many people regard as busy work. Right. Like I always say this, that over the last 15 years or so, every reporter editor was asked, in addition to be a storyteller, they need to be a content marketer for their own stuff.

(19:55):
Not only do you just write your story, you got to make sure your headline sings, you got to make sure it's great for search engines, you got to Get a thumbnail image that's going to stand out in a feed somewhere. And you got to do all these things. There are bigger organizations that offload that to specialists. But most, for most people, you got to. That's all you, man. You got to do that. And now AI can probably do 80 to 90% of that. Sure, yes, you need to still like check it. But even that's getting less true. Right. With, with AI getting better at writing. Yeah. Prompting getting better and then detection of hallucinations getting much more sophisticated. So I have no doubt at some point within the next five years a lot of that will be like 100% automated for sure.

I mean, agents are already being developed and tested for content creation is maybe one of the top use cases for agent creation and distribution and so on.

And so then there's. That leaves the middle part, the actual writing, the storytelling of it. And that's obviously the diciest place to deploy AI. Most newsrooms don't. They will have generally a blanket rule that if there's any generative content, it is forbidden from being put into a story. And those that might allow it will certainly require transparency about that fact. This is going to, I don't know if it's ever going to fully reverse maybe in a far future where literal robots are literally sentient. But you know, put getting outside of the sci fi scenarios, I think you'll gradually see more newsrooms doing robot written articles for basic stuff.

(21:31):
So one newsroom that is notably doing this is espn and they came out a couple months ago and said we are going to use AI written articles to make sure certain sports that we want to have coverage for. And I think it was like lacrosse and women's soccer or something like that. We're going to make sure we can do it. And you know, it's very specific. They have templates, they have editors checking the work that it allows them to sort of fill in the gap here on something they want to cover. Now why would you want to do that? Well, there's also, there's sort of the basic media incentives of like I want to get, just get some clicks and some audience for these articles. But I think there's actually a better reason to do this kind of thing.

(22:13):
And I think that the reason other orgs may want to start doing this for basically stories they don't have people for but are still relevant to their beat or their area of expertise, is to make sure their corpus of content includes those stories. And because those stories will be in their broader storytelling, their corpus Whatever AI system looks at that holistically, whether it's theirs or someone else's, whether it's a chatgpt or perplexity will rank them higher, so to speak. Now again, this is all early days for this kind of, this layer which I'm getting at, which is sort of an uber layer of AI over media in general.

(22:56):
But it seems clear that the more authoritative, definitive and comprehensive you are on a particular area that you will be, I guess you could say favored when an AI is looking around for information to summarize for someone who's asking about it. So that's sort of, I would say, the long term strategy reason to start thinking about using AI as an actual writer. Again, all caveated with. You gotta have a really good editorial process around it.

Absolutely, yeah.

Because not only the things to consider aren't just is it good to get it wrong? And that's bad. But even if it gets it right, even if it's hallucination free, what does that do to your audience's perception of your brand? Right? And we're just at the beginning of this, you know, we don't know. Like, I mean, I'm sure ESPN is going to be really excited to sort of see whatever first results they get from surveys on do they, do people care about these AI written articles and is it affecting their brand?

Well, I love the idea of catering to your kind of, let's say, peripheral audience. Right? The, you know, you have your core fans and the people who are really counting on you to deliver, you know, the best work possible. But why, you know, if they just want to know a little bit about, in this case, lacrosse and I don't want to downplay lacrosse, by the way, you know, we're both New Jersey dads. I know that we know a lot of people in the lacrosse world, but, and you know, it is a hell of a sport. But the fact is that, you know, not a lot of people are watching lacrosse or reading about lacrosse.

Well, my son plays Ultimate Frisbee and similarly, I'd love to see that written up and covered in more mainstream media.

Well, I mean, you know, Ultimate Frizzly, Ultimate Frisbee is dead when you are 22 years old and out of college. That's the way it is. Right. Nobody takes it seriously. Unfortunately, it's the same thing. Right? You want to see some content about this. And more than that, you know, if you're out there and you're subscribing to newsletters or you just want to see what's going on in the lacrosse world, it's better if it's delivered to you from places that you already trust and even if it's a little bit more shallow or if it's AI generated, you know, at least, you know, you're sort of still within the same garden. Right? You don't have to go to a different place and it's smart. And then of course, from the, you know, from the perspective of what are the, a, what are the AI.

(25:27):
I hate, I hate to call them crawlers because it sounds like such an SEO term. But you know, what is AI looking for when it is absorbing and learning from new websites and when it's being trained on new data, you know, it's going to be approaching it from a curious, from a curiosity standpoint, almost like it's asking questions. And if your website is, if your content is answering those questions, it is much more likely to be included in AI's answer to somebody's query. So very different than SEO, but in the, say at the same time somewhere because it's just because the content, you have to add the content to your website.

Yeah, yeah, I think the ingredients are a little different, the emphasis is certainly different. And it's very early days for this. Right. I don't think there's even really an agreed upon name for this, which, you know, is probably broadly called SEO for chatbots. But you know, I've heard generative AI optimization or all that stuff, AI engine optimization, you know, so the only thing that I, I advice I have, and it's very early days, is that your content to be favored by one of these summarization engines, needs to be definitive. And what does that mean? Well, it means that if you were to write up the ultimate summary of this topic, leaving out the information you have would be wrong, would be neglectful that you need that information. So it can't just be a duplicative regurgitation, essentially a reblog of something.

(26:59):
And even though I was sort of just almost defending that practice for AI bots, I meant that in the sort of web of coverage that you have. So in other words, if you're covering lacrosse to go with that example, you know, I think you can't just do sort of the, the regurgitated things. You would also need to occasionally do some kind of feature, some kind of deep dive on this that no one could get anywhere else. Right?

Yeah.

