dartos

37 ms·

601.
by dartos 2y ago
Could just as well be a typo. I’ve never heard anyone call diagnosing “sniffing”
602.
by dartos 2y ago
> which we did not and could not design, plan or engineer, but which emerged We literally designed, planned, and engineered the environment and mechanisms which created those weights. It’s just code. We can train models by hand too, it’d
603.
by dartos 2y ago
A constraint is still a constraint
604.
by dartos 2y ago
> The problem is, a computer has no idea what "improve" means unless a human explains it for every type of problem Not currently. I don’t really think AGI is coming anytime soon, but that doesn’t seem like a real reason. If we
605.
by dartos 2y ago
Really bullish on AI, huh? Designing an engine? “[snuffs] out common diseases”? Solving SEO? Idk cotton, let’s see how this plays out.
606.
by dartos 2y ago
I’ll believe what they say about O3 when it starts seeing use in the wild.
607.
by dartos 2y ago
Except also energy constraints. But I agree, there’s no reason to believe humans are the universal limit on cognitive abilities
608.
by dartos 2y ago
a little melodramatic at the end. As with every openai release, I’d take it with try a grain of salt until it sees wide public usage. So far every iteration of gpt since 4 has been pushed as a replacement for developers, but I have yet to f
609.
by dartos 2y ago
I don’t think that’s true for AGI. AGI is the holy grail of technology. A technology so advanced that not only does it subsume all other technology, but it is able to improve itself. Truly general intelligence like that will either exist or
610.
by dartos 2y ago
Nix is really weird and ugly tbf. It’s great at being a declarative configuration language though…. Even if it’s ugly.
611.
by dartos 2y ago
I mean… you still pay at a cheap buffet.
612.
by dartos 2y ago
I don’t really think you’re understanding my argument…
613.
by dartos 2y ago
> If the average user is given unfettered access to the entire source code of his/her favorite app, does he suddenly understand it ? That seems like a ridiculous assertion. And one that I didn’t make. I don’t think when we say “we u
614.
by dartos 2y ago
That makes a lot of sense. Binary success criteria has very little room for bias.
615.
by dartos 2y ago
> We don't understand how they work in the sense that we can't extract the algorithms they're using to accomplish the interesting/valuable "intellectual" labor they're doing. i.e. we cannot take GPT-4 a
616.
by dartos 2y ago
I gave it a fair skim, but I didn’t really feel like it refuted what I said. Is there a specific section that comes to mind?
617.
by dartos 2y ago
The intuition I’ve been building around LLMs is that they’re like search systems. They encode their training data in their weights and do a kind of search on it given a prompt. If leetcode problems are in their training data, then they’ll o
618.
by dartos 2y ago
Your whole argument falls apart at > We don't understand how they work, because we didn't build them. They built themselves. We do understand how they work, we did build them. The mathematical foundation of these models are sou
619.
by dartos 2y ago
Overfitting can be caused by a lot of different things. Having an over abundance of one kind of data in a training set is one of those causes. It’s why many pre-processing steps for image training pipelines will add copies of images at weir
620.
by dartos 2y ago
I feel like I have the worst of both worlds. I can’t work effectively without a todo list, but I have extreme trouble keeping with the habit.
621.
by dartos 2y ago
That leads to overfitting in ML land, which hurts overall performance. We know that unique data improves performance. These LLM systems are not students… Also, which students graduate and are immediately experts in their fields? Almost none
622.
by dartos 2y ago
Not as sexy to investors though
623.
by dartos 2y ago
> Productivity gains increase the standard of living for everyone This just isn’t true, necessarily. Productivity has gone up in the US since the 80s, but wages have not. Costs have, though. What increases standards of living for everyon
624.
by dartos 2y ago
You lost me at the end there. These statistical models don’t generalize well to out of distribution data. If you accept that as a fact, then you must accept that these statistical models are not the path to AGI.
625.
by dartos 2y ago
I’m curious how, if at all, the plan to get around compounding bias in synthetic data generated by models trained in synthetic data.
626.
by dartos 2y ago
That’s an interesting idea. What if OpenAI funded medical research initiatives in exchange for exclusive training rights on the research.
627.
by dartos 2y ago
They are. They’re just really really good at it. That’s how ML is. “Prediction” is just another word for really really good guess.
628.
by dartos 2y ago
Scammers can also buy aggregate information, just like advertisers
629.
by dartos 2y ago
They weren’t before?
630.
by dartos 2y ago
If only people know of the standards

More ›