The Monkey Patching Podcast: Going Bananas on AI, Data, LLMs & Tech

Murilo (00:08)
Hi everyone and welcome to the Monkey Patching Podcast where we go bananas about all things AI Slop, big deals and more. My name is Ruiolo, I'm joined by my today and always by my friend Bart. Hey Bart. I'm good, how are you?

Bart (00:21)
Hey, Miguel. How are you?

I'm good as well, just back from a week of family vacation, the autumn break. So yeah, I'm happy.

Murilo (00:34)
Cool,

cool, cool, cool. Recharged, ready for another week of use. Actually, I did say I'm good, but I am feeling a bit sick.

Bart (00:43)
Yeah, was thinking, does he want to go into it or not?

Murilo (00:47)
It was just my default. I'm I'm good. But no, I'm feeling a bit sick, but I'll survive. I'll survive.

Bart (00:52)
Yeah, we might keep it a

little bit shorter than usual, right?

Murilo (00:55)
just a tad.

We actually had quite a lot of articles this week. ⁓

Bart (01:01)
Yeah, actually

had, it was a it was a news heavy week. ⁓ But we're only gonna cover like six today, but we can actually add the other ones and the summaries to the newsletter. We'll do that.

Murilo (01:06)
Yeah.

That is a good point. Yes indeed. So maybe before we dive in, if we do have a newsletter, if you haven't subscribed already, it's a... How to candy subscribe Bart?

Bart (01:25)
you need to go to newsletter.monkeypatching.io and just leave your email address.

Murilo (01:32)
That's it, as simple as that. And then you get weekly news, weekly email, basically with the content that we covered on the show, some extra links, maybe something new in the future. So if you wanna stay up to date, just subscribe there. But without further ado, what do we have part?

Bart (01:52)
We have Google Labs and DeepMind launched Pommelie, an AI tool that builds a business DNA from a company's website to auto-generate on-brand social campaigns. It's rolling out as a public beta in the US, Canada, Australia and New Zealand, promising speed but raising some questions about originality and disclosure. ⁓

Murilo (02:12)
Cool, so what is this?

Bart (02:19)
And it's basically a tool called Pommelie and it looks at the website of your company and based on that it builds this, they call it a business DNA, a like a branding style guide for your company based on the assets that you already have. And then based on that style guide you can basically create new...

Digital creatives like some simple things like a flyer stuff you would want to promote to like image that you want to share on Instagram like these kind of things.

Murilo (02:53)
So there's two things. First, it says build, but you need a website. And then from the website, they'll create, how do call it? There's a name for this,

Bart (03:01)
Well, what I understand, I haven't tried it. It's a website

or like other type of assets that you have, right? Like you've maybe PDFs or whatever like that, that are really like embody your style as a company. Like what is your branding style, right?

Murilo (03:16)
So it's not like you'll create something from nothing, right? You have to have something and then it expects from that. I think there was a name for that, no? Like a marketing portfolio or something. So basically like what kind of fonts, what kind of colors, how is the text usually written.

Bart (03:19)
Yeah, exactly.

Yeah, but like a brand's guide.

Murilo (03:30)
Brand guide. That's how you mentioned to me once. So it extracts that and then once, I guess you can always verify that, right? And then once you have that, then you can generate campaign ideas or like actually generate campaign ideas. So you can, what I saw in the promo video is that you can bring your own. So it's like, I want to do this campaign, so you can do that. But also if you don't have ideas, they can generate ideas for you and then you can iterate from that, right? You say, okay, change the color of this, do this, do that.

Bart (03:50)
Exactly.

Yep.

Exactly,

Murilo (03:59)
So it's not available everywhere yet. So even if you wanted, you couldn't try, but yeah, this is actually, I saw this from Google Labs. So they call it an AI experiment. So actually, notebook LM, I think, also came from Google Labs, I wanna say.

Bart (04:14)
Hmm.

Lydia.

Murilo (04:17)
Even though

it's called an experiment, I there is a lot of, there could be a lot of business value, right?

Bart (04:23)
Yeah, this is what we've seen. It's already. There's also to me like a bit of a downside to this. I like a bit of question on how authentic is everything that you're going to see when it comes to like creative output of companies, especially in this case, when it comes to marketing. But we can discuss it, but I also at the same time, I have the feeling like the ship has sailed like six months ago.

Murilo (04:52)
Yeah, I think again, it's about efficiency, right? Maybe you can be more efficient, you lose a bit of the authenticity. I think it's a bit... I feel like you will converge to something, right? If enough people use it, you will start to be able to see patterns, like the same thing we see with JGPT and all these things, right? How like the wording is very specific, maybe there's gonna be fonts or colors, there'll probably be a bias, right, in these models.

Would you use this as if you had like if you had a small business would you use something like this?

Bart (05:26)
If it's good, I wouldn't know why not, right? Because I think that one of the challenges when you do want to do these kinds of things today with Gemini or Chess GPT, but like with image generation, basically, like it's very hard to like give clear instructions on this is my style, like this is my brand guide. It's very hard like in this tool, specifically does that. Like it locks down your style guide and like allows you to do things within that set of parameters, which I think is viable.

Murilo (05:49)
Yeah.

Bart (05:55)
I think also long term, it's what I meant a bit, the ship has sailed. The reality will be that everybody that is doing something in creative marketing will to some extent use these type of tools. And I think the difference again is like...

Murilo (06:07)
Yeah.

Bart (06:11)
mainly like the skills that they apply to that, like how, what is the creativity that comes from these people that apply these tools? Like maybe in that case, can compare it to some extent with, with like going from a pen and paper to Photoshop, right? And this is like a next step in reality will be that like 90 % of people will use this. Not necessarily Pommelie, but Pommelie related tools.

Murilo (06:25)
So I think it's like...

So it's almost like, it's like, are you gonna use it as a tool or are you gonna almost replace your whole marketing team with something like this, right? And I think.

Bart (06:43)
Yeah, maybe that's a good way to approach this. And I'm in favor of using it as a tool. Like, still have your own idea and feeling and strategy and creativity on what the outcome is, right? Not just try to let it do autopilot for you.

Murilo (06:48)
Yeah.

Yeah.

Exactly, yeah, no, I'm with you on that. I also, I think all these things are tools and I good tools make you more efficient as well, so it's not so not to undermine, right? Not to say it's just tools, it's a very powerful tool, but I still think that you need someone to really curate what are the things you wanna do, you wanna say, et cetera, et cetera, right?

So for example, LinkedIn posts, I think, is very easy, right? You see a lot of Gen.AI content. And it's easy to see, like, okay, this person is using Gen.AI. And I think a good...

Bart (07:33)
Well, it's easy

to see if people are not very skilled in using it.

Murilo (07:36)
Yeah.