And so whether you're a super hyper specialized site that is does that sort of thing or an espn, I think, you know, you have an interest in doing it. Yeah, either way. So. Yeah, so basically getting back to the SEO for chatbots thing, aside from sort of the ingredients of sort of getting into those summaries, which I think is an area of great interest for everyone, whether it's media or PR or what have you, is why would you want to, you know, like, and I think that's a fair question because your metrics for success, like what is your reward? You know, what do you get if you're summarized well, you're not getting conversions probably you're not getting a lot of click through. You know, you don't get numbers.

(28:10):
And because the whole point of a summary is to give what the person who asked the question what they want and they can get out easily, that doesn't mean nobody's clicking through. And apparently Google has said that the quality of the people who do click through is higher and that's, they're vague about that. But what that seems to indicate is they, if they go to the site, they spend more time, there's a greater chance of conversion, blah, blah. So there's that. But I think what is ultimately going to be the, you know, the quote reward for this is what is the economy that is emerging around this whole idea. And so right now there's a lot of emphasis on licensing.

(28:51):
So if you are in an AI powered search engine like the one ChatGPT just launched, there's others that are in the works like Pro Rata is a company that they're writing all these licensing deals with various companies. We're starting to see Microsoft and Meta and even Apple start making deals with publishers. Amazon, there's a report they're talking to people because they want Alexa to tell you news summaries next year sometime. Right. So all of this stuff is kind of good news for publishers in that, oh this thing that people were sort of half dismissing about a year ago. It wasn't even a year ago come to think of it. It was like six months ago because it was like, well is this going to save the media? Probably not.

(29:32):
But it does look like if these deals continue to get made with various well moneyed tech companies, it could become a significant line item on media companies balance sheets. That okay, here's income from AI engines. So that's ultimately I think the reward, you know, that's why you care about this SEO for chat bots and doing things in a form that wants to be summarized. And as we found out more and more about that, like I say, I think doing definitive content and making sure you're bringing something unique and valuable to your stories will be the thing that these bots ultimately reward. And to my mind, that's what we used to call good journalism. Right. So again, this is, I see a path here and I know this is probably a little, sounds a little Pollyanna ish, like it's a, this is all the.

(30:26):
If everything goes right and the incentives turn out the right way, this will better for journalism. Big, big caveat because there's one other thing like about this. So far these licensing deals have been almost entirely for big media companies, not small ones. Right. So that's, I see that as a big problem that needs to get solved that. Well, what about smaller publishers? What about independent publishers? No one like, you know, I run a newsletter like I can never get open. I'd answer my emails if they wanted my content and I'm sure they don't.

(31:02):
So luckily there are a lot of startups trying to figure this out, essentially doing a self serve scalable solution where you can essentially put your content on a marketplace and then, you know, AI engines, if they want to index it can then go to those marketplaces and you know, buy them or not. Yeah, that's a great idea. A bunch of startups are trying to do it. It just depends a lot on can they get the attention of these big AI companies which have mostly said they're not too interested in, you know, the small fish.

You're talking, you're talking primarily about licensing for these news kind of summaries, right? I mean by virtue of putting your content in the public domain and putting your newsletters out there, for example, it should be used somewhere by. Right to be training AI anyway. Right?

Well, that's the old system.

But yes, you're right, the old like as in just a couple years ago or three months ago, whatever it is. So in the, in that case for the, you know, for somebody who doesn't have the kind of cache or cash to make a licensing deal with chat GPT or with, you know, with OpenAI or with, you know, with perplexity or something, it's still the good old fashioned, write great content, share it, get it out there, be seen and be recognized as an authority. Right? And you have the number of people who are talking about you and quoting you and their own pieces. In other words, you know, what might have once been called, you know, still is called backlinking and all of that good stuff. But the links aren't as important.

(32:47):
It's more about like, are you, is your content substantively and significantly from an objective standpoint, a part of somebody else's content? Like, are they quoting you? Are they talking about you? Is. Is your point of view incorporated in their point of view? You know, all of this just increases the odds, let's say, of AI kind of understanding or knowing that there's a good probability that when somebody asks for, hey, who is the best tech journalist that I want to talk to about AI, right? It comes. If AI has been crawling recently or been looking, been trained recently on content out there, the more people are talking about Pete Pashel, the more likely it is that Pete Pashel will be one of the answers that AI comes up with. Because it's a probability game. Right?

Right. Yeah.

And that's, but that's not a game that any, that you can't, it's. You can't kind of beat the system on that by getting a licensing deal. That's all organic, right? That's, that is something that is just. We're figuring it out and the definite. Putting more definitive content on your site is one way to get more attention and kind of get out there with the engines and so on and, or the AI and such. And you know, it's a, it's a slog. In other words, it is.

And the thing is some of those incentives of SEO will carry over to some extent. So even like, you know, it goes back to the question I asked before getting summarized in an AI engine. Why would you even want to. If no one's clicking through? I think there is still a certain amount of exposure. Right. And again, if Google's right, and even though there aren't that many people clicking through, but those are high quality people, even for someone like myself, like, well, I might as well put my stuff out there for free because the content, I wouldn't call it a loss leader. I do have paid subscriptions, but I have a business, as you mentioned, of teaching AI, of consulting in AI. So being exposed to more potential clients is always a good thing, right? Exactly. So that will carry over.

(35:05):
I often look at this because I'm a tech guy. I always look at these things from the user perspective and how valuable this is would this be news summarization and how will it be used and how do we make that a good experience? Because. And it's beyond just new summarization, even though that's the broad term. I think what were just talking about, Google Notebook LM a bit ago and we have a now we're kind of on the cusp of what I would call reversioning content, right? Because I think NotebookLM has opened up people's imaginations to that. Oh, it's not just text summarizing text. You can actually take a text article and turn it into a podcast and well, why stop at podcasts? Why not make some TikTok videos or some long form videos or what have you?

(35:57):
And I think this is, I mean, certainly not just a possibility. This will certainly become a thing going forward where to some extent a user of, you know, a consumer of the news will be able to essentially get their news in whatever form they want with some limits, right? And then, you know, if you have a, you know, publication like the Times or whatever, and they only want their news done in the formats that they want reversioned, I think that will generally hold true. But at the same time, you can't stop people at the end point. If I put a New York Times article into NotebookLM manually on my end, I could listen to that as a podcast and there might even be hallucinations and stuff. And so I'm sure these are things that the Times and others are thinking about.

