Polymarket this is so close to being useful
Please make two CONDITIONAL markets for NVDA price: one if Jensen stays, the other if Jensen out
Then you (& shareholders) can see if the markets think ousting Jensen will raise or lower the price, before actually deciding
75% chance Jensen is out as CEO this year.
poly.market/2gD4gqP
Either you understood how Duck Hunt on NES worked, or you were gonna be confused anyway
iiuc CDT is fine at voting, it's just sometimes computationally confusing because the probability of being pivotal is tricky
but since UDT can be implemented via CDT+SIA (from @jessi_cata), you *can* turn out to vote using CDT just as well as using UDT
lesswrong.com/posts/5bd75cc5…
2
11
1,090
This post was 17 years ago when Dario was a grad student. He did end up giving the child vaccine charity $10,000 and he was a student not a billionaire yet. In the post he writes about how it's worth doing but still emotionally hard to put numbers on things like this
Readers added context they thought people might want to know
The excerpt is from Dario Amodei's 2010 GiveWell guest post as a Princeton physics grad student; he donated the $10k to VillageReach (child vaccines) and described the critical analysis as emotionally taxing.
blog.givewell.org/2010/06/03/my-…
nytimes.com/2026/11/05/tec…
*taps mic*
people wanna be told how to be artists themselves more than they wanna see anyone else's art
there's a joke in photography spheres even before the internet that if you want to make money in photography, your best bet is to make a paid workshop called "how to make money with photography", where you explain how to sell paid workshops instead
1
14
592
Tentative Plan A scenario:
- Solve deep interpretability for current-ish models and create a DEEPLY TRUSTWORTHY but only-slightly-superhuman agent (no strong-super generalization yet)
- Etch it into some chips, temp ban all other models
- Hopefully it can solve superalignment
Basic idea is just Do Things That Don't Scale
This kind of alignment very likely WOULDN'T persist across self-improvements or work at all on a sufficiently intelligent AI
BUT could it get us a *somewhat* superhuman researcher that is at least as trustworthy as a human upload?
1
26
767
Hooray I love standard time, the sun was coming up way too late this past week, now morning is saved ☀️
3
23
798
Also seeing people say, "what if we prove P=NP but it's not constructive!"
But if you prove P=NP it will immediately be constructive, because Levin's universal search algorithm solves 3SAT in polynomial time iff P=NP. (But it's galactic so then we are in this situation again)
Seeing a lot of talk this week like,
"Maybe P=NP but it's a galactic algorithm so it doesn't really matter"
I think Scott A (the complexity theory one) has the right take on that -- it wouldn't answer our actual question!! We'd just need a new conjecture, like EasyP ?= HardP
.......... PERIODICALLY ???????????
I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model.
Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys Intelligence Index at 59. It also leads across key cybersecurity evaluations scoring 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching.
We’re seeing significant improvements in long-horizon tasks even compared to Argon, and we’ve strengthened frontier safeguards.
As we did with Argon, we’re going to roll it out gradually + responsibly. Some trusted cyber defenders will get access starting this week. We'll be expanding it to more partners through our Fairwind Program, ahead of wider availability soon, so hang tight!
I think it's a fair take on the current robot capabilities, but also...
2024: LLMs can't even count the R's in raspberry
2026: *gestures around*
2026: robots can't change batteries
2028: hmmm
I see how some people can watch flashy robot demos on the timeline and think with a shiver that the robot apocalypse is around the corner, but reality check:
There is no robot in the world that can change the batteries in a jack-o-lantern.
If you have a robot friendly door handle it might be able to walk outside, pick it up, and carry it back inside. The robot could open the drawer with your batteries in it, but then things get tough.
Opening the package of AA batteries?
Extracting a precision screwdriver from its case?
Unscrewing four tiny screws that require a hard press and twist?
Popping batteries out of the case?
Yes, you could build and train a specialist robot to do that today, but there is a very long way to go to replace the full range of human capabilities. Near term doomer scenarios should stop at “secretly controlling humanity” instead of “wiping humanity out” 😀
ppl in 2016: if your arguments are so rational why arent you winning
ppl in 2026: beware their arguments because rationalists are SECRETLY the most POWERFUL group in the multi bajillion dollar AI industry
ppl* never: object level takes on the arguments
* except a few cool ppl