Exactly, that's the thing. So it's like if people are skilled, they curate the stuff, they use it as a tool to speed up their efficiency, then I think it goes a bit unnoticed and I think you can be very productive and have a lot of very interesting stuff. But you see a lot of people with like huge posts with like a whole bunch of links and a whole bunch of like bullet points and emojis and doing this and the language that they use, right?

Bart (07:50)
Hmm.

And it's very like,

it's visibly clear. Like this is Yennai, right? Like something, like there's a part that is in bold and there are, there are some itemized stuff and like, there's like, yeah.

Murilo (08:04)
Yeah.

Yeah, exactly. There's like some emojis in there and there's always everything

is like, yeah, indeed. And I think when that happens as well, I feel like maybe lose a bit of credibility as well, right? It's like... ⁓

Bart (08:22)
Yeah, I fully

agree. Like I don't take it seriously anymore. When you have people that like used to post once every six months and now suddenly they are posting some bullshit message every day, trying to be an influencer. mean, that's, it's not very authentic, right?

Murilo (08:31)
Yeah, exactly.

Yeah,

yeah, indeed it's not. I think I've also seen some, I think this is more recently actually, it was like ⁓ a inspiring message that has nothing to do with tech, or like it was more about personal development. And then there's like five images, GNI images clearly, with just some subtitle.

And then like a huge message on the text, right? It's just kind of saying like, small habits are the ones that make the biggest impact, don't give up. And it always starts with like, I was unemployed for five months and then this and this, and this is how I turned my life around, you know? And then I actually saw like two or three posts from the same person, exactly the same, and I was like, come on, you know? Like, it's a bit, yeah.

Bart (09:16)
Yeah, that's the inevitable rise of AI slop.

Murilo (09:20)
Yeah, nice show, nice show. Alright, what's next? Next we have the...

Wall Street Journal argues that Anthropix's enterprise-first focus is paying off as corporate customers drive the bulk of its revenue and adoption. A reported multi-billion run rate and strong coding share sharpened the contrast with OpenAI's mass market push, setting up a strategic showdown over who monetizes AI best.

So I came across this article, I thought it was interesting, because they basically saying, they're comparing a bit the two different strategies, right? Like OpenAI, according to this article, it's more mass-based, right? So it's more for the everyday person. And what they say is that Entropic is more of a enterprise business approach, right? Even though they are less flashy, in the sense that they're less on the news, they're less this, that.

They actually brought some, they show some numbers that actually I think I saw that entropic, the revenue, actually they do say that the revenue per user is actually more than opening eyes, right?

Bart (10:32)
It is rumored, it's difficult to get a good view on what the revenue of OpenAI is, but indeed, yeah.

Murilo (10:36)
Yes,

indeed, it's rumored.

Bart (10:40)
So

maybe to give you a bit more numbers there, like, Anthropic is expecting,

9 billion revenue by end of 2025. OpenAI today is rumored, but some altman has recently denied it, to do 13 billion in revenue this year. But let's even say let's... 9.

Murilo (11:01)
So how much is in traffic again?

Nine and 13, you said.

Bart (11:07)
But

Sam Oldman recently said in interview that it's like way under, it's way more, but it's not really clear what it is. But even let's say it's going to be not 13, but like double 26, right? Like if you would say then nine of Entropic versus 26 of OpenAI, if you then look at the active users, so OpenAI reports around 800 million weekly active users.

Murilo (11:15)
Yeah.

Yeah.

Yes.

Bart (11:37)
Anthropic, basically with Claude, they have 19 million active users. Which is a fraction, right?

Murilo (11:46)
It is a fraction. Yeah, yeah, yeah. So I think.

Bart (11:48)
So,

if you look at those key metrics like the revenue per user on foreign traffic, it's way better,

Murilo (11:58)
yeah, for sure, way better. But I think there's also like who's using OpenAI, right? And maybe it's more a lot of people kind of using, and that's also what they talk here, right? Like that a lot of people are using OpenAI kind of as a Google replacement. So they're just doing search. It's free users, right? For OpenAI to really end up ahead or to make money from this, right?

Bart (12:12)
Mm-hmm.

Murilo (12:22)
would they do ads, right? Also then they are competing with Google as well, if they're doing a Google replacement, right, for ads. So it's not an easy...

challenge, right? There are other things that I, that's why, okay, I was looking, was listening about two strategies, right, more high level and I was a bit like, okay, how does this translate to concrete things? And then one thing they mentioned towards the end of the article, I guess, is like that opening eyes is moving more towards a hands-off approach in terms of what they censor, what kind of conversations and stuff, and not specializing models and doing things that are more quote unquote edgy or for fun, right? Like generating images and all these things while Entropic is really focusing on coding.

Bart (12:55)
Mm-hmm.

Murilo (13:00)
It's focusing on like I think they did there was a study that they mentioned about Legal documents as well. So something like this is very very specific for enterprise, right? They're really specializing their models towards this which does I mean also Microsoft which was Historically quoted I'm saying historically but it's like a couple years, right? They would be very tied to open AI and now they are looking to Use more of entropic models as well think or pilot

Bart (13:30)
Yeah, my, this is just good feeling here. Is that, that I think like this difference, like you, I don't necessarily think it means that one of them, like if I need to bet on who will win the AI race, because I think there will be one of two very big ones and like then maybe five small ones and then that's gonna be it in terms of land providers. I would still put my money on OpenAI because they have their hands in so many things.

Murilo (13:50)
Yeah.

Bart (13:55)
And I think, I think what the Wall Street Journal here is saying is that Antropic is really focusing on the enterprise, which is probably true that they have a more focused aim at that, at corporate clients. But you can't say that OpenAI does not have it. But I think the only thing that where Antropic is basically like it rivals everybody is, is code generation. And it has, it has rivaled code generation for them. I don't know, last two years, 20 years and a half.

Murilo (14:25)
Yeah.

Bart (14:25)
And that's

where it is today very good and like if you want to have the state of the art in code generation you use use Claude basically. And I think that is where they are at.

Murilo (14:33)
Yeah, no,

I agree, but I was also questioning, are we in a bubble, and I, right? Maybe if we were lawyers and we were analyzing legal documents, we would say, yeah, but in topic is much better, hands down.

Bart (14:46)
But the IAM, and again, that's why I saying, good feeling, I don't agree. Like I use it a lot to get extract numbers and stuff like this from annual accounts and stuff. I feel that OpenAI in thinking mode is way better than Claude has today, for example.

Murilo (14:59)
Okay,

I see. Yeah, okay. Interesting.

Bart (15:03)
But I don't think that you necessarily have one model clearly surpassing the other. But you do have it in coding, right?

Murilo (15:10)
except for coding.

