generative AI

“It's probably, I feel like there's a whole second wave of what you can do with this stuff. And it will get to video, and it will get much more specific and precise and controllable, and it will answer a lot of sort of make me 20 options of X and Y and Z that h…”

“Now there are some people who think this is another step towards AGI, which is like, I think Sam thinks that I'm not, I didn't think I'm, and I'm not an AI scientist, but I'm not sure there's a consensus of the majority of people would think that.”

“You could make a shitload more Mozart. You wouldn't, that wouldn't get you Beethoven. And you might stumble on Beethoven, but you would know, it wouldn't know it was Beethoven to do more of that. And it would be buried in the other billion things that you've d…”

“And so, yes, you will, you would expect to see kind of a great chunk of this stuff getting automated in some way.”

“This is going to, you know, never, no, not the biggest of issues, but this is going to be, we could be pretty disruptive to the porn industry. Because you will be able to just kind of tie stuff in, you know, for the people who look like this, who doing this, a…”

“So basically what this is, is if you think very, very crude terms of machine learning is pattern recognition on a massive scale. This is take that pattern and make another one. So where image recognition is to say, does this, which pattern does this fit? Does …”

“You're going to have people generating content in response to what they think search traffic is automatically. So you will have automatically generated video with automatically generated voice targeting search queries at massive scale. There will be a new floo…”

“I think there's a, you can, one can dismiss very quickly the concept, oh my god, this is like the end of art. No, no more than Photoshop was. It's another form. It's another thing that you can do.”

“And now suddenly this isn't just writing funny bits of text. This is making, or this is automation of the creation of something. And I think that that we're going to see a lot of this sort of stuff.”

“And clearly you should be more, if you look at it and you say, well, it makes cats, but the cats have four, have three legs. If you say, well, yeah, it can make some text, but the text isn't very good and it's got mistakes in it. You're frankly, you're an idio…”

“And so this is why you have GitHub saying, well, we're going to call this Copilot, and like, it's writing code for you, but you're a coder, you can look at it and see. You can look at it and check. And so what are the domains where You can tell it's wrong. Ver…”

“What the four images and the 10 links are doing is communicating uncertainty, whereas the one answer that's in text presents itself as the certainty far more than you get with than you get with the others, and”

“This is a two by two matrix, which is, you know, does the error matter and can you spot the error?”

“The point is that, that images are both much higher bandwidth, but also much higher noise. And so the mistakes can be kind of hidden in the noise that, like, the clouds aren't quite right. If you're not a meteorologist, you don't know, and you don't, and you a…”

“If you're going to, you know, build a new system that will just tell you what the story was without linking or giving you credit, that's very different from you know, what Google and Facebook do, which is you search for a new story and they tell you where to r…”

“It's making a pattern without any structural understanding of what the actual thing actually is.”

“It changes an awful lot of how tech works and what tech products you'll use. And there'll be all sorts of stuff that's happening inside stuff you use you don't even know is using this, which is also what happened with machine learning. So at the minimum, it's …”

“It's also going to happen to porn. It's just going to demolish the porn industry. Because again, there's vast amounts of training data.”

“If I've now looked at millions of pictures of wizards and I draw something, do I have to pay them? Like, no. If I copy your thing, do I have to pay you? Well, yes.”

“Where is the, like, where is the creative energy at the moment? And it's not in social media anymore. Like, obviously there's a, yeah, absolutely. Yes. Yes. That otherwise absolutely all attention is now totally maniacally focused on on generative AI.”

“The challenge is as I always, I've always been saying, if you use this as a general search, then the error rates might matter and you can't tell.”

“As soon as you move away from make this shot better, you're probably moving away from a prompt box, and you're probably moving to buttons, and switches, and options, and palettes, and lists of possibilities, and a screen that says, okay, here are six different…”

“But the idea that like it might cost, it's like it actually has more, a consumer internet company has real tangible marginal cost every time a consumer does something. That's kind of a new thing.”

“The second question is right now, if you ask this a question and the result is three or four paragraphs, one of those paragraphs will be completely wrong.”

“And it says we don't have an Excel yet. You've got the PC as this sort of substrate. Suddenly everyone has a computer and it's only, it's amazingly cheap.”

“If you want to build a big, complicated thing, there is not going to be just one button that says go that produces a very specific, complicated thing you had in mind.”

“And the company creation is going to be in the middle, which is how do you create actual like tools with lots of stuff on top of this?”

“We haven't got to the point that we've got a mental, even if you think it's a platform shift, we haven't got to a point that we've got like a kind of a useful mental model of what would, what is this? Whereas for the last wave of machine learning, we're now ki…”

“What is generative AI trying to do? Which is, is not actually trying to reproduce any individual thing in the training data. So, you know, when we did image recognition models, 10 years ago, you give it a billion pictures of cats. The purpose is not that it ca…”

“There's a difference when you do, something can be the same theoretically, just at greatest in principle, but a much greater scale and practical and accessible and everything else. And in fact, making something much greater scale and much more accessible, but …”

“One of my good friends is a filmmaker, and AI has completely shifted the way that she can create, for example, the trailer. It used to, you know, you'd used to need, what, a million in budget and a couple of months to create a good trailer for a movie? She can…”

“So Accenture revenue from generative AI projects in the last four quarters, a hundred million dollars, two hundred million dollars, 450,000,600. Well, over, over 450 and over 600. So they've done 1.4 billion probably in the last 12 months, and 1.1 billion in t…”

“It seems to me that there were sort of four use cases. Use case one is coding. Secondly, brainstorming and suggesting ideas. Thirdly, and this is more like a corporate thing, knowledge management pointed at an internal corpus of thousands of documents or PDFs …”

“So you need to have a use case where either it doesn't matter if it's not a hundred percent right, which is, you know, Give me 50 ideas for slogan for a new brand of toothpaste. Like there's not, you're not, there's not like a complete. There's no right answer…”

“They asked about hallucinations, and they said, do you believe that generative AI always produces accurate answers? And the question is still, like, a third of people say yes, even if they've tried it.”

“It seems like we're now sort of at the point that every big company's got this, everyone big company sort of is thinking about what they do with it, and no one's got any production yet.”

“Every big company's got a trial. No one's really using it very much, or not many people are using it very much. There's a small number of fields where you had this immediate obvious use case, like it's great for coding, it's great for marketing, it's great for…”

“And so Accenture is now doing a billion dollars a quarter in what they call generative AI bookings, doing pilot projects and building stuff for big companies.”

“These models are very good at automating certain kinds of things, but not others. They get stuff wrong, or rather what they do is they say, what would an answer to this question probably look like? That's about the best way I can come up with formulating it ri…”

“Every consumer has tried this, but well, a very large portion of consumers, something like a third to a half of all consumers in the developed world have now tried this, but only like two, three, four percent of people are using it every day.”

“There are no network effects in building the models yet. There may be in the future, but at the moment we don't know.”

“Because what generative AI does is it says this is what a good answer probably looks like. This is what most people would probably say. This is how most people would probably make it.”

“The narrow problem with that, of course, is that that was built out ahead of demand, whereas this is being built out behind demand.”