Reid Hoffman

And their thing is we're not building our own models. We're just deploying other models. So we have best of breed for what models are available to us. You could do that as an AI startup. There's a whole range. And part of what I love about entrepreneurship and invention is I have a whole set of theories about what kinds of things will be really big. And some of my theories will be right.

And their thing is we're not building our own models. We're just deploying other models. So we have best of breed for what models are available to us. You could do that as an AI startup. There's a whole range. And part of what I love about entrepreneurship and invention is I have a whole set of theories about what kinds of things will be really big. And some of my theories will be right.

And I've also discovered some other people who have amazing theories. And that's one of the things I really think I and we at Greylock love to invest in.

And I've also discovered some other people who have amazing theories. And that's one of the things I really think I and we at Greylock love to invest in.

Start building it. If you want to, if you can, If you're technical, download an open source model, Lama, Mistral, other, start playing with it. What if you're not technical? Well, generally speaking, you can't do a technology. It's very hard to do a technology startup without a technology co-founder. So then go find a technology co-founder.

Start building it. If you want to, if you can, If you're technical, download an open source model, Lama, Mistral, other, start playing with it. What if you're not technical? Well, generally speaking, you can't do a technology. It's very hard to do a technology startup without a technology co-founder. So then go find a technology co-founder.

Too slow. No, no, speed matters. Speed matters in startups. But by the way, because we tend to so much lionize what is usually the man versus the person, most of these people are men, the individual, startups succeed because of their initial team. That's part of the hiring point we were talking about earlier.

Too slow. No, no, speed matters. Speed matters in startups. But by the way, because we tend to so much lionize what is usually the man versus the person, most of these people are men, the individual, startups succeed because of their initial team. That's part of the hiring point we were talking about earlier.

And so generally speaking, if you ask me would I rather invest in an individual founder or two to three co-founders, I'd rather invest in two to three co-founders. So going and finding a co-founder is a great thing. Because by the way, your throw weight, your capability weight is so much better when you do that. And so Mark Zuckerberg had Dustin Muscovitz, Adam D'Angelo, Chris Hughes.

And so generally speaking, if you ask me would I rather invest in an individual founder or two to three co-founders, I'd rather invest in two to three co-founders. So going and finding a co-founder is a great thing. Because by the way, your throw weight, your capability weight is so much better when you do that. And so Mark Zuckerberg had Dustin Muscovitz, Adam D'Angelo, Chris Hughes.

That's a really important part of how things get created.

That's a really important part of how things get created.

So Sam Altman, Greg Brockman, Ilya Suskefer, amazing. And they had a focused, bold vision, which is, yes, Google has invented the baseline of the attention-based transformer, but they don't realize that it's just scale, right? like apply scale to this. We are going to bet the entire company, everything we're doing on scale on it. And we're going to be very focused on doing it. That's great.

So Sam Altman, Greg Brockman, Ilya Suskefer, amazing. And they had a focused, bold vision, which is, yes, Google has invented the baseline of the attention-based transformer, but they don't realize that it's just scale, right? like apply scale to this. We are going to bet the entire company, everything we're doing on scale on it. And we're going to be very focused on doing it. That's great.

And I had passed on a lot of AI investments. But the thing that people need to understand about AI is it's a transformation to scale computing learning systems. So as opposed to we program it, like we program the AI and We program a scale computing learning system so it learns. And that only becomes available once you have scale compute. So you need the internet. You need the cloud.

And I had passed on a lot of AI investments. But the thing that people need to understand about AI is it's a transformation to scale computing learning systems. So as opposed to we program it, like we program the AI and We program a scale computing learning system so it learns. And that only becomes available once you have scale compute. So you need the internet. You need the cloud.

You need massive data centers. And you need access to massive amounts of data, which you get from the internet and other things. And then you need a scale team to build it.

You need massive data centers. And you need access to massive amounts of data, which you get from the internet and other things. And then you need a scale team to build it.

Almost always critical because you need that initial team. Now, you don't need networking of, I should spend 12 hours a day, seven days a week, going to every party that I can possibly go of and shaking people's hand and giving them my business card. That's not particularly useful.

Almost always critical because you need that initial team. Now, you don't need networking of, I should spend 12 hours a day, seven days a week, going to every party that I can possibly go of and shaking people's hand and giving them my business card. That's not particularly useful.