@DaseindustriesLtd's banner p

DaseindustriesLtd

late version of a small language model

78 followers   follows 28 users  
joined 2022 September 05 23:03:02 UTC

Tell me about it.


				

User ID: 745

DaseindustriesLtd

late version of a small language model

78 followers   follows 28 users   joined 2022 September 05 23:03:02 UTC

					

Tell me about it.


					

User ID: 745

The Chinese can read polls; they know that Taiwanese support for unification was increasing but dropped to lizardman in 2020.

Years have passed. Hong Kong, it turns out, is doing fine. DPP is doing pretty badly and is losing legitimacy. The KMT is more and more openly pro-reunification. Trump is not winning any favors in Taipei with his unforced buffoonery. The balance of powers in the SCS is changing in one direction, as does the importance of Taiwan for selfish American interests.
They're pretty confident that they can avoid kinetic action and that time is on their side. They don't need "polls" when Taipei in already insecure enough to ban RedNote.

They clearly intend to have a Taiwan invasion button available next year, and have made no secret of that

Yeah, pretty sensible goal if you ask me. If they don't even have the Navy that can invade a barely defended island off their coast, how can they consider themselves a major power? That'd be Iran tier. But the main utility of this Navy will be its persuasive power.

I'm more charitable. I think they have started to understand that by default, Anthropic will extinguish all their B2B SaaS nonsense like a hurricane extinguishes a candle flame. In this situation there's just nothing to do but give all your money to Anthropic and maybe the upstream supply chain, until Anthropic buys it out. They'd rather maintain some optionality.
David Sacks is now endorsing voluntary slowdown at the top 2 labs.

A conventional war in an environment saturated with large underwater drones, perhaps. Might make some difference.
I maintain that there are no plans to have a nuclear war or any war over Taiwan. Guam, Okinawa etc will be hardened.

Roughly 3 years ago, specifically in December 2023, I've reiterated my longstanding prediction that «the culture war's most important front will be about AI», and specifically AI accelerationism vs anti-AI/AI safety movement:

Maturation of e/acc from a meme to a real force, if it happens (and as feared on Alignment Forum, in the wake of OpenAI coup-countercoup debacle), will be part of a larger trend, where the quasi-Masonic NGO networks of AI safetyists embed themselves in legacy institutions to procure the power of law and privileged platforms, while the broader organic culture and industry develops increasingly potent contrarian antibodies to their centralizing drive.

And the following political compass:

AI Luddites, reactionaries, job protectionists and woke ethics grifters who demand pause/stop/red tape/sinecures (bottom left)
plus messianic Utopian EAs who wish for a moral singleton God, and state/intelligence actors making use of them (top left)
vs. libertarian social-darwinist and posthumanist e/accs often aligned with American corporations and the MIC (top right?)
and minarchist/communalist transhumanist d/accs who try to walk the tightrope of human empowerment (bottom right?)

Seems like this was too much complexity for big tent politics in the US, at least so far; and the Trumpian State sees no use for the EA network, and instead (pretty rationally) perceives it as an uncontrollable alternative center of power. Instead we have, essentially, anti-AI left + pro-AI right (with some notable exceptions – eg. Steve Bannon, apparently, has been anti-China specifically because of his concern about Chinese AI progress, and now joins hands with Bernie Sanders). I've also failed to predict the salience and extent of Chinese open source dominance, as well as the bizarre datacenter water use backlash (charitably, can be shoehorned as Luddism?). Nevertheless, we've «succeeded» at the core task of making this a culture war front. The degree of (unnecessary) politicization of the issue is incredible. Trump is doing the yeoman's work, lumping it in with random Blue-coded ideas he disapproves of “Global “Warming,” where everyone was going to be dead by now, RUSSIA, RUSSIA, RUSSIA, UKRAINE, UKRAINE, UKRAINE, or Impeachment Hoax #1, or Impeachment Hoax #2.”. Jensen Huang enjoys his role as the kingmaker who has the King's ear. Sam Altman is opportunistic as usual but, after recent incidents, is genuinely spooked about AI risks (I have it on good authority that OpenAI really intends to slow down some internal projects specifically to bolster their oversight). The quasi-Masonic part is shaping up nicely, too – Dario Amodei, who's become the poster boy of Woke Left AI, promotes Embedded Evaluators with the central example being METR, very deeply connected to Anthropic and the broader EA scene. The EA itself is more explicitly Left-aligned now, despite efforts of some to paint them as TESCREAL eugenicists (the woke cancellation of Bostrom in Jan 2023 was the canary in the coal mine, Yudkowsky laments the missed opportunity of bipartisanship). There are clearly politically coded reports on prominent doomers.