(36:49):
And if someone were to do that in a systemic way, that would end up being like, well, maybe you need to sign a licensing deal, right? So this is the whole thing, the crunch of where Perplexity is at, where they don't have licensing deals. They have introduced a rev share program for publishers, but they are operating on what I sort of identified a few minutes ago as the old system, the old rules of like, oh, if you crawl the Internet, you can just use that content. And that's honestly what everyone's doing, right? Because even though OpenAI is signing deals with these big publishers, they are still ingesting the entire Internet and training their AIs on it. They just want to make sure their search engine can surface current news and they don't want to get sued for it, right? So it's all defensive.

(37:34):
But this, as you think about it, and you know, I'd think about this daily on, you know, what this means for media. It's such a contradiction. It's like, well, is it okay or isn't it? You know, and it's to some extent it's sort of being figured out with all these lawsuits and everything else. But I do think we're going to need some kind of ruling in the next year or two of to what extent is it okay to just crawl, index and ingest to an AI engine as a kid? Shouldn't be zero. I don't think anyone wants that. Impossible for an AI company to index anything without signing a licensing deal.

(38:15):
I mean, go ahead, index geocities all you want, but again, when there's a replacement essentially of the economy that old, that media has relied on, I think there needs to be some kind of recourse beyond just depending on the largesse and the generosity of a big tech company.

It's an interesting kind of conundrum, I think, that's happening here because there's, on the one hand, you have this industry that is being seriously disrupted and a lot of people who make their livings that way and a lot of companies that make their money that way, you know, they don't want to be snuffed out and nobody wants to snuff them out necessarily. You don't want to put all those people out of work, but at the same time, you know, can't, you can't stand in the way of progress for too long. And, you know, it's more like it feels almost a little bit like, okay, we're going to, you know, come up with some way to ease the transition, but the transition is going to happen. So, you know, we're. You, you'll make enough. We'll figure out these licensing deals.

(39:25):
You will, we'll make sure that there's enough cash flowing into your coffers. But if, but slowly but surely that you're going to have to kind of, you know, either upskill your people or, you know, transition them to different kinds of roles, etc, but it's kind of like buying time in some ways. Then on the other hand, or maybe on the same hand, there's a, I was reading about this recently that none of these court cases, none of this litigation is actually ever going to go anywhere because, you know, this is almost a little Tin Hattie maybe, I don't know. But, but the fact is that AI is a, is a national security issue in some ways. There are, you know, there are competitors in China and in Japan and everywhere else.

(40:06):
Not that, not that they're all, you know, hostiles or that they're kind of, you know, enemies of America. Or any of this kind of stuff. But, but many places around there, around the world are developing different LLMs and other kinds of other kinds of tools. And you know, litigation is one way that we will slow the process of innovation. So, so there's a, there's a sort of game that has to be played on this knife's edge to, you know, to make sure that innovation happens, but at the same time, you know, recognize that people do need to be recognized and compensated for what they've done and for their real original work and so on.

(40:48):
So look, I don't envy the legal geniuses who have take care of all this stuff, but it's a, it's not as clear and cut and dry that licensing is ever going to be really required eventually. And it's kind of a weird future to look at. But I wanted to get back to something you said earlier about, you know, you, it just sort of triggered something in my head about the writing part of the process. Right? We have the, you know, you have conception, writing and distribution. When you're talking about the job of a journalist or of a writer of anybody really who's creating content and that newsrooms had created rules that, and guidelines that there will be no generative content in, you know, in this, in these articles, etc.

(41:42):
I think that they're playing on a, they're playing a game on a field that was, you know, already overgrown several months or years ago. Like they're behind there. There's always a lag, there's always a lag between where technology is and where companies are. Right? But in this case the gap gets fat, gets wider and wider as AI advances and what. The only reason I'm really kind of thinking about this is because of Claude's newish style creator, right? Where, you know, newsrooms, etc. Newsrooms can, looking at the stuff that they were putting out there with, you know, GPT 3.5 and GPT 4O, that's when they're making the rules. They're making the rules based on that level of, of writing and in some cases laziness.

(42:34):
But when you look at something like Claude, where you can, you know, basically train it to write in a certain style in a very short amount of time, you know, that is going to pump out a piece of content that is 80% there, 90% and the consumer A would never know and B, might Even benefit from this because, you know, the content is good. So when you add in the skilled writer and editor on, you know, sort of on both ends of that process where you're teaching the, teaching Claude your style, giving it the instructions and then massaging what comes out in the other end, isn't the argument, isn't there a good enough argument there? Well, that you're actually creating good factual content and it's based on what you do anyway, sir, I don't know.

No, yeah, yeah. This is the existential sort of issue that I think a lot of newsrooms are going to start to face. And it sort of ties back into what I said earlier. AI is super easy to use or, sorry, super easy to abuse. Harder to use. However, it's getting easier to use, you know, and I think with the feature of that you mentioned about Claude, I think that's a good indicator of that. Right. Because those old rules were not old rules that are actually really new rules. But I mean the rules based on older versions of AI are well founded, you know, AI out of the box, particularly Chat GPT, they tend to default on this sort of formalized marketing speak and everyone sees coming a mile away now it's an issue. There was a recent ChatGPT update where it's supposedly better.

(44:19):
I don't actually see it, but maybe others have. I always have thought of Claude as a better writer.

Yeah.

And now with, you know, better prompting and basically your own data and training, you can get it to get like you say, in terms of style, like 90% there, like.

Sure.