Yeah, yeah, yeah. No, I haven't used as much of the other ones, but I do think, again, I also think that OpenAI has a lot of momentum, right? They are, I mean, everyone knows OpenAI, everyone uses it, like I said, like customer base, like the weekly users is...

much bigger, right? So I also feel like if they lose the AI race, it's because it's not gonna be like one mistake, right? It's something that is gonna be like over time, like more series of consecutive mistakes for them to really fall behind because I agree with you that they are, I do think they are ahead, right?

Bart (15:50)
But I think OpenAI just serves a much broader market. And you can also like at the same time, you can call that like they're not very focused. So that's pros and cons, right?

Murilo (16:01)
Yeah, but I think, for example, you see they do a lot of stuff on commerce, Like deals with Stripe and stuff. They also have their own servers, right? The data servers that they're investing on as well. So I feel like it feels like they are investing more and I think if you say you wanna mitigate risk, right? If you invest in a lot of different things, it's also a healthier.

I think, right, if one of them goes wrong, it's okay, quote unquote, right? And I think the main thing, which is the mind share of people, I do think, I mean, my impression is that OPINAI is ahead, right? If I talk to my parents, which are not technical at all, they definitely know CHPT. I don't know if they know Anthropic. If I say Claude, I don't know if they do, So.

Bart (16:45)
Yeah, I think that is a fair point. That's a fair point.

And I think that maybe that's a bit what I'm questioning here in this article is like a bit like the chicken or the egg. is it true that Claude is focusing way more on enterprise clients or is it simply true that two years ago they became by far the best at GenAI coding and thus are today more relevant for corporate clients?

Murilo (17:12)
That's true, maybe AI for coding is a very clear, I mean, because also AI, the development of AI is closer to coders as well. Maybe that's also the first very clear productivity boost, right? But I do think there will be other productivity boosts and the Tropic will not be the better one, as I said earlier, So.

Bart (17:28)
I agree with you.

I guess.

Murilo (17:35)
Time will tell. We'll also, I think we'll talk a bit about this, or I'm not sure if this is like a different job than we had over the years. But we'll see, we'll see. will tell, time will tell. More on OpenAI, what do we have apart?

Bart (17:49)
OpenAI is restructuring into a public benefit corporation with Microsoft retaining a 27 % share, clearing the way for bigger capital raises and potential IPO plans. The PACT resets rights and revenue sharing through 2032, signaling massive data center ambitions and tighter alignment amid escalating AI costs.

Well yeah, it is restructuring into a public benefit corporation, ⁓ more or less a for-profit. I think... ⁓

Murilo (18:18)
Do you know the differences?

Like what's the difference between a for-profit and a whatever they call the year?

Bart (18:25)
Well, that's an interesting one as well. It comes after, I want to say, six months of rumors around this, right? That they wanted to transition basically to a for-profit organization because it would make their life a bit easier, I guess. it becomes easier to raise capital. So this news broke last week. To be honest with me, it didn't really come as a surprise. think the only surprise is that it took that long.

⁓ But it is not a for-profit, it is a public benefit corporation which is basically something like a... ⁓ should I put this? Try to find the right English words. Like a non-profit. So a non-profit typically has a certain ⁓ mission for the greater good. So this mission for the greater good needs to be there. But it is a for-profit.

Murilo (18:47)
Hmm

You

and non-profit.

Bart (19:12)
Something like that doesn't exist in Belgium. I'm not aware it exists in any of European countries today. But it is basically a...

Murilo (19:17)
So.

So basically like a non-profit is just a mission for the greater good, but the goal is not to, I think we talked about sometime about this. Basically there's no, the company doesn't profit, right? Like everything is either reinvested or other things or paid employees or something. Actual non-profit, yeah, yeah, But you also have the mission for.

Bart (19:32)
in an actual non-profit.

Profits are

reinvested. That's how it should look. They're not paid to, like there are no shareholders to which profits are paid out to in a non-profit.

Murilo (19:39)
reinvested, yeah, yeah,

Yeah, okay.

And there's also a mission for the greater good.

Bart (19:49)
Typically.

Murilo (19:51)
And then for profit, so the other quote unquote extreme is just like you don't have a ⁓ quote unquote mission for the greater good. You can just try to basically try to put money on people's pockets, which is fine.

Bart (20:02)
your let's

be honest your mission in a for-profit is to create value for shareholders however you do it

Murilo (20:07)
Yeah, okay.

And then this one is a bit in between that you're trying to create profit for your shareholders but align with your mission, which should be something for society. Okay.

Bart (20:19)
Yeah, that's how I understand

it. And I'm paraphrasing here. The mission that they're stating here is that they want to develop AGI for the greater good of the global population. That's more or less our mission.

Murilo (20:41)
Which is also a bit of a... I don't know, AGI for the greater good is something you don't hear often, right? Usually when people talk about AGI, it's a bit with fear, with... It's gonna replace jobs, gonna replace people, it's dangerous, it's gonna try to, I don't know... Okay.

Bart (20:55)
True. Yeah, true.

I mean, you can think about the mission, but what you think, but maybe an interesting point in that is that, well, they basically become for-profit, which there's also a lot of rumors about a potential IPO in the coming years. An IPO is basically they go on the stock exchange and they become publicly traded. Most likely at some point they have to do it. It depends a bit on like...

Murilo (21:10)
Okay, and maybe for people, yeah.

Bart (21:19)
What is the profitability they can reach in the coming years themselves, the revenue they can get versus like how quickly do they need to raise more money, et cetera. The challenge of the public benefit corporation of going to an IPO is that you need to basically keep aligning to that mission that you had when you go on the socket change, when you become publicly traded. And it's not always easy to match that with

Murilo (21:43)
Okay.

Bart (21:47)
shareholders have to have a lot of expectations.

Murilo (21:51)
Yeah,

I see. Do you think this could be a first, well, I guess if they were going for profit, they would probably just call it for profit, It doesn't make sense to...

Bart (22:00)
No, I think it's very difficult.

I think they couldn't. They couldn't. This was the only option to go sort of for profit just because of how it was originally structured. And they got a lot of pushback from us from Elon Musk in the restructuring. was a lawsuit. I think this was the best they could get without becoming an actual for profit.

Murilo (22:04)
Okay. ⁓

Okay.

Okay.

Yeah, and then by doing that also there's a bit of restructuring that I read somewhere that Microsoft, like they're still, they have a 24 % stake, right? So they still own, actually it's not via directly. So there's chat GPT, not chat GPT, OpenAI non-profit that org that controls or that owns the quote unquote for-profit, I forgot the name.

Bart (22:27)
Yep.

Yeah.

