Is the Future of AI Too Fast for Humans to Handle

Sometimes people have very pie in the sky ideas um about where the future will go. And I do think a lot of that is stuck in what they'd like versus like maybe practically how [laughter] the world's going to go. And um you can never really predict that kind of stuff. So I appreciate that. That's fair. >> Yeah. And and I think and just just one more quick thing there is like the become the world has become a lot more nonlinear because of how things are progressing right it's just like the the the your prediction cycles are so it's like it's like of course you can never predict even tomorrow but I mean just in terms of macros the the the level of dynamic trends that are concurrently happening at the same time is never been more like each each of these specific trends are at

an accelerating pace, right? Robotics is at an accelerating pace. Computer vision is at an accelerating pace. LLMs are at accelerating pace, right? So, there's just like this in if you if you consider that as part of the infusion and the equation to trying to see the world, it it's kind of really u myopic to kind of assume that you can actually even build a a year plan for a company, right? I think my job is like to just take into account everything that's happening in AI and almost be willing to go not 180 as much because doing that twi like every two years I'm going to [laughter] question my life and you know what it means but like to truly every every sort of like 90 days to be like hey do we have the right set of principles or fundamentals assumptions in terms of

where the world is going right um yeah so I think uh I don't have three to five year plans I I only operate on 90-day plans now. >> H that's fair. I I think it's hard like you said that the time just does not feel the same that it used to. Like I I was talking about it with somebody recently. There's a insane level of uh quick um sort of turnaround on on model improvement at at a minimum, right? Like we're we're talking text inside of AI generated images was impossible until like March, right? And now Nana Banana comes out and you're like you can basically Photoshop anything in a sense like in the the colloquial term of what Photoshop means. >> You can replace anything with anything. You couldn't do >> video at all in a reasonable manner. Now, there's multiple competing models that

are killing it, that are making better, funnier meme videos than you could make practically because people can come up with crazy ideas and it'll bring it into fruition. Um, you have models that at first coding was like maybe helpful. Maybe you can do formul like it help you with formulas and stuff in certain products, but you know, like is it really going to do a great job when it comes to coding? No. Now, we got models like Claude 4.5 Opus that send you from like a recurring can't figure it out cycle to insanely good suggestions and insanely good speed of coding. Um, I could go on and on and on, but you get the point. Like, I thought text inside of images was for, to use the kids lingo, cooked um, and was going to be for a while. But uh now we're in now

we're in a position where it's where it's cooking. Yeah, exactly. That was good. That was a good layup. Yeah, it went from like that was [laughter] dude cooking. Like it's it's insane the 180 >> and you don't know when it's going to happen. >> You don't when the models are going to reach certain levels. These companies are just continuing to innovate in ways that I would have never predicted. Um, like for example, Gemini is one that a lot of people aren't considering, but they're visual uh onscreen context thing, incredible.