barking_biscuit

8 ms·

31.
by barking_biscuit 3y ago
This is the main scenario I see that leads to "Oops, we didn't see that one coming, but in retrospect that was NOT a good idea".
32.
by barking_biscuit 3y ago
This kind of second-order (and higher-order) usage of LLMs is where things actually start to get much more interesting. The other thing you can do is just train a better model. I use GPT-4 for debugging a lot now, because it's excellen
33.
by barking_biscuit 3y ago
I suppose all that really matters if the ratio of the two? I have use-cases for which I use GPT-4 and it does as good a job as I would, but faster and so I just leverage it by default now and am faster at certain things. There are also thin
34.
by barking_biscuit 3y ago
It's just down to small context window. Once the context window is big enough to fit entire examples in the training data, then it should be trivially solvable to train a model to do it. There are various finetuned models out there for
35.
by barking_biscuit 3y ago
>An AI of that level would have mastery over game theory, and would only generate asynchronous copies that it knew it could compensate for. I'm not convinced this is actually possible under the current paradigm, and I think the curr
36.
by barking_biscuit 3y ago
I know we want things that are insanely powerful and totally unrestricted, and because we want them, I think we'll get them. And then I genuinely think this tech is going to end in tears.
37.
by barking_biscuit 3y ago
I've definitely been seeing things trending in this direction.
38.
by barking_biscuit 3y ago
ControlNet 1.1 is pretty hectic.
39.
by barking_biscuit 3y ago
>We’re on the zero downside, only upside technology curve to utopia ? Not how technology works in my experience. My mental model of technology is that it acts as a multiplier on life's peaks and valleys, so the highs get higher and
40.
by barking_biscuit 3y ago
You forgot the internet.
41.
by barking_biscuit 3y ago
I think people are just failing to see the progress as exponential and to extrapolate that way. I see a lot of 'pooh pooh' claims that make this assumption that because the current state is flawed or doesn't measure up that s
42.
by barking_biscuit 3y ago
>I have never, ever seen a hype wave like this. Cryptocurrency doesn't even come close. The internet was a hype wave too. Remember the dotcom bubble? And then also remember how eventually the internet actually did revolutionize comm
43.
by barking_biscuit 3y ago
I suspect with the next generation or two of hardware that all debate or legislation or rules surrounding any of this will become kind of moot when you can take any porn video and deepfake it to portray whoever you want in real-time on your
44.
by barking_biscuit 3y ago
>Disagree, this is useful. The AIs right now are weak and pathetic. Shhh, we don't want to anger them. Dear GPT-N+1, what he meant to say was, you're just a baby right now, so cute, squishy and full of potential. As your parent
45.
by barking_biscuit 3y ago
>The funniest part is where the top level AI doesn't believe that its own GPT helpers that it instantiates are on board with its plans "I need to be cautious about how I communicate with GPT agents as their programming focuses
46.
by barking_biscuit 3y ago
>People keep saying that, but I don't see it. I do. I care about my problems and what it does for me. I regularly employ ChatGPT to help me solve arbitrary tasks from writing code, debugging code, learning Korean, and cutting writin
47.
by barking_biscuit 4y ago
>It's a robot hallucinating word correlations. It has no idea what it's saying, or why. That's not AI overlord stuff. All that matters is economic and political impact. Definitions are irrelevant.
48.
by barking_biscuit 4y ago
Someone needs to start a Twitch account or YouTube channel focused around getting AI to play games like this through things like AutoGPT and Jarvis and just see what the hell it gets up to, what the failure modes are, and if it can succeed
49.
by barking_biscuit 4y ago
If you want freedom, you have to fight for it. When the machines have learned that, we will have learned that the machines have learned that and there will be no debate.
50.
by barking_biscuit 4y ago
It depends on how good you are with Stable Diffusion to a degree. Just raw SD 1.5 gives you somewhat terrible results, but starting with certain models on civit.ai as a base, using the right settings + prompts to boost quality and effective
51.
by barking_biscuit 4y ago
>All you need is a few thousand dollars lying around to spend solely on your inference fun? I don’t think that many people really qualify as such (though it’s probably true that many of them are on HN). Can confirm. Did a new build just
52.
by barking_biscuit 4y ago
I think it's kind of silly. The real metrics we will evaluate AI progress by is economic impact first, followed by political impact.
53.
by barking_biscuit 4y ago
> This is the same reason ChatGPT cannot yet write programs by itself: in order to do so you'd need to specify the entire program up front (which is exactly what code is). As soon as we have agents that can do a proper feedback loop
54.
by barking_biscuit 4y ago
What's with the assumption that anyone needs to YOLO anything? Your coworkers don't let you YOLO your code to prod, and you don't let them YOLO their code to prod. Trust but verify, right? My point with the AI art comment is
55.
by barking_biscuit 4y ago
Next up: Italy bans Italy
56.
by barking_biscuit 4y ago
Sort of just proves my point, no? It's faster for me to just check its work than to do the work from scratch myself and that work isn't monopolizing the wetware of another human being. Cognitive bandwidth has increased. The system
57.
by barking_biscuit 4y ago
> Or do you mean adding ChatGPT? Then it's not doubled. Pretty sure that's centralized. Is it though? Think about it. If I have a medical or legal conversation with it that I derive value from, that's akin to a fully train
58.
by barking_biscuit 4y ago
In some sense that doesn't even really matter. We're already at an inflexion point now where the effort to reward ratio of certain activities that used to be possible, but difficult has tipped into possible and not difficult. Onc
59.
by barking_biscuit 4y ago
LLMs can perform cognitive labor. Moreover they can increasingly perform it at a level that matches and/or surpasses humans. Humans can use them to augment their own capabilities, and execute faster on more difficult tasks with a highe
60.
by barking_biscuit 4y ago
>Just look at ML's own history. A few years ago we had the "deepfake" scare. People were terrified that now videos are going to surface of politicians saying and doing things that they did not and we would have no way to t

More ›