So, Microsoft had an important stake in old OpenAI. In the new structure, they will have 27%. There is still this clause that exists ⁓ that from the moment that they will reach AGI, up to an external board, more or less, to decide when that is reached, that hard ties with Microsoft are cut.

then they are completely free. And those hard ties with Microsoft basically mean that ⁓ today and the coming years, they have to give 20 % of their yearly revenue to Microsoft.

Murilo (23:19)
Damn. So okay, I'm gonna repeat what you said to make sure I follow, because this seems like a biggie. They have to give 20 % of their profit to Microsoft, or have to me, okay? Yearly revenue. Until they reach their mission, which is about AGI, and who decides if they reach their goal or not is an external board.

Bart (23:21)
the significance.

Yeah.

yearly yearly revenue, not profit yearly revenue, which is even more.

Murilo (23:45)
And then when that happens, so OpenAI has 27%, but do they take that as the evaluation at that point as an exit or?

Bart (23:55)
No, from that moment onwards,

Microsoft will probably still own the 27 % shares, but they do not have to pay the 20 % revenue to Microsoft anymore.

Murilo (24:04)
Tim.

Bart (24:05)
Microsoft did a good deal a few years ago.

Murilo (24:07)
Yeah, I was gonna say like it sounds like someone had opened their eyes and opened their eyes fucked up. Like, Jesus.

Bart (24:10)
But also

in this new restructuring, they are already less dependent on Microsoft, so they just announced a big new deal with AWS for compute, where before they had to rely or had to go first through Microsoft to see if they could contract Microsoft for extra compute and only then go to third parties, so they have a bit more wiggle room there. ⁓

Murilo (24:33)
And there's also

something with, because OpenAI, they are also investing in data centers and all these things, right? And I think now there is no, it's OpenAI and Microsoft doesn't have any say or any stake in that, right? It's OpenAI owned.

Bart (24:47)
Well, to

do these big investments in, for example, in data centers, they have to raise a lot of capital. And before this reorganization, Microsoft basically could more or less veto whether or not it was going to happen or not. So now they have much more autonomy and say in their own decisions.

Murilo (25:04)
See you.

But they still have like board seats, I guess, for the parent group and all these things.

Bart (25:09)
Yeah, I assume that Microsoft is still on the board. And it's only

because the nonprofit organization will keep existing and the boards will be the same. It will have the same board. What exactly the nonprofit is going to do, I don't know, but it will keep existing.

Murilo (25:22)
Yep. Yep.

Yeah, yeah, it's like, I remember I saw some explainer video kind of, it's like you have the non-profit that owns the for-profit or what today is the not-for-profit, but you know what I'm saying. And how everything plays together and all the moving parts, it seemed very complex, yeah. And yeah, like Microsoft shares rose again and I think it's $4 trillion. More big numbers.

Bart (25:45)
Yeah, it's complex,

Murilo (25:58)
for big numbers. All right. Next we have AWOL. Is that how you say it? AWOL? A-W-L, okay, AOL. Let's go with that. AOL is changing hands again with Bending Spoons agreeing to buy it for about $1.5 billion.

Bart (26:08)
Ewel.

Murilo (26:20)
as the Italian firm keeps scooping up legacy internet properties. The deal comes with $2.8 billion in new debt financing and claims of 30 million monthly active users.

Bart (26:31)
you been an AOL user in the past?

Murilo (26:32)
didn't even know.

Bart (26:32)
Do you still remember

the sound? If you don't remember the sound, then you're too young for this.

Murilo (26:36)
Is it

Bart (26:37)
That was also part of it, but it's like, you've got mail.

Murilo (26:39)
yeah, you've got, I was like, you've got nail.

Bart (26:41)
Yeah, but that's the one that's from AOL.

Murilo (26:43)
you had a dial-up internet, no?

Bart (26:44)
Yeah,

Murilo (26:45)
I remember the things of like if someone's on the phone, the internet's not working, and like all these other things. I didn't know it was Italian either, so.

Bart (26:52)
No, is not. Bending Spoons, which is sold to us, is Italian. Bending Spoons is like basically this sort of this software conglomerate, their existence in 2013, and they basically buy like these mature consumer internet brands. And like typically they do a lot of cost cutting to really bring it down to basics and then start refactoring to make it more profitable. think there lot of people that are very negative about this as well.

Murilo (26:55)
⁓ Ben is supposed to...

Bart (27:22)
Like you have for example examples of this are Evernote for example or Meetup.com that are, I think both are still owned by Benning Spoons. And ⁓ well, I think you can argue like Meetup is a good example. I think they've been neglected for a while and there's clearly been a very limited investment in them. But at the same time you can argue like it still exists and like they got a like place to live and.

it seems to be, well, there seem to be running in a still big community. mean, pros and cons to this approach, guess, haters and lovers.

Murilo (27:56)
I guess

if you say that it wouldn't exist if they hadn't acquired, I think people would be like, okay, fine. I think it's a bit like what if, right? I think hindsight is a bit, if there was no acquisition, but people are still working on this and trying to make it, then maybe.

Maybe people wouldn't be so happy, but yeah, it's to start to know what would happen. But I looked here up quickly. Evernote, Meetup, there's also Remini and Komoot, which I don't know, but Streamyard and WeTransfer actually is also from them.

Bart (28:26)
Yeah, and it's a very, it's a very typical like the commut is in the outdoor community like routing and like you can find like, like navigation instructions, stuff like this. were, they want to say acquired two years ago, wheat transfer was acquired and like their typical play is that like for wheat transfer, for example, like they've signaled that they're gonna cut like 75 % of the staff really do.

really make it super lean and that basically means we're not going to invest super heavily anymore in new features. This is what it is today and we're going to really try to market the current set and then only grow with what we really need to invest in, what are really critical things. ⁓

Murilo (29:07)
Yeah, so basically

it's like bending spoons owning it is really just kind of saying we're just going to stabilize just to the bare minimum to keep things afloat. We're not a growy face.

Bart (29:16)
Well, I would say that

that is the pessimistic view on it. I would say that. I think the optimistic view that they probably themselves would say is that they are an evergreen holding where something like Evernote can live for a very long time and they will invest when it is needed. But they will ⁓ cut all the things that are inefficient in the original organization. That is maybe the more optimistic view.

Murilo (29:20)
Okay, what's the optimistic view?

Okay. And evergreen holding,

yeah, evergreen holding is like when a group basically acquires these companies without the intent to sell and without intent to necessarily change how the company is working. Is that a good description of it?

Bart (29:58)
Without the exit. think that is more or less correct. Yeah. Yeah. I think not necessarily that they don't want to change anything to the company because they clearly do it. Like they want to make the company super efficient, super lean. Right. But they don't want to do what they value is the, is the brand and the community around that brand.

Murilo (30:07)
Yeah.