And so here's where I think there is an important, it's important to think about what your rules are on the article level and the content corpus level. Right. So as I was sort of saying earlier, I do see a world where we're going to move into where some articles and this could be like, say there's a reporter with a beat. Whatever it is, you know, it could be business, it could be. Well, it probably wouldn't be business. It'd be some, you know, the oil companies or whatever I cover big oil. And so whatever announcements are happening within that beat the news day to day, maybe a good chunk of that is AI assisted in the way we're sort of talking about. I have a bot that's trained on my style.

(45:24):
It has the corpus of coverage and a lot of facts about the oil industries that I'm covering all to its, in its, in its training data. And I can, I have an editorial process attached where I check it and get these things out. I. Or an editor. And that is. That sounds fine to me. You know, that sounds actually like, this is how you would do it. The issue becomes like, well, that's intended to free you up to do the bigger stuff. And how much should AI play a role in writing that stuff? The stuff that's more evergreen but important, the investigations, the enterprise reporting, the interviews, the stuff that requires. And then I would argue, very little AI should touch that.

(46:11):
Because the thing that always rings true to me is something I don't think this guy, this is a quote from him, but it's true. An old editor of mine, a colleague of mine at coindesk, would continually say, writing is thinking. And he's right. The act of writing is not separable from the storytelling phase of it. And I think every writer knows this. As you're writing a story and you're, you know, you're figuring out transitions and things, you're like, oh, right, I should say this or I should say that. And usually the end product of that story isn't. Might be way off the map of what you originally thought you were going to write. Then you sort of have to go back and change your headline and maybe even your premise and refine it some more. And that's all good. That's natural.

(46:55):
That's actually how you get what you would call in the business, like scoops of insight. Right. Like, oh, I put these together, I thought it through, and I brought in other stuff. And now this story is unique and differentiated and, you know, as a bonus, the kind of thing an AI indexing thing would probably want to ingest. So I think that. So it's important to have an AI policy if you're using it in the writing process, to make sure you're making distinctions between these. At least these two things, probably more, but certainly, like, if something's AI assisted, then it's a certain class of story, but the thing you want to be really known for is something else, and that's got to be human.

(47:40):
Because the whole consequence of the AI bringing down the value of content generally is that human content is going to be more valuable than ever. Because the pie of content is constantly expanding now, massively with AI, but the amount that humans can do is always going to stay the same. Right. Or at least not scale nearly as much. Yeah, certainly with a lot of people leaving the business now because of layoffs and whatnot and all these altering economies. But that again, like, in a. In a world I Don't see any world where that content isn't more valuable than it was before.

I'm with you 100% on that and I have been speaking about that recently, that, you know, it's as we're recording this, the end of, towards the end of 2024, and you know, it's always like prediction time. What's, what's happening next year, where is this going, blah, blah. And you see this all over the place. And I think, you know, by the time this is published, it's not going to be too far from now, but It'll be early 2025. And my feeling is exactly what you just said, that good writing, talented writers, skilled writers actually are in an advantageous position for the future, but they have to be really good at it.

(48:57):
Like, you have to have people who are just excellent at the job or at the, you know, either they are naturally great writers, well trained great writers, but people who can actually construct an argument, who are critical thinkers, who are curious, who can, you know, go back, revise, rethink, re argument, re augment, you know, all of that is not going to go away, but, or maybe, and I should say, and that person, those people who also have a process that is kind of supercharged by AI are going to be in an, in a great position. I don't think they have anything to worry about in terms of getting their, in terms of their professional careers, in terms of making great content.

(49:51):
They're just going to be able to do it a little bit faster, have maybe more research at their disposal, be able to ask themselves questions and get answers far more quickly than they used to be. The writing process itself, you know, the process of thinking as you write is still thinking as you write, but there are like, you know how it is. It's, it's these little things that add up as you're writing a longer piece or as you're, you go through your, you know, you put down your draft or you get your outline and then you start to go back and you start to question things. Or you know, you, as you get into a transition, you're like, what's the transition here? Should, what should this be?

(50:28):
You know, in those instances where you're trying to make those quick decisions, you might turn to AI and say, you know, what am I missing here? What is it? What, what's a transition here? That, that might work and then you get some ideas and they'll spark something and you get it. Done a little faster. But the quality of what you're doing is not suffering and in fact you could probably do more. So anyway, I think that's, that is a construct for, for a writer of the future or of the very near future in the now, actually combining the different tools we have, reversioning, as you said, you know, content in different ways.

(51:05):
A great writer will be able to do that, whereas somebody who is relying too heavily on AI to write their stuff and doesn't upskill isn't able to get out of the mediocrity of dull writing. Make your bed. That's all I have to say about that. You need to just be able to understand where there's, where there is wheat and there's chaff and nowhere you lit where you are.

Yeah. I think what you've just outlined is sounds good. And the thing that, I don't know if you know, the knowing the nature of people a bit like there's a fine line between use and abuse. There's a fine line between just, you know, getting a little AI help on writing a story versus it, just doing it for you.

Right.

And the thing that probably I worry about the most in this whole process for the, and what it means for the future of newsrooms, I think about what are the interns going to do and the junior level people. Right. Because what I've outlined in terms of my sort of oil writer person, that's someone who's on the beat, they've cut their teeth, they paid their dues and they're probably very skilled and have a lot of expertise in the field they're covering. Whereas if you're brand new doing a lot of those little stories on your own with the mentorship of an editor like that is how people get good at that and become senior writers. And you know, you go from assistant editor to associate or whatever. And should we mandate a. Those are AI free processes.

(52:48):
You could, I just don't think in a future where it's there and you have the quality of a clod that's even tenable. So what is the right process? What is the way to think about an organization where you don't lose the pipeline of people to become the senior people, you know, three, five, ten years from now. That's, that's a thing I worry about. And I haven't heard a lot from the newsrooms I talk to in terms of thinking through that problem again, other than they just blanket forbid AI content.

Yeah, well, I mean, let's go broader than newsrooms on that. And maybe this is a good chance to start looking at the PR side of the coin. But what you're talking about there is, is a question that, you know, I've been trying to noodle around a bit with some of my guests and just think about, you know, we're approaching it from this paradigm that says, okay, you start a job at the entry level and you do this good work and you get promoted and you grow and your career goes, you know, to certain directions. And if there's 100 people who start off as interns in a PR department or on a news desk, then, you know, many of them will enjoy long and healthy careers in the industry. They keep going, right? But that.

