undefined

points

[-]

I elaborated on my use of "dark forest" in another reply. We're headed for a dark forest--not amongst interstellar civilizations, but in intellectual work.

I agree that we have not solved the Fermi paradox; I disagree that comments highlighting immature behavior from people who wield enormous power in our world are unproductive.

by AlexErrant16 hours ago|

parent|

[-]

This clarification substantially changes the flavor/nuance of your OP; may I suggest an edit (assuming the locktime hasn't passed)?

Separately, I disagree that intellectual work has ever been free of "dark forest"-style secrecy. Scientists everywhere have worried about being scooped; AI just magnifies that (as all tools have; e.g. Leeuwenhoek lenses).

And thirdly, if you want to make a stronger case for "I feel even less confident in them as a team to be shepherding this much capital and compute", you should give citations and arguments. From what I've seen, there's drama, it's much OpenAI trying to avoid scooping, and Tristan being stuck in a game of telephone, and Levent being incommunicado.

If you have a better analysis, you should say so instead of being vague.

by recitedropper15 hours ago|

parent|

[-]

I don't comment on HN much, and I don't really expect HN comments to hold to rigorous standards. This forum is more casual than other places on the internet where people expect heavy citations. I also wasn't expecting this to blow up, although it is interesting to see that a lot of people react to this announcement with a negative sentiment.

I appreciate your upholding of ideals, and since I respect that, I will honor with final replies:

1. Locktime has passed.

2. Yes, intellectual work has always had elements that incentivize secrecy. If you want to say we were already in a "dark forest", so be it. My suggestion is that the multiplier AI adds to the possibility you get scooped is a step-change, and therefore we now enter a new "dark forest".

3. This is a big thread, and the twitter antics are well-documented, so I would assume someone else has cited them. If not, I think most are aware at this point that the online antics of AI researchers, especially when announcing or citing mathematical advancse, have regularly been childish.

by AlexErrant15 hours ago|

parent|

[-]

2. Fair, and further I would agree that OpenAI not knowing if prior user prompts were part of training data is concerning and will only lead to more secrecy.

3. ctrl-f "x.com" in this thread only yields https://x.com/sama/status/2097385167002415140 https://x.com/SebastienBubeck/status/2097379411691516310 https://x.com/OpenAI/status/2097375276384567642

and frankly, I'm unwilling to give Elon any more traffic to dig up drama that ultimately doesn't matter. I'm not seeing anything especially childish, but y'know... I'm not sure I care.

by zem13 hours ago|

prev|

[-]

the projectnash link claims it's mathematically valid, the noahpinion link says that it's invalid and has a marvellous proof that the non-walled section is too small to contain.

by AlexErrant12 hours ago|

parent|

[-]

Ah derp; that's what I get for moving too fast. Genuine thanks for calling me out on my bullshit. (and this is also why I prefer auditable citations instead of casual "my reading of twitter is...")

I retract the projectnash citation; I grabbed it from the Cool World's youtube description, thinking it was a blog version of the video. It was not. I suggest watching the video instead.