Yeah, I think that's

what I was thinking. They don't wanna change the brand or change the, the stuff that works, they don't wanna change. I don't know how it is working for people, but maybe the culture, the brand, the atmosphere, the public image, they don't want to change that. Don't know about this.

Bart (30:22)
Exactly, yeah.

Yeah, exactly.

So yeah, they bought a well. And I looked into a bit like what because I didn't know he was still around. And I looked a bit what it is. And so if the a well portal, which is a bit like like you have some news, you have some finance information, you have a better information of entertainment, so basically like more or less a new site, you have some desktop applications, there are there are some security stuff like ID protection, data security.

But I think all of that is like peanuts. What it comes down to what they have is that they have 30 million active email users still. So you have basically still a lot of people with an AOL email address.

Murilo (31:09)
What?

a lot of people with, you've got nil.

Bart (31:15)
Yeah,

and probably and I think reading that I also understand like this is actually interesting because probably a lot of people that still today have an ⁓ email address they're like, it's very hard to switch. But I probably should switch because like, how long will the well still live? Alright, you're probably

Murilo (31:30)
Maybe a question as well.

Maybe this is a bit of a dark, but how many of these people are alive?

Bart (31:36)
Well, they're 30 million active users. Active users. Yeah, yeah. Are you calling people that use AOL old now?

Murilo (31:38)
active users. so for sure this is... ah, okay. Okay. Okay, okay, okay, okay, okay. No, no, no, not

at all, not at all. was just, I don't know what I thinking actually. Okay.

Bart (31:48)
But

that basically means that AOL is still very big email provider, right? And I think it would be interesting to have a big modern revamped email provider. Because today, you either use something for free and then you're the product, right? Like you use Gmail or you use Live or something like that, or Outlook. Or you go for something like Proton or...

Murilo (31:52)
Yeah, yeah, yeah, for sure.

Bart (32:14)
But there are not a lot of them. It's hard to run an email service, right?

Murilo (32:21)
It could be cool.

Bart (32:21)
But actually AOL

is still a very big one, they're not, I don't think they're really, like people don't think of them like, I mean, Neemel sort of like, let's check out AOL. Like I think there's still a lot of value there. if you like freshen this up a bit, bring it up to 2025 standards, really reposition this in the market.

Murilo (32:35)
Yeah. Yeah.

Yeah, I think it's an interesting brand, let's say, in the sense that people don't think of it now, but if you mention it, everyone knows what it is. So it's like, if you say AOL, you don't have to explain what it is and what it does, and I think there's a lot of people that probably have fond memories or something like this. It became a bit of an icon, so it's new in the sense that you can reinvent, but it's not new in the sense that you already been ahead on the...

Bart (32:48)
Exactly. that's a point.

Yeah.

Murilo (33:06)
the acknowledgement from people, the brand and all these things, right? So it could be, it's true. think there is an opportunity there.

Bart (33:12)
Yeah.

Of course it goes a bit like it's a bit, um, it doesn't really match with the typical playbook of bending spoons, right? Like where they cut a lot of stuff and say, is just what we have like to reposition it. Like it really has to be like revamped and build up the speed. like, I mean, let's, let's see a year from now if more people around us know of AOL.

Murilo (33:30)
Yeah.

Let's see. Yeah,

they also mentioned they have a very loyal customer or user base, right?

Bart (33:42)
But is

it? wonder, to me, like, AOL was huge, right? Like, AOL was huge in the US. I I lived in the Netherlands, even we at AOL, like everybody in the world had AOL at some point. And I think a lot of people, I mean, the only reason I still have a Gmail address today is because it's fucking hard to say to everybody, I've got a new email address, please contact me on that new email. Like, it's super hard, like. ⁓

Murilo (33:46)
Yeah.

Yeah.

Yeah.

Yeah.

Bart (34:06)
It has nothing to do with my loyalty, it has to do with my laziness.

Murilo (34:09)
Yeah, it's true. It's very sticky, right? Like everything's set up there. Like what about all the the accounts that you created using that email? What about this? What about that? Yeah. When I was a kid, this is a bit of a side story. I had some some emails that had, I was trying to be edgy, but I was a bit like a bit rebellious, right? So I had some some emails that they weren't, there was no profanity, right? But it was a professional and

Bart (34:12)
It's super sticky here.

Murilo (34:36)
I remember at point some people were like, you should create a new email because like if you start going to university and start doing these things that you need to actually use serious email addresses, you cannot use this obviously, right? So I created a new email and actually I'm super happy that I did because imagine you just, just, if I was more stubborn, I had just stuck with it. And now I have like all these things, all these accounts and all these places that like.

Bart (34:57)
Uh-huh.

Murilo (34:58)
can't go back. I think I was like young, right? So it's like, I didn't have a lot of stuff there.

Bart (35:02)
Yeah, yeah, I the same thing when

I was young. Yeah, I also had like my gamer name I had in my email address. ⁓

Murilo (35:07)
your game name. Trying

to be edgy, you know, like, find your identity and all this thing.

Bart (35:10)
I still have it today actually, like in my Gmail address.

have my... I didn't think about it at that point. Like my name, it's very common name in the Netherlands. So it was already taken. So I just put like the last two numbers of my birthday here behind it. And now I feel a bit weird about it.

Murilo (35:19)
Yeah. Yeah.

Okay, and that worked.

I thought that's okay. I thought they're still okay, right?

Bart (35:31)
Right?

That you can still pull off, shouldn't go further than that.

Murilo (35:37)
Yeah, but like

imagine if you put the last year of your girlfriend back then, you know, like, I don't know. That could be worse, that's what saying, right? Like, I like the birth year is okay, but there's some things that... But actually Bart and then the two digits, that was enough for you.

Bart (35:44)
yeah, that was bit awkward, yeah. It could be worse, yeah.

No, it's Bart and my last name, Smith. Not too much.

Murilo (35:59)
Okay,

okay, okay. I have a very common name in Brazil. First name, I mean, they're not that common, but it's common enough that creating a new email address, especially these days, I need to be creative about it.

Bart (36:12)
Well, I knew that there were other Barthesmaids alive, only a few weeks ago I checked it on LinkedIn. You can type in the exact name, it's like pages of people with my exact first name and last name, pages and pages and pages. It's crazy. Any feeling of any uniqueness just went out the window.

Murilo (36:24)
Yeah.

Yeah,

exactly. nowadays to be creative. think finding a good email address is not easy these days. yeah. What else do we have? Your turn now.

Bart (36:42)
No, no, it's not easy.

I can take it. Wharton's latest adoption study says generative AI has moved from pilots to measured ROI with usage now weekly for most leaders and daily for many. Standout stats like 72 % tracking ROI and three quarters see positive returns. They set the stage for 2026's push from experimentation to performance at scale.

