Searching...
16 results for “generalist approach”
...generalist tasks. But, really, what's happened is the models have gotten so good that the generalists are no longer needed. Like, what they really need is experts, experts across every area that the models are focused on. And really, you could think
10:25 / 11:31
from transcript10:25 – 11:31
“Why going narrow on domain expertise beats being a generalist VC”
is I found going narrow on domain. Right? And that may not necessarily help you with reading the personal, the kinda more fundamental level, but definitely helps you in calibrating a founder's underst
18:10 / 18:46
tactical advice18:10 – 18:46
useful in more settings than this obscure fact that it has. So I think that's always attention at the same time. You also don't want your model to be kind of completely detached from, you know, knowing stuff about the world. Right? Like, it's probabl
51:00 / 52:10
from transcript51:00 – 52:10
street signs and some things. So I train a street sign recognition recognition model or I want to decode speech recognition. I have a speech model, right? I think now the era of unified models that do everything is really upon us. And the question is
48:48 / 49:50
from transcript48:48 – 49:50
from specialists, maybe that won't be as necessary. I don't know. Maybe you have some thoughts. So here's my current mental model. There there's two personas. There's persona number one is the expert in the space, and there's persona number two is th
44:50 / 46:13
from transcript44:50 – 46:13
“Why going deep beats going wide in today's mature startup market”
ability to be successful in that domain when it comes to surrounding themselves with the right advisers, the being able to track the right employees, being able to tap into the right customers' networ
18:29 / 19:03
tactical advice18:29 – 19:03
domains of intelligence? Like, unbundling the models into multiple experts in different areas, etcetera? More directly. Yeah. Instead of just MOE that we have no exposure to. Because that can be, like, confusing as a user from the outside Uh-huh. Whi
29:15 / 30:16
from transcript29:15 – 30:16
architecture in general if you assume that that, you know, there's we're never gonna get to a point where the model is just a 100% perfect. Right? And so it might also be the right, kind of architecture design because at some point, you're gonna have
41:41 / 43:14
from transcript41:41 – 43:14
It's eight out of two fifty six. And there's different implementations for mixture of experts where you can have some of these experts that are always activated, which this just looks like a small neural network. And then all the tokens go through th
45:44 / 46:56
from transcript45:44 – 46:56
...the generalist foundation models will just get so good at a global set of intelligence that they're going to win these categories where, all you have to do is basically synthesize information and then make a recommendation. And then really where the
13:22 / 14:33
from transcript13:22 – 14:33
...it used to be generalist to work. Like, a lot of the market before the model started to get better was leveraging talented international lower cost labor
9:43 / 11:00
from transcript9:43 – 11:00
So there's there's two main techniques that they implemented that are probably the majority of their efficiency, And then there's a lot of implementation details that maybe we'll gloss over or get into later that sort of contribute to it. But those t
35:10 / 36:15
from transcript35:10 – 36:15
...model is good for a generalist purpose. Or you can, like, vaguely understand different niches. But that's true now. Right? Yeah. Like, don't you think in ten years a generalist model would be everything? Exactly. So when it especially when it comes t
20:35 / 22:14
from transcript20:35 – 22:14
...from, like, generalists, where you are predominantly leveraging, like, inter international
18:54 / 19:57
from transcript18:54 – 19:57
Do you think so? Or do you think it honestly just builds up habit? You get used to talking to someone who's not a human. I would very much like it to be the former. That is going to be ultimately up to the users. What do you think is the hardest prod
11:11 / 12:15
from transcript11:11 – 12:15
We don't usually come to engineering summits because we usually go to vet summits and, like, talk to the they're they're, like, you know, they're they're they're literally the I'm sure it's a massive pain point. They're willing to pay a lot of money.
34:22 / 35:30
from transcript34:22 – 35:30
Have a podcast?
Run your podcast through Clypt
Get ranked clips, hooks, and ready-to-post copy from your own episodes. Free to try.