Legend2440

3 ms·

31.
by Legend2440 1mo ago
They explain why here: https://openai.com/index/how-two-settings-tripled-our-arc-ag... TL;DR all the other models are being crippled by limitations of their harness. >First, we noticed that after each game action, a
32.
by Legend2440 1mo ago
>This suggests that deeper isn't always better for looped transformers, which leaves me less worried about a race to the bottom toward looped transformers with hundreds of recurrent loops. I disagree with this. Deeper will always be
33.
by Legend2440 1mo ago
Greed and money aren't helping. There is no single cause for anything. But the fact that both liberals and conservatives think of the other as 'evil' is a large problem.
34.
by Legend2440 1mo ago
Well, historically, it's sort of worked out on its own. Automation displaces workers but also creates new jobs in other sectors. We have maintained full employment at higher wages. That said, some displaced workers have experienced sub
35.
by Legend2440 1mo ago
It's not going to end humanity. We're going to use it to do incredible things that would otherwise be forever beyond our ability - a post-work economy, science fiction technologies, advanced mathematics, etc.
36.
by Legend2440 1mo ago
>Anybody who disagrees with the ban on superintelligence in principle, maybe not in practice, is a traitor to humanity Oh, so anybody who disagrees with me is evil? This is the kind of attitude that ruined US politics.
37.
by Legend2440 1mo ago
If people like him their way, we'd still be farming with shovels and weaving our clothes by hand. Banning automation is shortsighted and stupid; you refuse to let people increase the size of the pie because you're worried it'
38.
by Legend2440 1mo ago
Honestly, yeah. I wouldn't eat a wild mushroom. The ones in the store are good enough for me.
39.
by Legend2440 1mo ago
You should never trust any image-based method for mushroom identification. It is simply not possible to identify mushrooms from image alone, no matter how good your LLM is.
40.
by Legend2440 1mo ago
Is it possible to reliably (>99%) identify mushrooms from a single image, with any method? All image datasets have an intrinsic minimum error that comes from the limit of information in the data, regardless of what model or method you
41.
by Legend2440 1mo ago
There are like three dozen companies selling similar services for generating AI training data.
42.
by Legend2440 1mo ago
They talk about that in the article: >"A centrifuge plant does one thing, costs billions, and takes years to stand up. Our machines cost a few hundred thousand dollars, produce material within months, deploy anywhere, and are able t
43.
by Legend2440 1mo ago
>Private enterprise enriching uranium is a seriously worrying development... Or seriously exciting. We might actually get affordable, clean nuclear energy this century.
44.
by Legend2440 1mo ago
>And I thought enriching uranium is something that counties with nukes go to extreme lengths to make it not accessible. It's not that difficult from an engineering perspective, the tech is almost a century old. It's just that
45.
by Legend2440 1mo ago
Apple is working on their own system, which is expected to launch with IOS 27: https://www.macrumors.com/2026/08/10/ios-27-apple-reference-... >Images captured with an opt-in Reference mode can be authenti
46.
by Legend2440 1mo ago
My bet is they do considerably worse. Digital cameras are not designed with security in mind. Arbitrary code execution has been achieved on many DSLRs and there's even been open-source firmware projects for some.
47.
by Legend2440 1mo ago
>Do you think that 100% of people whose job is displaced find another job in a reasonable amount of time and on the same order or pay / status? Do you not realize that the only reason we are not impoverished dirt farmers right now i
48.
by Legend2440 1mo ago
Did you know that not everyone who disagrees with you is a bot? Some people are genuinely excited about AI.
49.
by Legend2440 1mo ago
>Does anyone remember the awe that was felt the first time we encountered an iPhone. Or an iPad? Yes, and when I first saw LLMs and image generators I experienced a similar feeling of awe. We haven't had anything like this in decade
50.
by Legend2440 1mo ago
A lot of those environmental claims are not accurate, or substantially overblown. Heavy metals in the water table is a new one I haven't heard before, and I'm skeptical. Most of this is motivated reasoning by people who don't
51.
by Legend2440 1mo ago
No thanks.
52.
by Legend2440 2mo ago
Most peptides don't seem nearly as dangerous as insulin, luckily. Injecting insulin without medical supervision is a good way to give yourself hypoglycemia and die - blood sugar regulation is not something you want to mess with! People
53.
by Legend2440 2mo ago
You also wouldn't want to randomly inject yourself with insulin outside of medical supervision, because it can kill you pretty quick. Injecting poorly-studied grey market peptides seems like a bad idea.
54.
by Legend2440 2mo ago
Seattle did this. It didn't really work out for the drivers: https://www.king5.com/article/news/local/seattle-rideshare-d... There were two problems: 1. Increased wages got passed along to increased pric
55.
by Legend2440 2mo ago
>In a reasonable society, most individuals' earnings should go up all the time. I don't think this is a reasonable expectation at all. In the absence of economic growth I would expect the average individual's earnings to b
56.
by Legend2440 2mo ago
Coding agents are increasingly popular right now, and they can use millions of reasoning tokens per hour. Basic ChatGPT queries hardly use any tokens by comparison.
57.
by Legend2440 2mo ago
I don't think your premise is true. Remember John Henry, worried that the steam engine would put him out of a job drilling holes in rocks? https://en.wikipedia.org/wiki/John_Henry_(folklore) Ludditism has a long h
58.
by Legend2440 2mo ago
That seems really speculative. I think you're assuming the maximum possible evil intent for no reason, mostly because you hate AI and anyone involved in it.
59.
by Legend2440 2mo ago
>What worries me about the trends is inevitable sanitization of content or straight out falsification. That sounds really speculative, and not inevitable at all. I think you're just trying to invent things to be worried about becaus
60.
by Legend2440 2mo ago
>even if there was enough power there’s no use case for this much spend According to them, demand for tokens already exceeds supply. Coding agents are very popular but burn through millions of tokens per hour. The reasoning models that h

More ›