A 2025 report by Wharton that goes a bit in depth on the use of Gen.AI in enterprises with much more positive results than the study we saw a few months ago by MIT. Thank you.

Murilo (37:24)
MIT.

Bart (37:29)
so I checked a bit in,

Like what is the audience, who did they survey?

they interviewed 800 US enterprise decision makers between June 26 and July 11th in this year. The previous year, so there are also some trends there. So in the report previous year, it was also 800. So that's 2024. In 2023, they interviewed 670. So quite significant numbers, right?

Murilo (38:01)
Yeah.

Bart (38:02)
They're all decision makers from US commercial organizations with more than 1000 employees and at the very least more than 50 million in revenue. So relatively big. Very varied like HR, IT, legal, marketing, sales, operations, engineering, procurement, blah, blah, blah. A lot of different domains. What is also interesting that they look at different tiers for the results. you have...

Tier 1 is companies that are at size 2 billion revenue and more. Tier 2 is 250 million to 2 billion. Tier 3 is 250 million and lower. So you also have some interesting differences between the size of the companies. And the main finding, which is already important one, that tracking of ROI on these projects, on these YANY projects is...

more or less standard. Like 72 % of these enterprises say that they formally measure the ROI on their generic projects. Which is a lot,

Murilo (39:00)
which is maturity, post-maturity.

which I guess,

yeah, yeah. And it shows that AI is really becoming more mature, right?

Bart (39:16)
Well, I would argue that the next result is because 74 % already see positive ROI overall.

Murilo (39:28)
70%, that's bigger.

Bart (39:29)
74 %

of these companies. That's a lot, That is like to me the complete opposite than the study from MIT, where it was like 5%.

Murilo (39:32)
Okay.

Yeah, maybe question

also. On the study from MIT, did they also have a multi-year analysis of this? Or was it just a, yeah, I don't remember either, but yeah. Well yeah, indeed.

Bart (39:47)
I don't know to be honest.

They also make some like 74 % seen positive ROI overall. They make some comparisons between sectors. So you see digital heavy sectors, they show higher positives. So for example, tech and telecom 88 % positive ROI, banking, finance and professional service 83%. But like manufacturing and retail, they are respectively 75 % or 54%. Meaning that there's definitely an ROI manufacturing, but retail with 54 % like it's still a bit like...

We still need to prove this stage. Also interesting, tier two and tier three, so that is the small to mid-sized companies, they skew to be more positive in ROI results. Tier one, those are the very big companies. They're more stating like, it's still early to tell whether or not we're gonna get a positive ROI from this.

Murilo (40:47)
So the smaller companies are benefiting more, like at least concretely. That's what I understand.

Bart (40:55)
That's also how I understand it. This...

feels intuitive to me just from my experience that in like the still very young technology and very very big companies like the tier that they're defining here is 2 billion revenue more is very big like they typically also move very slow

So I don't think it feels intuitive that these companies like will lag behind in reporting a positive ROI.

Murilo (41:23)
But because they said also they're not sure, but I guess it's more, I know they just say it's too early, right? But okay. But then the expectation would be that the ROI will come, but it's just that these enterprises are bigger, so it takes time for the ROI to catch up. Yeah. The intuition. that's correct, then we should see the ROI for the tier one enterprises coming up in the next years. Yeah.

Bart (41:38)
Exactly, well that's the the vision behind it.

Exactly.

Murilo (41:51)
Indeed, it's just really, well, way more positive than the MIT report. I think also they measure about how much the leaders, right, they believe in GEN.AI as well. So it's not just the actual ROI, but also how much they want to, how much trust there is, quote unquote, which is also high, right?

Bart (41:57)
Definitely.

Yeah. And you could, you could be skeptical on these results as well, because you're, you're interviewing leaders in these respective fields and they're in their respective companies. They are probably leading, leading a lot of these projects. Like it's also their head on the line, right? Like they also have a lot of incentives to say like, it's going very well. That's creating a lot of ROI. So there you can be skeptical, but at the same time, like these are overwhelmingly positive results, right?

Murilo (42:32)
True, true, true.

Yeah, which is very contrasting with what we saw from the MIT report, which again, fair, like I have heard some criticism over the takeaways that a lot of people have drawn that the MIT reports, like basically MIT report, the headlines was 95 % of projects fail.

Bart (42:46)
I think too.

Murilo (42:59)
And some people came back to actually the report doesn't say that it says that 95 % doesn't show clear ROI, but there's a big chunk of those that is just inconclusive. So that doesn't mean the project fails. So like there was some criticism on that, right? There's also criticism on how the report was made. was some controversy there. And now we have this one that kind of says 80%, 74 % with clear ROI and all these other things, right? So.

Bart (43:09)
Mm-hmm.

Yeah, definitely.

Murilo (43:27)
Yeah, this is also for my college knowledge at Wharton. I didn't know Wharton, but I know University of Pennsylvania, right? So this is a business journal from the Wharton School of the University of Pennsylvania. So it's another big university in the US, reputable as well. And also a GBK collective, but I don't know about this collective here. So I think on the first page they also talk a bit about it. Do you know them or not?

Bart (43:38)
Yep. Indeed.

But again, it's a report, it's based on their survey, it's not a peer-reviewed article, So that's again, like you can, but it is a positive sign. Let's look at it like that.

Murilo (43:57)
Thank

After having seen these two things, before we move on, do you lean more towards one or the other?

Bart (44:03)
Well...

You mean whether or not AI will have a positive ROI on companies?

Murilo (44:20)
Yes, is how much height?

Bart (44:21)
I think.

I personally think that especially in digital heavy sectors, I will have an overwhelmingly positive effect on ROI.

I think we're already seeing it today. I think we will even see it more in the future. But I think it will come at the detriment of jobs.

Murilo (44:45)
Yeah, there's actually one article that we're not going to have time to cover today. They talked a bit about this, right?

Bart (44:52)
We have an alternative one,

the next one.

Murilo (44:55)
And that's true. So what is this next one? And I think that's my turn to take a look. A massive scrape of job posting finds steep drops in creative execution roles like computer graphic artists and writers, while machine learning engineers surge 40 % year over year. Watchlist, medical scribes roll 20%, hinting at AI note-taking tools biting into entry-level healthcare admin work. So again, I think...

nice segue like we have the the gloomy reports, we have the very positive report and then we have something that is really data driven right like it's very just looking at job postings right which is from many many many different places. This is in the US though I want to say or at least focus in the US but still I think it feels more tangible right feels more concrete and data driven than what we've seen so far.

Bart (45:50)
Yeah,

it's interesting. I think I would suggest people to have a look at the article. It's interesting how they did it. They work with a data provider. They also worked with Amazon Mechanical Turk to label jobs into certain categories.

It's a very rich data set, right?

