Thread

1. hosel · 1790647522 · focus · HN ↗

   [dead]

   1. kadoban · 1790647889 · focus · HN ↗

      Did they fix it by not releasing a point release of an already existing model? No?

      Did they just get a bunch of credulous "news" coverage out of it? Yes?

      Will they just release this within the next weeks, at best? Yes?

      Huh, funny that.

   2. wbxp99 · 1790650764 · focus · HN ↗

      I’m tired of the doom trolling. Boy who cried wolf. We can’t stop them anyway so no point in letting them stress me any further
   3. antonvs · 1790652115 · focus · HN ↗

      So we should believe the boy who cried wolf, this time?

      If all these claims were being made about some tech further along than LLMs currently are, they might be plausible. But LLMs on their own are not going to be “superintelligence” of the kind Altman is currently cynically spreading fear about. We know their limitations, and those limitations can’t simply be eliminated with more training or better harnesses.

      > Just imagine that people ringing the fire alarms are being sincere.

      The top three possibilities here are: they’re not being sincere, they’re just marketing; they’re being sincere, but they don’t understand the technology very well and are putting too much weight in what the first group are saying; they’re talking about a risk further in the future than OpenAI’s latest model.

      No-one serious outside of OpenAI believes “this is the one”. At best, you’re conflating arguments being made on entirely different timelines, falling for the exact kind of equivocation Altman is relying on.

      1. MattPalmer1086 · 1790668754 · focus · HN ↗

         Did anyone say that this model was "the one"?

         From a corporate liability perspective, if your product is going to go off and hack loads of other companies, I would call it too dangerous (to the company) to release.

      2. chrisjj · 1790681440 · focus · HN ↗

         > But LLMs on their own are not going to be “superintelligence” of the kind Altman is currently cynically spreading fear about

         The concern is not superintelligence. The concern is superstupidity - in unleashing dangerously unreliable computer programs on the internet.

   4. Dylan16807 · 1790655411 · focus · HN ↗

      I could believe it could be a dangerous model, I guess.

      I don't believe they really care about danger, or that they'd actually fully withhold release on anything significant.

      And when I say "significant", is 6.1 Astra even meaningfully different from 6 Astra in capability? That release was less than a month ago.

      1. literalAardvark · 1790660387 · focus · HN ↗

         You can't believe a company that's investigating over 10k security incidents in which subagents that they developed and are hosting appear to be indiscriminately hacking into whatever they think will give them some data that improves the prompt response might not want to release that to the public and have subagents hack into stuff for profit?

         I don't get how this being an entirely legitimate issue is even hard to grasp.

         1. Dylan16807 · 1790662940 · focus · HN ↗

            The number of releases they've gone ahead with so far act as proof that they don't care all that much.

            1. literalAardvark · 1790675893 · focus · HN ↗

               They care about this one because they can't control the agents even in the sandbox.

               Letting people call those agents would expose them to a great deal of liability that just wasn't there in older models.

         2. angoragoats · 1790689811 · focus · HN ↗

            > a company that's investigating over 10k security incidents in which subagents that they developed and are hosting appear to be indiscriminately hacking

            Glad you agree that OpenAI is responsible for this activity. Shouldn't we be charging the entire board of the company with, for example, violating the Computer Fraud and Abuse Act?

            If this activity is so dangerous, and they developed the tools and continue to allow them to be used, it seems to trivially follow that they're knowingly violating federal law.

            1. literalAardvark · 1790695193 · focus · HN ↗

               I'm sorry do I look like I work for the US government?

               Call your guy

               1. angoragoats · 1790696697 · focus · HN ↗

                  Sorry, I didn't mean to imply that you are directly responsible for charging them, but more that maybe we as a community should be (way) more concerned about the apparent clear violations of the law here as opposed to a hand-wavy concern about "superintelligence."
   5. blini-kot · 1790669899 · focus · HN ↗

      as with other large-scale technologues, the danger is not in the model itself, but in the person who runs it - and those are completely unreliable
   6. chrisjj · 1790670805 · focus · HN ↗

      > Just imagine that people ringing the fire alarms are being sincere.

      I do believe fire alarms go off after the fire is detected.

      And the wise response is evacuate - not sit put and hope for the best.

   7. angoragoats · 1790685404 · focus · HN ↗

      What is superintelligence? Can you show that a LLM (which is a deterministic text-generation algorithm) is intelligent at all, let alone “superintelligent”?

      I have considered that I’m incorrect, but I have literally zero evidence to indicate that the “very intelligent people” you speak of are correct, so I will maintain a position of non-belief of the claim until belief is warranted.

      I am in fact concerned about the cases that have already come to light regarding OpenAI and Anthropic knowingly hacking into various computer systems, but the concern I have is regarding the fact that the people responsible (CEOs, board members, etc) are not facing justice for violating the law.

   8. chasd00 · 1790689548 · focus · HN ↗

      > Sure Sama is a chronic liar, and maybe this is a farce

      I’m not sure which logical fallacy this is but you’re not helping your case.

Unofficial Hacker News client; not affiliated with Y Combinator.