(54:16):
It's never a straight like 100 people start, 100 people finish. Because it can only. Because it is, it does narrow towards the top and people exit the profession or they stay in different levels, you know, or they just kind of, you know, find different things to do. You know, I think it's no longer even remotely realistic to say that 100 people start and 100 people finish. I think it's going to be more like, you know, 10 people start because it's a less attractive area for people to get into because AI can do it, just supply, demand. So, so you're starting with a smaller group, which actually is good for the newsrooms and the PR teams because they're, you know, there's more to do for ten people than for a hundred who are all trying to do the same thing. And then, you know, that.

(55:01):
Where it goes from there, though, that's a, that is something that people need to, that we as a society and certainly as a profession need to rethink because that career path, as it were, is completely unknowable. Now only one person's going to be a cco, but is there going to be a need for a CCO 10 years from now? I mean, what does corporate structure look like? Job descriptions are going to change. Like the, you start going down these alleyways and pathways and rabbit holes and, you know, the answers just are less and less clear. I mean, I don't think companies or corporate structures are going to go away. I just think that there's going to be a very different way to assessing resources, meaning people assessing value of people in the jobs that they do. I don't know what that is.

(55:55):
Yet, but anyway, like from the PR side of things, right? There's, we can only look at the now we can look at what we need to be doing for the next few years. You know, and we've talked about newsrooms. You got the other side of the coin, people having to deal with those newsrooms. I've been on the side saying, well, you know, PR comms teams have to stop putting so much weight on that media relationship that you know, it's more important to have a much broader view of the audience content consumption habits where audience gets the news, right.

(56:42):
So instead of focusing all your time and money saying I gotta get a hit in the New York Times, for example, it's far better spend of your time and money that to hey, let's look at podcasts, let's look at bloggers, let's look at these great sub stackers, right? Let's broaden the horizons and get, you know, we're not going to get a billion impressions from a. Which, which are kind of useless anyway. We're not going to get this reach that we might have gotten from one hit at one time, but we will get lasting longer attention earned. Earned attention and engagement by going across channels where people actually want to read content and view content and listen to content. So from the PR side of things, you know, they need, they still need to be like, how are we going to get the earned attention?

(57:36):
But then there are always going to be, you know, larger companies, highly regulated companies with financial concerns, like these kinds of things where the mainstream, the main legacy media will always be critically important. So like, it's not going to go away, it's just going to change is my point.

Right? Yeah. And I think the value of all of those things is going to change and the ease of reaching them and then, you know, making sure that you have kind of a comprehensive picture of where like basically who is going to respond to, you know, what you have to say is I think the AI brings a little more honestly clarity there. But I mean it's all potential in that. Oh, it can, it is probably better than you are at like understanding the trends and the motivations and where, where audience is essentially. And so one of the things that is slowly becoming a consequence of these licensing deals and the lawsuits is like what these AI summarization engines will have access to.

(58:48):
And that's going to affect a lot of things because you know, I know folks in Gen Z who are starting to essentially go like, oh, I'm Just going to turn my search into chat, GPT or perplexity. And well, if you are a publication like the New York Times, which is pretty, you know, forcefully said it will not allow AI engines to index its content, that means you won't have access to those articles in those places. So I think this is a little bit existential for the Times and others. Now the Times is one of the very few media companies that has its own platform that is reaching so many influential people that it's a, it's still obviously very valuable. But the, that's not the case for many others. Right.

(59:34):
And so it becomes this, you know, a bit of a devil's bargain on do you want to be in these AI engines and potentially if you're not a big company not really reaping the financial benefits, but then you do get to be, you know, front and center for potentially for Gen Z. Right. It's going to be. Going to be tough. And as a PR person sort of thinking about like who should I pay attention to and where should I concentrate my fire? I don't have all the answers to that, but I do think AI is essentially altering what you might have done a couple of years ago so that you know your where you're. Where you're basically directing your resources and what you consider a win is going to be very different.

Daniel Nestle (01:00:23):
Absolutely. Yeah. I, I'm both impressed and a little scared by the proliferation of AI tools for PR and for communicators to really just take the burden of pitching and the time of all of that and the kind of research off of your hands so that you can focus in on serving your clients better or writing the stories you need to write. Because fundamentally PR comms is also about writing and about messaging and about excuse about understanding connecting strategy to story and all that good stuff. And if you're spending your time kind of pitching, it's for that one hit, it's counterproductive. And the AI tools that are out there now, there's like PRofit or like propel or there's. There's a variety muckrack does this kind of stuff. Your, you're doing that.

(01:01:25):
You're leaving the heavy lifting of the reese of the research and in some cases even the pitching to these, to. To the, to the tools. Now I don't know any journalists who want to keep. Be inundated by p. By pitches from, by AI generated pitches and. But they don't want to Email pitches anyway, they don't want anything except for, like, if you're that kind of like convinced that your story needs to be, you know, on channel X. I don't mean X like Twitter, but I mean like on channel. Channel X is like just hypothetically and you're leaving it to AI to pitch channel X. Well, then you're. You're committing some malpractice there.

Pete Pachal (01:02:08):
Right.

You should be if you know that's your channel. But most time you just don't know. There's like loads and loads of alternate media out there. There's. There's so many sources, you know, you can't possibly get to them all at once. So there's value there too. It's just kind of finding the balance. And like you said, it's knowing it's the difference between use and abuse. What is the right balance? I don't know. Yeah, I mean, we talked about pitching, but.

Well, no pitching. Yeah, I'm definitely familiar with a lot of these sort of auto pitching tools. Some are better than others. And I think, you know, it's like you say in the use versus abuse here, I think the lines are a little different because. And it really depends on how you define success. Is it a numbers game? Do you need so many, you know, whatever it is, podcast interviews, for example, or whatever on something, and then you can certainly ramp up the automation on a pitching campaign and get numbers. But to what extent are the people who are ignoring those pitches because they can smell an automation has that hurting it overall? Right. So it becomes, are. Are you selling hamburgers or steaks? Almost.