Murilo (46:09)
Actually, is the chief, I think I saw, he's the chief of research actually at this place. So they didn't necessarily, yeah, had a research at the place. I think he's part of his, but I thought it was interesting in any case. So he also shares again all the methodology, the data. He also used Mechanical Turk. So he didn't just use AI to add taxonomies, right, for the AI.

Bart (46:17)
okay, I see. He is the chief research at the data provider. Okay, interesting.

Okay, interesting.

Hmm.

Murilo (46:36)
I like the transparency and it's very understandable how he got to these things. So walk me through this.

Bart (46:37)
That's very interesting.

Well, overall global postings, indeed, like you already mentioned, like we're talking US, but I think.

we can more or less, it's a dangerous statement, can more or less guesstimate that we see similar trends in the EU. Maybe not the same, maybe there some lags, there similar trends. Global postings, they fell 8 % in 2025 versus 2024.

Murilo (47:09)
Yeah, and I think he also mentions that it's important because he uses this as a benchmark. if overall job postings fell 8 % and you see that for example, creative writing dropped 8%, that just means zero delta, right? It just followed the trend, the overall trend. That's how he approached this, right?

Bart (47:27)
Yeah.

There are some big numbers. So for example, computer graphic artists minus 33 percent 2025 photographers minus 28 percent writers minus 28 percent journalists and reporters minus 22 percent. That's big,

Murilo (47:49)
Yeah, big numbers.

Bart (47:50)
Big changes, not all creatives are down. So yeah, for example, more the strategic and lead roles, creative directors and producers, they more or less like they hold up versus the last year.

Murilo (47:54)
Yep.

Which does make sense when you look at AI, Like if AI can generate content faster, the amount of people you need to do quote unquote the grunt work, right, of actually creating the things, it is smaller. So there is something that you could explain with AI. But the people making the decisions of the campaigns and what's good and what's not good and signing off, that's still very relevant.

Bart (48:08)
Hmm.

Yeah.

Yeah.

I think the difficult thing when looking at these job evolutions, you have a lot of things going on at the same time. You have this upcoming capability on AI, GenE, ILMs. You also have this economic instability, economic parameters that are less good inflation, et cetera. There's geopolitical imbalances at the moment. But I think all of the things that we're looking at here...

like journalists, photographers, et cetera, like intuitively you link these two AI becoming better. Right? And AI becoming better means a decrease in these jobs. And if we like take this further, like there is another one on more regulatory work, like this is typically very paperwork heavy work, like compliance specialist minus 29%, sustainability specialist minus 28%, environmental technicians minus 26%.

Murilo (48:57)
Yeah, indeed. So again, we could.

Yeah.

Bart (49:18)
Chief compliance officers minus 37 % sustainability managers minus 35%. Again, because it's very paperwork heavy, like intuitively you can link this to AI becoming better at these things, right?

Murilo (49:34)
Yes, I would also add, and this is also something he mentioned as well, this is US based, and I think if you look at the geopolitical movements as well in regards to sustainability, feels like the sustainability agenda also dropped a bit. So having people using these things is not as top of mind, right? But they did mention that was, I think, a trade compliance specialist group, right? So again, geopolitical things.

Bart (49:44)
Agreed.

Yeah, agreed, agreed.

Yeah. And there you see

like, like that's probably because of the tariffs and there you see like, even if AI is better at these things, like when you need real expertise, you're going to hire an expert for this.

Murilo (50:08)
Yeah, exactly, exactly. Which is a bit, I mean, maybe we're skipping a bit, but there's also the mid-level managers and the top-level managers, which is a bit, yeah, the men for senior leadership is much stronger than middle managers plus individual contributors, which...

It feels a bit like similar trend, right? Like the more quote unquote grunt work, which is a bit of a pity because a of times that's where people learn a lot, right? When they're executing the job that someone else thought about, right? And they learn some things. So it's also something that...

Bart (50:35)
Hmm.

Agreed, agreed.

Murilo (50:44)
you see as a trend, right? So maybe just to add some numbers to this, individual contributors minus 9%, middle manager minus 5.7%, which actually is technically above the baseline, but senior leadership, it's 1.7 % less, minus 1.7%. So you see a bit of the thing there.

Bart (51:03)
And maybe on the other side of the coin, like if we say, if we do think all of this is linked to AI becoming better as a skill, if you look at the jobs that are needed there, we basically see a positive sign. So yeah, for example, machine learning engineers plus 40%.

Murilo (51:17)
Yeah.

Bart (51:21)
even after doing already 78 % in 2024. Robotics engineer up 11%, research and applied scientists up 11%, data center engineers up almost 10%. So the opposite side of this, right?

Murilo (51:35)
Yeah, and we'll see 10 % is 10 % on increase, right? So it's even more than the benchmark. So it's true, it's true. I was also wondering like data engineering, machine, I mean, data engineering I can get, but like machine learning engineer, which is the top one, there are some people that argue that to build AI, GenAI systems.

You don't need a machine learning engineering background. think the data here, well, maybe for people listening, right? The argument is that machine learning engineers, train the models. They have the data, they need to experiment, they have different models to try, different parameters, and they need to basically see if this model is good or not. But basically, you create something from scratch, whereas with Gen.AI or a GenTech AI, you take these off-the-shelf models from Entropic or OpenAI, so there's no training, so you can just use it. And this also means that usually the infrastructure

So where the model runs is not on your local machine, is on very specialized hardware, which is also very expensive. And with that change, there's a lot of people that argue that for doing AI, so gen AI work, you don't actually need to have a machine learning background or a data science to understand all these things because the skills are different. But if you look at this number here, just purely looking at the job positions,

That's not true, right? You still need people with the background to be able to build the systems. Unless you say that AI and general AI just increase the popularity of machine learning and that's why people are being hired, right? But yeah.

Bart (53:06)
What I think is happening here, like I think the companies that are building foundational models, like Entropi, like OpenAI, I mean, they're still hiring more people. I don't think that is the growth that you see here. I think the growth that you see here is indeed people that are using these models are building a wrapper around OpenAI and the logic around that.

They used to be called software engineers and simply because they touch LLMs now, in some way, a fashioner and their job name switched to machine learning engineers. I think that's a big part of the shift here.

Murilo (53:34)
I see what you say.

Huh, so you think that a chunk, again, they're speculating here, but you think that a chunk of these jobs are just, like what I'm thinking when I heard machine learning engineer five years ago, is not the same thing that these people are talking about. This is more LN, API calling on these things. I that's a more plausible explanation, actually. So yeah. The thing that is a bit to me, again, maybe the taxonomy and the methodology, right?

Bart (53:52)
No, I don't think so.

