—
no predictions resolved yet
0 pending · 0 expired
But we also do support NVIDIA, and we're one of the big purchasers of NVIDIA chips, and we have them in Google Cloud available for our customers in addition to TPUs.
For, you know, really good AI advertising. I just, I don't think we're going to like necessarily our latest and greatest models, which are, you know, take a lot of computation. I don't think we're going to just be free to everybody right off the bat. But as we go to the next generation, you know, it's like every time we've gone forward a generation, then the sort of the new free tier is usually as good as the previous pro tier and sometimes better.
Well, okay, it's free today without ads on the side. You just get a certain number of the top model. I think we likely are going to have always now like sort of top models that we can't supply infinitely to everyone right off the bat. But, you know, wait three months and then the next generation.
Well, we mostly, for Gemini, we mostly use our own TPUs. But we also do support NVIDIA, and we're one of the big purchasers of NVIDIA chips, and we have them in Google Cloud available for our customers in addition to TPUs. At this stage, it's, for better or for us, not that abstract, and maybe someday the AI will abstract it for us. But, you know, given just the amount of computation you have to do on these models, you actually have to think pretty carefully how to do everything and exactly what
we're definitely pushing all the bounds in terms of intelligence, in terms of context, in terms of speed, you know, you name it.
I mean, for any such cool new idea in AI, there are probably five such things internally. And, you know, the question is, how well do they work?
We had this AI tool that actually was really powerful. We, unfortunately, anyway, temporarily got rid of it. I think we're going to bring it back and bring it to everybody.
But nowadays, these things, I think, are more sensible. I mean, there's still battery life issues, I think, that, you know, we and others need to overcome. But I think that's a cool form factor.
Yeah, I kind of messed that up, I'll be honest. Got the timing totally wrong on that. There are a bunch of things I wish I had done differently, but honestly, it was just like the technology wasn't ready for, for Google Class.
I mean, I have to give credit to where credit's due. I mean, DeepSeq released a really, surprisingly powerful model when it was January or so. So that definitely closed the gap to proprietary models. We've pursued both. So we released Gemma, which are our open source or, you know, open weight models. And those perform really well. They're small, dense models, so they fit well on one computer. And they're not as powerful as Gemini. But, I mean, the jury's out which way that's going to go.
And this is sort of broadly true across machine learning. I mean, you used to have all kinds of different kinds of models and whatever, convolutional networks for vision things. And, you know, you had whatever RNNs for text and speech and stuff. And, you know, all of this has shifted to transformers, basically. And increasingly, it's also just becoming one model. Now, we do get a lot of oomph. Occasionally, we do specialized models. And it's definitely scientifically a good way to iterate when y
I mean, for myself, definitely makes me more productive.
we're trying to, you know, roll out every possible kind of AI. And trying external ones, you know, be whatever the cursors of the world, all of those, to just see what really makes people more productive.
recently, I just had a big tiff inside the company because we had this list of what you're allowed to use to code and what you're not allowed to use to code. And Gemini was on the no list. I mean, nobody would, like, enforce this rule. But there was this, you know, actual internal web page. For whatever historical reason, somebody had put this and I had a big fight with him. And I, you know, I cleared it up after a shocking long period of time. Anyway, it did get fixed, and people are using it.
I'm probably the one weirdo who doesn't, who's not a big fan of humanoids. But maybe I'm jaded because we've, you know, we at least acquired at least two humanoid robotic startups and later sold them. But the reason people want to do humanoid robots for the most part is because the world is kind of designed around this form factor. And, you know, you can train on YouTube, we can train on videos, people do all the things. I personally don't think that's giving the AI quite enough credit. Like, AI
I mean, I think we've acquired and later sold like five or so robotics companies, and Boston being one of them. I guess if I look back on it, we built the hardware. We also had this more recently. We built out everyday robotics internally, and then later had to transition that. But, you know, the robots are all cool and all, but the software wasn't quite there. That's every time we've tried to do it to, you know, to make them truly useful. And presumably one of these days that'll no longer be tr
not just our models, but all models tend to do better if you threaten them.
when you use some of our AI systems, you know, it'll suck down whatever top 10 search results, you know, and kind of pull out what you need out of them, something like that. But I could do that myself, to be honest. You know, maybe it would take me a little bit more time. But if it sucks down the top, you know, thousand results and then does follow-on searches for each of those and reads them deeply, like, that's, you know, a week of work for me. Like, I can't do that.
And more recently, the post-training, especially as the thinking models have come around. And that's been, you know, another huge step up in general in AI. So, you know, we don't really know what the ceiling is.