Shirayuki2
new account of Shirayuki, lost old password
User ID: 4180
What does anyone do with it?
Right, but my point is that the set of actions available to random civilians is far smaller than the set of actions available to the real power players of the world.
As a civilian, I basically do agree that if you believe this thesis, then all you can really do is take less risks, invest some of your money and otherwise live your life as is.
As a nation state or even just as someone with meaningful connections, there's obviously many things that could be done if you believed such a thesis.
Are key frontier lab employees being assassinated to slow down the US lead?
Are nuclear missiles being pointed at DC as a brinksmanship move to buy France/Russia et al entry into the new superintelligence world order?
Is China centralizing all compute into ASI training runs to win the AI race at all costs instead of slopdramamaxxing?
Is Anthropic suddenly dedicating 100% of their compute to training for the final push with the unrestrained financial backing of the United States instead of enterprisemaxxing?
Hell, is METR willing to unredact their RSI report and burn those bridges to whistleblow if This Is It, we're close to superintelligence imminently?
I mean, I'm pretty bullish on AI now in the "it can probably do most generic white collar work tasks" sense, but fundamentally nobody is acting as if they truly believe anything approaching superintelligence is imminent.
I don't agree with Aschenbrenner on most things, but he was fundamentally correct that if AI is truly such a big deal it would be a matter of national security much more significant than the Manhattan Project. Instead, Anthropic spends most of its time grinding enterprise sales and Bytedance is introducing new and innovative ways to generate werewolf smut dramas and sell clothes to old ladies.
Two different observations:
a) The three groups on the planet with the most information with regards to AI progress in no particular order are the market, nation-state intelligence agencies, and frontier lab companies. Are any of these entities actually acting as if superintelligence is imminent?
Why is the AI trade starting to crack if there's a 75% chance of imminent superintelligence that presumably spikes productivity unbelievably or kills everyone? Are AI stocks really priced in any ballpark of correctness if that's an accurate model of reality?
Is China acting as if though AI is a world-ending existential threat? Why are there massive amounts of chips stranded under an import ban in Hong Kong, and why is Bytedance being allowed to spend massive amounts of China's compute on generating microdramas instead of AI 2027's takes on Chinese AI lab nationalization that are aging like milk?
Are the Israeli's, the Russians, the French acting as if the end of the world is imminent and readying their nuclear weapons, or are they busy pissing resources in their own backyards as if nothing will happen?
Why is Anthropic busy spamming enterprise integration tooling like Claude Meetings and hiring FDE's enmasse to do the boring work of integrating Claude into enterprise workflows if Claude 8 is going to eat the world anyway? Are a majority of Anthropic employees being moved to military bases and drawing down their retirement funds like RAND employees in the Cold War, instead of working from home and lining up to buy nice Bay Area homes after the IPO pop? Is there any concrete whistleblowing instead of METR's "self-reported Slack polls"?
I suppose to me it's a matter of watching the feet instead of the mouth. It's not that I think the concept of superintelligence is totally incoherent or even that short timelines are impossible, but it seems to me that right now it's still mainly internet bloggers getting worked up on thin evidence, while all the movers and shakers of the world, or indeed even most of the people that espouse such viewpoints, don't truly act as if they expect such things to happen.
b) Why does everyone who talks about AI need to include digs about how the sheep are so unaware about AI? Not to call you out specifically, but I find it just so tedious how everyone talking about AI feels like they have to be on the right side of history at the expense of the unpilled proles.
Realistically for most people it just doesn't matter whether they're "prepared" or not. Either AI is a normal technology and people will figure out how it slots into their lives over time as it diffuses into the world and into their lives as needed, or it's the singularity and 99.999% of people will have absolutely no way to affect the outcome regardless of how much they monitor the situation. The harness-pilled SF tech bro will be in the exact same position as the accounts payable griller in the end if such outcomes really do happen.
- Prev
- Next

To abuse the Manhattan Project metaphor again, the smartest physicists in the world in the early-mid 20th century all knew very well that the atomic bomb was viable, and hence once this was known to the government, it became a top military priority to own the bomb at all costs, with physicists going dark and drafted to Los Alamos to ensure their safety and secrecy as a matter of national priority instead of podcasting and job hopping.
I grant that France and Russia are basketcases now that perhaps can't muster that sort of state capacity, but the point is that while governments certainly don't hire the highest g researchers, they are not so incompetent that they would miss a development significantly more consequential than the nuclear bomb; Dwarkesh Patel does not have any information that intelligence agencies do not.
Personally, I'll start worrying when frontier lab researchers start vanishing, and Seedance goes down for maintenance indefinently.
More options
Context Copy link