Yeah.

And if. I think most people that would value the stakes, and in that case, I think those tools can be helpful at sort of discerning things, analyzing things, and helping you get a little closer to the finish line. But you really need to go in and give it, you know, first understand who and what you're pitching, obviously, and then give it that human touch. Right. And so, yeah, these tools are all, you know, kind of very different. And, but there's also, there's always a certain amount of how much do you let the AI write? Yeah. And I think again, with stuff like Claude, that line is sort of slowly but surely getting, you know, bigger and encompassing more stuff. But I think we're still a ways away from 100% for sure.

And, you know, I think also people overthink it sometimes. It's like the date, the big issue, when you're putting AI written AI generated content out is not telling people it's AI generated content. I mean, it's really, it's an ethics and transparency issue. And I think that a lot of people out there would be receptive to, you know, if they know it from the get go and they agree, like the, like classic opt in type stuff, they say, look, I'm agreeing to get this information that's AI generated. Therefore it doesn't have to have some kind of crazy, wicked, funny style that's like your style. It doesn't have to mimic it. It just has to present the, present it in a digestible way. And there are newsletters out there that do exactly that. Right.

(01:05:00):
It's where you start to say, oh, I'm going to make the AI write exactly like me and don't tell anyone that you're AI assisted, that you start to run into some problems. I think maybe people won't know the difference, I don't know. But I think ethically it's an important thing to say, hey, look, I hope you're enjoying this. It was 85% AI generated, right? And just be, be open about it, you know, and same thing with the, when you're pitching or look, I'm a podcaster, you're a podcaster, you know, you must get pitches all the time from, right? From people who are like, hey, dear host name. You know, been listening to your show. That, that episode with fill in guest here was particularly insightful and here's somebody you might want to consider for your show. And then it gives like a little rundown.

(01:05:57):
Now you could be offended by getting AI pitches, but look, I know it's an AI pitch. I'm going to look at the guests that they pitch me. I mean, you never know. There's, there's diamonds in that rough sometimes, but I'm not offended by it. I don't care. I have a delete button. It's cool.

Well, podcasts are almost, I wouldn't call it unique, but because of the low barrier to entry and there's so many of them now, you know, you can't expect everyone to know every podcast out there, of course. And so, you know, automated pitching. I think in podcast world there's probably always going to be a certain amount of it that you need tolerate. And so the thing about sort of the transparency is such an interesting issue and I sort of alluded to it earlier and like changing the relationship with your audience if you sort of tell them that, you know, some of These articles are like AI written or something visceral in us that reacts to, oh, this was written by AI. You feel a little bamboozled, right? Like just inherently.

(01:07:02):
And then you have to sort of take that extra second to think like, oh, well, maybe it's not so bad in these particular use cases. So I think it's worth, you know, kind of unpacking what that means. And there was this great New Yorker article a few months back, it was right around Labor Day by Ted Chang, and he wrote about how, like, what AI is essentially, and it's outsourcing a certain amount of decision making on any particular project and getting what is thought to be the average of what humans would actually have decided on that. And I liked that because he got into sort of what that means. And because AI doesn't have intent inherently, there's no real human there. It feels like, well, why should I even engage with this thing now? The human behind it has some kind of intent, right?

(01:07:55):
So I think you can, if you are taking an extra second to think about it, sort of think about that and then decide whether you're align with it or not. But I think that's an important thing to understand both as a recipient of AI content as well as someone producing it. Because at the end of the day, it's the human that you're that's being assisted, whether it's the audience or the person. And I think you need to think about like you are always the person judging the AI content or rather in control of the content and sort of getting the AI to help you as opposed to wondering if you like what the AI created. The AI is not really creating anything. You know, it has no intent to it. It's like you're creating it.

(01:08:48):
So I think it helps on both sides of the equation to think about it in a couple of layers that you're not necessarily, you know, fooled, although you might be. Again, this is why we care about the transparency. But it's going to take a little while. I think it's going to take a little, a couple of years for AI content and AI, even like AI emails and stuff to be a little more normalized before more people understand that sort of other layer, the human layer on top of it and think of it that way.

Yeah, I agree. That's a good way to look at it. And also just going back to, you know, your point earlier about like, certain kinds of content should probably just not be written by AI. Like, just like when you, those kind of deeper Thought pieces, the, the kinds of things that, you know, that require point of view, that require research thinking to display what you're doing. Like, you know, maybe there's, maybe it's a word count thing, I don't know. But, but I think in many cases just common sense. I can't have AI write a, you know, an, A piece where I want to create an emotional reaction that is aligned with or that is opposed to what I'm trying to say. You know, like I want to evoke a reaction from something. I'm not going to outsource that to AI.

(01:10:12):
I might get AI to help me, you know, clean it up, you know, at the beginning, the end of the process, as you're saying. But when it comes to the core of it or what's important and even within an article or within a newsletter or something that I'm writing, some of it will be just fine, AI generated. But the parts of the newsletter that I know where I want to establish an argument, my point of view and kind of close it out, you know, that's got to be me. And, and I don't think that's ever going to really change, you know, because only I know what's in my head. And that's. If that's a value, if I think that's a value to my audience, that's what I need to do.

Yeah, I mean, honestly, it's like, if it's AI all the way down, what are we even doing here? Well, sure, let's all just have the robots take over.

Then we're being middlemen, right? Then we're just like, okay, I'm gonna just create this automation and just guides you're. You become a manager. You know, that's not. And you know what? That might be a model for some future PR teams where, you know, especially when you, once you have agentic AI really come in and we don't know. I'm, I'm a little Pollyannish about some of this. I'm also doubtful that it is going to, you know, be able to take over everything. I don't think that's true. Like, it's. As with any technology, when something new comes up, new tasks evolve, new roles will be revealed. You still need somebody to kind of oversee things. You'll have the manager role will be different. The manager will be managing a team of people, but also team of agents, AI agents.