Murilo (54:06)
Today, if you ask me, and the AI engineer is something that really works with LLMs, it's very specialized in that, and I think there is a specialization as well on that. A machine learning engineer is someone that builds models. But also the fact that there's no AI engineer here, and you see data engineering, you see all these things, think also, I think it speaks to your point as well, right? think maybe the taxonomy, the methodology. Yeah, exactly, so, yeah. Yeah.

Bart (54:22)
Hmm.

bit of a bucket definition.

So yeah, guess this is the only reason for which, the only people for which this is positive is people that are today in the tech sector and doing something where they are.

Murilo (54:45)
Maybe one thing that also stood out to me was the influencer marketer that also grew. I think there is a bit of a, well, again, you could say that is this AI, is it because of ⁓ using AI tools for marketing, like we mentioned? One of the arguments that the author puts, which I thought it was interesting,

Bart (54:46)
For the rest is, yeah.

Murilo (55:12)
is that there is a trend with JNI in the sense that there's so much JNI's law created these days that...

ads and marketing campaigns and all these things, they don't feel really genuine. But having influencers that are very relatable to you to showcase products, to try out things, to see a video of someone that like kind of the same demographic as you, it's very genuine. And I think the, and what he argues is that maybe the explosion of AI is lot.

makes people crave more for the genuine thing, you know? Like this is really someone that is trying these things. And that's why he was hypothesizing that there is a growth there. What do you think of that argument?

Bart (55:57)
Let me paraphrase. So what you're saying here, this very positive strength on like that there are more like influencer marketing specialist hired is because people value authentic stuff more. That's what you're saying, right?

Murilo (56:08)
I think authentic more in the sense that it's not AI generated. It's an actual person behind it.

Bart (56:13)
Yeah,

don't believe it. So I do believe that we will see actual human created creativity in contents will get a higher value in the future. I think it's already happening. I think it's already happening. Like if you can stand out by who you are, the things you can think of, you can create, etc.

Murilo (56:15)
You don't think that? Okay.

Like, yeah, it's true.

Bart (56:36)
helps you to stand out from the all of the generic stuff that is out there and which is becoming more generic because everybody uses the same tools. I think the trend that you see here in influencer marketing is simply that from the moment you start marketing as an online brand or as an online or an e-commerce or clothing or whatever like the channels today are Instagram, TikTok if you focus on the

the 50 to 70, you do a bit of Facebook, you do a lot on podcasts, like, and all those you can capture on there. You, if you want to go viral there, you need to be an influencer. To me, it's just like, it's a bit of a shift of the older marketing channels to what today the biggest ones are. And like the new name that comes with it. That's what it looks like to me, to be honest.

Murilo (57:23)
Yeah,

so not necessarily AI, it's just that people are not watching TV much, they're watching their shorts, right, like TikTok or something. That's just like a natural evolution that doesn't have street correlation with AI.

Bart (57:35)
Yeah. And then TikTok is a very good example. Like TikTok,

need to, like, you don't put your company video there, right? Like you need to have an influencer, somebody that talks about you, like someone that, that built something around your brand, like some, some identity and like you need an influencer. Right.

Murilo (57:48)
I see. True. One thing that I heard, I don't know if you've seen this with your kids or friends of your kids or parents that, you know, whatever.

I heard that kids today, like when I was a kid, actually, when asked like, what do want to do when you grow up? And I grew up in Brazil, so take that in mind. It was always like football player. All the boys wanted to be football players. I think on my dad's time, or maybe, I don't know, or maybe other places, they wanted to be astronauts, right? Nowadays, what I heard on a podcast is that the podcasters' kids, they wanted to be influencers.

They want to be YouTube creators. Yeah, I know. And that was a bit the whole, yeah, exactly. That was a bit the whole discussion, right? Like, should we, because like, it breaks it down a bit. Like, why should we try to inspire people to do hard things, right? To, you know.

Bart (58:28)
That's depressing.

Murilo (58:39)
And maybe actually the motivation is more about the fame. So it wasn't like people wanted to be astronauts, not because it's so challenging, maybe it's because they're famous. And now the influencers are famous, right? But yeah, but like, I don't know if you felt any of this with, I don't know, personal experience. Yeah, no. Okay. But it would be, it would be interesting to see. And yeah, like the influencer, think it's, I do think that quite a lot of kids, they would like to be.

Bart (58:44)
Mmm.

Yeah, that's probably true,

With my kids now, I think they're still little bit too young for this.

But there's maybe a

good insight that you have to like, like, like it sounds like going from astronaut to football players already stepped down and down to influencers, tumble down the stairs. But maybe it's just like, like which people are at one point in the global society, like elevator, like these are our heroes. Like, and probably like in the, in the sixties, in the seventies, these were astronauts and football of course, a very big thing. And now it's influencers.

Murilo (59:17)
Yeah.

Exactly. These are the people who venerate, right? Yeah.

⁓ Yeah, again, speaks a lot to where we are.

Bart (59:38)
Interesting.

The pessimistic thing about this report is of course, there's a big, big decrease in a lot of different categories. That means that there is a lot of hiring stops. It's also like we got this confirmation also from Jerome Powell, the chair of the Federal Reserve, a few days ago that job creation is very low. Job finding rate for people who are unemployed is also very low.

Murilo (59:54)
Yes.

Bart (1:00:00)
And all these hiring freezes basically also means that it will take longer for these people to gain experience. It feels a bit like a downward spiral. Right? So I'm...

Murilo (1:00:10)
Yeah, which is not

a cool cover today, but we will send on the newsletter. So that article and the other article. So that one is from someone from in the US government, right? So someone that hasn't looked at the numbers and he has said that like when you analyze the numbers and you dive a bit deeper, we should be concerned, right? And the other article that we're not gonna be able to cover today, do you remember what it is?

Bart (1:00:15)
No.

It's the one about the bot farm that's in recent horrors invested in.

Murilo (1:00:37)
Yes, yes, yes.

Basically it's like a company that they got a lot of money from A16Z, so big venture capital. it's basically, well, the click baity title is like a AIslob creator, right? So we'll let people.

Bart (1:00:55)
Yeah, they're basically investing

in a huge bot farm that will create even more AI slope.

Murilo (1:01:04)
and also interact with users, right? So if someone says a question, there's some AI bots to reply, which is also illegal for most social platforms, Reddit, Meta, Instagram, they don't allow for AI spam, let's say, and also to make money off of this, but that's their own business model. Yeah, if people are interested to have a look, subscribe to our newsletter.

newsletter.monkeypatching.io, that right part.

Bart (1:01:32)
Yeah, that's right.

Murilo (1:01:35)
And I think that's it we have for today, unless there's anything you want to leave everyone with part.

Bart (1:01:44)
thank you for listening and we'll check in again next week.

Murilo (1:01:49)
Thanks everyone, see you all next week.

Bart (1:01:52)
Ciao.