cubefox

11 ms·

1.
Dan Selsam (OpenAI Researcher) Personal Statement on AI Risk(docs.google.com)
3 points by cubefox 2d ago|3 comments
2.
by cubefox 3d ago
> Any interruption or automation to what Im currently doing will just open the door to exploring new and different things. But that's the rub: At some point, possibly sooner than most people think, there won't be anything left
3.
by cubefox 4d ago
> Yet that journey is funded by society (esp. for domestic students) and society benefits from the results. The average person benefits when progress is made. Society doesn't benefit from results in research mathematics because it m
4.
by cubefox 4d ago
It's a graphical problem. Inverse kinematics is purely a graphical technique. It has no effect on gameplay. That doesn't mean it's irrelevant, except insofar graphical fidelity is irrelevant in general.
5.
by cubefox 4d ago
I was recently wondering why foot sliding is still a common problem in fighting games. Surely with Inverse Kinematics it should be possible to fix this problem by now. I have the suspicion that people who are accustomed to fighting games ar
6.
by cubefox 5d ago
Where is the difference?
7.
by cubefox 5d ago
Doing the last step of solving an open problem is the yard stick. But not for much longer: https://davidbessis.substack.com/p/the-fall-of-the-theorem-e...
8.
by cubefox 5d ago
> Dr. Tao said the same thing. Apparently he has since changed his mind.
9.
by cubefox 6d ago
> Surely some understanding of the lean proof is required, to make sure it proves what it claims to prove. Yes: > The only way the Lean proof could still be wrong is if the conjecture was formalized wrong via misleading definitions (i
10.
by cubefox 6d ago
So an LLM (or more realistically, a huge swarm of agents) should check his work, find a mistake or gap, or, if there is none, provide a formal verification.
11.
by cubefox 6d ago
> Not necessarily applied to OpenAI's solution to Navier-Stokes, but what happens if and when an AI genuinely appears to solve an extremely difficult problem but humans cannot independently verify the solution because understanding
12.
by cubefox 7d ago
It was just chosen earlier.
13.
by cubefox 7d ago
> So far all the foldables I've seen get destroyed quite fast. I doubt you did see any from the last few years.
14.
by cubefox 7d ago
They are addicted to teleprompters. Steve Jobs didn't use them.
15.
by cubefox 7d ago
I'm at a loss why anyone thought this asymmetrical design looks good.
16.
by cubefox 7d ago
What's your take on the aspect ratio question?
17.
by cubefox 7d ago
This article is not up-to-date. There have been various benchmarks (some of which published and acknowledged by OpenAI, see the charts in this thread: https://xcancel.com/tomekkorbak/status/2095596839886274689 ) sh
18.
by cubefox 8d ago
These three forms are trained in order to approximate the ground truth, and they perform significantly better than their non-neural counterparts which do the same thing but worse. Moreover, neither of them is a pure postprocess, since they
19.
by cubefox 9d ago
This should have happened years ago.
20.
by cubefox 10d ago
> Well I think the point still stands that a superintelligent AI would still be constrained to the digital realm. No it doesn't! As I said, it could control robots, just as animals (including humans) can control their bodies. It wou
21.
by cubefox 10d ago
> >a superintelligent AI will inevitably destroy humanity > Can someone please explain to me how an LLM is going to "destroy humanity"? He said "superintelligent AI", not "LLM". LLMs will for sure help
22.
by cubefox 10d ago
Perhaps too many are aware, because the more people fix the auto-editorializing manually, the less inclined Dang will be to eventually see a reason to remove it after all.
23.
by cubefox 10d ago
Many subject matters are initially so confusing and inaccessible that I have often wished for an audio-visual video lecture series in the following format: Part 1: Music theory explained in just 5 minutes Part 2: Music theory explained in j
24.
by cubefox 11d ago
I first thought it was a neat idea, but the exception for his shop makes the closure look a little hypocritical. "This site is closed on Sundays -- except the shop of course, the actual type of thing that is usually open or closed in t
25.
by cubefox 11d ago
Exactly. The WMF spending growth is completely decoupled from its effect on Wikipedia. This only works because the donations pop-up makes people think they are donating to Wikipedia, not to a foundation which mostly does totally unrelated s
26.
by cubefox 11d ago
Otherwise known as a "zero-day exploit".
27.
by cubefox 11d ago
"only"
28.
by cubefox 11d ago
> That's a threat to civilization. No it's not. Even if it were, they never said it is a threat to civilization.
29.
by cubefox 12d ago
They never said Mythos was an imminent threat to civilization. You are constructing a straw man.
30.
by cubefox 12d ago
> Anthropic mostly did it to themselves That is absurd, the US government was mainly at fault, not Anthropic.

More ›