(01:11:49):
And how that rolls out is Going to be fascinating to see. I don't know.

Yeah, well, it's like I said, when AI is just giving you the average, there will always be people who want to pay for above average, for sure, and probably cheaper to do it as a human, at least for now. You know, there's ways to coax AI to get better than average, but it still doesn't have the spark. It still can't quite put things together. Maybe that'll change someday. Maybe that's AGI, maybe that's super intelligence. Above my pay grade. But for the foreseeable future, I think humans will always have a lot to offer and there will always be someone willing to pay for it.

I certainly hope so, because that's my future. Sounds like yours too. You and I are doing a lot of the same. A lot of similar things in different ways. And I'm so glad to have met you and to. To really kind of hear more about your thinking. You know, there's so much more we can talk about and there's so much more I wanted to talk about. We're sort of up against it right now. So. So how about we just wrap this up and why don't. First of all, why don't you. You know, were talking about AI, like using AI earlier. Let's just end with something sort of fun and like, what are the three things, the three tools, let's say that you're excited about and that you think people should be using.

(01:13:10):
You know, keeping in mind that, you know, two, three months from now, will they still be tools? I don't know. But like, what are the, like if you had to say here, do these three. These. Use these three tools? Do these three things. What are, what are they?

Yeah, sure.

Let me, let me just kind of narrow that down for PR marketing people.

Right, right.

Yeah.

Okay. So I think there are, there's something. I'm actually pretty excited about that anyone who is involved in content that I think is very helpful. And we're just kind of scratching the surface of. And it's honestly a feature in chat, GPT and Claude. And it's like Canvas.

Love it.

So if, you know, yes, Canvas is really good. If you haven't used it, use it. Claude. It's called Artifacts. And so on the most basic level, this essentially opens your content in another window and then you can sort of do surgery on it as opposed to having to completely rewri the thing every time. It's super useful. It's I think just how it should be done. Right. The back and forth on the chat was always limited. But the even better way to use it is with analysis and telling it to specifically create charts and dashboards for you on the fly for like really specific things. Because I think this is where I think teams are really just starting to scratch the surface of what this could do for you.

(01:14:34):
So you can give it like, I don't know, a bunch of data on how a social media campaign or something and various responses you got on X or LinkedIn or whatever. And if you put all that data in there and just tell it like, okay, create a chart for this on this specific thing and make it customizable with dropdowns and everything. It can do that. It'll just code it for you right then and there. It's pretty. Again, this doesn't, you don't need to know a lick of code. Like you can just go and get to do this. So I think these kind of like highly customized interactive experiences are now super available to everyone. So if you haven't messed around with ChatGPT Canvas or Cloud artifacts do. It's super useful. I'm starting to use it more and more. Okay, getting more specific.

(01:15:19):
There is a new app called Ask News and I believe it's AskNews app. But it is a really cool dashboard essentially for tracking news topics. Now they're public facing product is essentially a news hub. It looks like a homepage of a news organization but all the stories are assemblages essentially. So you can go into a story and it's taken all the stories that it could find on this topic, giving you a summary of it, but also giving you a lot of data around that story. So who's covering it? What are the political leanings of those places and what are the contradictions among these places? So for political stories, again, I don't mean to make it's not explicitly about politics, but you can see how interesting analysis tool like this is for news, for political stories.

(01:16:21):
Because obviously they, there's different perspectives, shall we say, on stories and so they, but they can identify those contradictions and how sort of it's being covered on the right and the left and wherever else. But the best feature they have is the dashboard I talked about earlier where you can actually give it some, a very specific or very general area of coverage. It will go out, find all the news on that and you can sort of track it over Time actually use this for media and AI, which is super nice. And then you can actually converse with the coverage, you can refine it over time. It'll show essentially maps about how the stories are mapped to each other and what terms are trending and stuff. Really, really smart product. And then there's a couple of sort of custom AI solutions that are very similar.

(01:17:13):
So I'll just name check a couple of them. One is called Mind Studio, which I use often and that's a great tool for basically creating custom AI tools for yourself. So if you have sort of reached the limits of working with ChatGPT or sort of one of these more generic chatbots and you want something that not only is more customizable with you know, like repeatable prompts and stuff, but also you want to tie into automations. So that's I think where a lot of the low hanging fruit really becomes a true efficiency, a true ROI. When it's like okay, great, yeah, anyone can get ChatGPT to write social copy, but if you're just copying and pasting back and forth various windows and doing it on a case by case, like what is that really doing for you?

(01:18:02):
What you really want is something with an automatic trigger, then feeds that trigger into a generative system and then produces the output whether it's social copy or whatever and then take that copy and put it where you want it to be. And that's just one very basic example. But that's exactly what mindstudio does. So it's a little bit like Zap here, but with AI as inherently the center of it because that tool has every single AI model imaginable. It's basically a dropdown and you can, you know, experiment whether Claude is the thing you want or Llama or you know, Mistral or whatever and then you know, you're, you obviously are then in charge of the compute costs and stuff but you know, with smaller models and whatever else and depending on your application you might actually end up saving money in the long term.

(01:18:48):
So you can obviously create as many automations as you want. I use them all the time for things like social media, for things like spinning up formal email pitches and whatever else. And it's just a really smart platform for turning AI into what you might think of as your basic level agents. Right, that, yeah, putting tasks and function calling involved. And there's another one that's very similar. I didn't, it's similar to mine, Studio, but it has a little more pre built stuff and is more explicitly for teams. So mine Studio can be for teams but you can also use it as individual. But the sort of teams and sort of advanced version of that with pre made templates is called LVX and that's a. They've just gotten a little bit of funding and this is.

