Skip to main content
Maty Bohacek

Maty Bohacek

Nov 16, 2026

8:07
Let me turn it over to our speakers and let's start with Maddie and I'm going to magically call up your PowerPoint.
8:18
All right, well, thank you so much for being here.
8:20
It's great to see you all.
8:22
And thank you for the very generous introduction.
8:24
My name is Maty Bohacek.
8:26
I am at Stanford.
8:28
And my goal here today as the computer scientist on this panel will be to try to frame this in the sort of current work that's happening in AI, both in the sort of research and more academic circles and also in the very pervasive and salient AI circles in the industry.

49 MINS LATER

58:15
But I ask you, in the work you're doing now, and you're up to your eyeballs in very valuable research, what is the imputed revenue model? How do you think these guys are going to be, these people are going to be making money?
31:52
So as such, Mati, can you say with any certainty if this third of new websites which is using AI, who's making them?
31:59
I- I- it's very difficult to say.
32:01
Uh, we have seen a couple of patterns that I, that I wanna bring in.
32:05
Um, one thing is that there's a lot of these sort of propped up sort of mass scale news websites, or seemingly news websites, where there's this pattern or this sort of template that is being used to then sort of localize the news or seemingly localize the news to individual towns and cities.
32:21
We've seen this mostly in the US where there's this sort of same template for a website, and then it's copied over and over and over again hundreds of times.
32:28
And so there's one for, you know, every village in, say, New Jersey or every town in New Jersey, and it seems like it's, it's sort of local to you as a resident of that town, where it's still the same content or AI content.
35:22
... whoever is here, what will you, what answer will you give them?
35:26
Well, it, it's very funny, but, um, yeah, so, so we're already seeing, uh, increased rates.
3:01
How is this different from how people perceive that artificial intelligence is changing the web?
3:06
This is really interesting because we, uh, tested these hypotheses, uh, against a representative pool of Americans, or American adults, I should say.
3:17
And, um, we, uh, put forward six hypotheses in addition to this diversity of ideas or topics and the, uh, sentiments, this cheerfulness.
3:27
We also tested, um, impact to truthfulness, so this, this idea that AI might be making the internet less accurate factually.
3:36
Also, the idea that AI might be, uh, providing fewer references or citations and links.
3:42
Also, uh, the idea that, um, AI might be just generating so much text that really is just, just extra fluff that doesn't have a lot of meaning.
3:50
And finally, um, this idea that AI might be, um, making the diversity of style, of the way, uh, the text is written, less diverse.

5 MINS LATER

9:18
You know, as it advances, do you think the websites it creates could present a greater range of perspectives and also maybe be less peppy? [laughs]
44:51
Mm-hmm
44:51
... might happen is the following.
44:53
I, I think that so far we've been, um, really exploiting the, the scale of data and computation that we have, and when I say data, I really mean, you know, these companies and, and, you know, many different entities in AI just, like, scraping anything they could find on the internet and using that in their training data.
45:11
So, so that was one big factor or one axis along which you can scale, and they've mostly exhausted, like, you know, all the data that is just, you know, that, that can be humanly found on the internet and even, you know, like archives.
45:23
Like Google or Anthropic, like, bought books, right? To scan books to get additional data because they couldn't find anything else on the internet, right? So, like, they, they've really maxed out that aspect or that axis, right? So that's, that's the first axis, and, and that has been yielding really, like, major improvements, you know, along the way.
45:39
The same thing happened with compute, right? So this, this talk about, like, GPUs and, like, chips and NVIDIA, like, that's what they do.
45:46
They, they provide specialized chips for AI to be trained and run.
47:18
But, uh, what is your, your read of this, having worked on, um, this season of the show?

We value your privacy

We use cookies to understand how you use our platform and to improve your experience. Click “Accept All” to consent, or “Decline non-essential” to opt out of non-essential cookies. Read our Privacy Policy.