(01:19:32):
These are the guys actually know them a little bit. They are the founders of Parsley. Okay, how do you spell E L V E X. And they have a lot of pre built AIs similar to what I described in Mind Studio. So you don't have to go in and actually like drag and drop all this stuff. By the way mine Studio, it sounds like it would require a lot of coding. It doesn't. I use it, I'm not a coder. But LVX does has sort of an even other even more convenience in that they've pre built a lot of this stuff. Again choose whatever model you want, adjust these templates to how you work.

(01:20:09):
But it's, it's a really smart platform that can get you some really extensive results from just you know, some pretty simple pro, some prompting and some editing of what they already have. So it's pretty cool.

That's terrific. I mean like I never heard of, I haven't heard of LVEX or MindStudio yet and I'm really keen to try these out and I know that many of our listeners will do the same. I would, I would only add to your list that you know, we are very heavy Notebook LM users. I know we talked about it before. I think, I think people really need to get on NotebookLM or find a similar type of application that allows you to, you know, as you said earlier, create a folder and ask of resources and then just have conversations with those resources. It is such a game changer for reversioning, repurposing content and for coming up with some novel brainstorming based on whatever topics that you choose to add sources in for.

(01:21:15):
I mean it's a fairly and so easy to use and I'm sure it's going to evolve and change. We don't even know if it's going to be here three months from now. It is Google after all. They tend to put things up and take them down. But I know, I think it's a keeper.

Yeah, they've got a roadmap. I know that, I know they're planning some upgrades. But notebook LM is an amazing thing to do. The basic thing that you just said, which is, you know, converse and work with some specific content. And then it's good to do because you will run up into the limits of that. Right? And then once you do that for a while, you're like, well, what, if only I could do this, and this. And then you'll find maybe more productive ways to customize things in something like a Mind Studio or an lvx, which can do similar things. So.

Yeah, yeah. And keep in mind, the output from Notebook LM is not the kind of thing that you put out in front of the public. I mean, it's, you know, it's. But it, be, itself becomes a good source that you can then use in whatever other tools that you are using to generate more content or, you know, or just as a reference for yourself. So always excellent, excellent things that I don't, you know, it's funny, I don't even use the, the audio Deep Dive kind of generate podcast part of it. But you know what? It's fun. You, you could and you should, and you know, it's a good gimmick and it's, it could be quite valuable.

(01:22:40):
But even more valuable is when you generate those Deep Dive audios, you download the, grab the, the audio files, and then you drop the audio file into a transcript, or then you take that transcript and use that transcript as a source for something that gets to be pretty valuable. I realize that we're going in this kind of circle, and I hope that gets automated at some point, but I, I, I love doing that. And it's kind of fun to see, you know, how you can find even additional insights that you might not have known before. Assuming it's not a hallucination.

Yeah, I think there's actually another app that I didn't mention that's similar to NotebookLM in this respect and that it creates audio books on topics, but it works a little differently because you can give it specific URLs. So in other words, if you want the POD or, sorry, the audiobook version of, say, a big feature article that you don't want to read the whole way through, it'll do that for you. And again, this isn't just a spoken article. It's essentially adjusting the length and getting the gist of it, giving you something more custom. But you can also search a topic, whether it's like nuclear physics or whatever, and it will go out to essentially create that audiobook summary, and then that's something you can listen to.

(01:23:57):
I've been using it a little bit, it's handy but there is a lot to be said for the notebook LM approach of a conversation which does evoke a little more basically gets your attention a little better.

Sorry Pete, what is that tool called?

Called anytopic. Anytopic IO I think is the website.

Well look on that note anytopic IO now I know that my next 7 days are just now going to be to go vanish down a rabbit holes of anytopic I.O. That's glad, always helpful. I, I, I know we ran a little long and I don't think that, you know, I don't think that we could have that we could stop if we wanted to. I mean it's just there's so much to talk about. I really enjoyed listening to your perspectives. Anybody out there wants to find Pete, please, please do go to Media copilot IO sorry MediaCopilot AI and of course look for Pete Paschel on LinkedIn. His name will be spelled properly in the episode notes and in the title of the episode. Where else are you very active Pete that people can find you?

Well, that's about it. Yeah, I guess I, I'm so bad at X I'm Pete Paschel on X but I, I was lousy at Twitter when it was Twitter and I'm lousy at X when it's X so I'm still on it but I don't do a lot there. I'm on Instagram as Mirror Pete mostly my personal stuff. You want to follow me? But if there is by any chance a Doctor who fan out there at. My other podcast is called Pull to Open. I've been doing it for a number of years now with an old colleague of mine, Chris Taylor at Mashable and we have a lot of fun Going through the whole television run of Doctor who. I was always determined to put my encyclopedic knowledge of the show to good use at some point.

(01:25:56):
Decided to kick off a podcast a number of years ago. So it's super fun if you're a fan. Hope you, hope you like it.

Media Copilot AI Pull to open Podcast. Two very different things. But we'll, you'll get the multifaceted Pete Paschel in the doses that are, are probably digestible in that way and you know, anybody please go out there and reach out to Pete and read Media Copilot if you are in our industry and even if you're not, because it's brilliant writing and, you know, I'm jealous. It's like, it's like the newsletter that I want to be when I grow up kind of a thing. It is. You know, it's just. But, Pete, thank you so much for joining me. And you know, I will welcome you back here with open mic and open arms anytime.

Thanks so much, Dan. It was super fun. Thank you so much for the incredibly thoughtful conversation.

You got it. Thanks for taking the time to listen in on today's conversation. If you enjoyed it, please be sure to subscribe through the podcast player of your choice. Share with your friends and colleagues and leave me a review. Five stars would be preferred, but it's up to you. Do you have ideas for future guests or you want to be on the show? Let me know@dantrendingcommunicator.com thanks again for listening to the trending Communicator.

All Episodes

Reinventing Newsrooms: AI as Friend, Not Foe - with Pete Pachal

Episode Transcript

Popular Podcasts

On Purpose with Jay Shetty

Crime Junkie

Ridiculous History

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}Reinventing Newsrooms: AI as Friend, Not Foe - with Pete Pachal