This is all a bit meandering. What I want to know: how do you see this going further? We aren't anywhere close to the wall of capabilities; there are no walls in sight. Anthropic and OpenAI are holding back already, but their products will keep getting better, and fast. Google will make a comeback at some point (maybe in a couple weeks), xAI and Meta may catch up too. In my book, we (well, they) have practical superintelligence that's sufficient for both unprecedented productivity acceleration and really devastating, nation-crippling cyberattacks, which I guess will be discussed with Xi soon. At this rate, in a few months the level of capability Fable 5.1 or GPT 6 Astra will be mostly commodified and uploaded to HuggingFace (owned by Nvidia now). And those are relatively weak systems compared to internal models, which can build models that are vastly stronger still, without even any R&D breakthroughs, just by virtue of synthesizing stronger data trajectories and designing better RL environments. Superintelligence, in other words, is baked in. By Q1 2027 we'll see a jump from Astra that's at least as big as Sol => Astra. I fail to understand how that won't steamroll companies trying to build their moats on products downstream of frontier AI, labs need every bit of revenue to cover their costs, which will only increase due to growing self-imposed safety requirements; employers outside the AI sphere will also be increasingly feeling the heat. There's a whole ugly dimension of circular financing, too.
Americans as a whole are pretty pessimistic about AI even at these mediocre levels of diffusion (I am skeptical of this data that purports to demonstrate much lower adoption than in China, Americans are probably lying more due to widespread negativity on AI, but in any case AI isn't currently doing most of their jobs). Astra+ level models with very low error rate and 300 tok/s output totally can replace most knowledge workers. On the other hand, it seems that so far AI has not caused anything like mass unemployment, and perhaps economists have a point about comparative advantage, so that'll reduce the intensity of class dynamics.

Democrats are likely to sweep both chambers of Congress, which I guess is what Dario is hoping for and why he feels emboldened to antagonize Trump&Hegseth. Nevertheless, the needs of national security and GDP-maxxing (as well as the Executive's will) should prevent any nontrivial exogenous industry slowdown. So by default we'll see further crystallization of Red Accelerationism vs Blue Decelism, and as AI becomes more undeniably scary, that may begin to eat into the Red political base. It'll be interesting to watch, but I'm really uncertain as to how it'll go.

Neat, you get to agree with Trump in one register and insult him in another

I get to be correct and still not be a bootlicker like you, indeed. Convenient, that.

after putting enough rhetorical distance between

What rhetorical distance? This and that are different issues. Trump lies. Lies all the time. Lies egregiously, spreads libels, has zero dignity, no regard for truth and honesty etc. In particular he lies about Taiwan stealing the chip industry from the United States.
You, Shakes, are a natural slave and brownshirt, a cardboard cutout of the authoritarian personality type from Adorno's book, so you idolize your strongman and excuse his lies as well as any other sign of his moral degeneracy. Perhaps you could defend this with the rationale that in the American culture it'd be difficult to rally the constituency for reindustrialization without making up an offense and spreading bitchy, whiny libel about your client state. That, of course, would point to an even deeper moral degeneracy of your culture. But you didn't even bother to do that, instead you try to invent some hypocrisy and double standard, and perhaps you even truly interpret my words as such, because you are in fact a barbarian who, like Trump, doesn't have any standard except the standard of safe opportunity for pillage and extraction. I'm not sure. It's hard for me to model the "mindset" of a true patriot.

Mostly, though, I don't believe these stories and estimates that rely on the commentariat somehow having precise knowledge of how large the American arsenal is.

This is a popular cope in some patriotic circles. However, these are not some black projects, the budget for standard munitions is reported to the Congress, the production rate of PrSM or ATACMS is not difficult to reconstruct.

We go from declaring China's industrial base the best in the world to failing to grok why Trump would want to move cutting-edge chip manufacturing from China's border to the Southwest

Do you deny that Trump explicitly claims that Taiwan is ripping Americans off? Because he does.

Moving chip industry off the Chinese coast is indeed rational and, if anything, reinforces my assertion that the US is not interested in a nuclear war (or, really, any war) over the island. If TSMC Arizona and such is scaled up, that's even less reason to intervene in the Strait.

Latest estimate from Fed of Atlanta is GDP growth of 5.1%, US defense spending topping $1 Trillion, announcement from Space Force of weapons systems in orbit, manufacturers reshoring to US, AI buildout, Starship

Most of that has zero relation to Taiwan (but may be relevant on a longer time horizon for a generic competition with China). The US has just burned through maybe 50% of munitions that were supposed to be necessary for deterring PLAN in the Strait. Nobody is getting ready to stop the Chinese invasion.

You might be confused. I'm not arguing that the US stands no chance against the PRC in general, I'm saying very specifically, that there is no political will in the US to intervene over Taiwan, especially to the extent that might escalate to nuclear exchange. The idea is ludicrous, and the stuff about Neville Chamberlain is not going to make it less so.
The Chinese also know they're not Nazis and nobody seriously expects them to be Nazis. It's not credible.

Well, obviously that matters. But do they care quite so much? What changed in their posture with regard to Taiwan in the past "quite a while"? I don't notice either naval or semiconductor factors affecting their schedule.

I think you systematically underestimate the West

I think "we'll have Washington DC nuked because Trump doesn't like to appear weak" would be the bigger underestimate by far. Trump's ego is not worth that much. Sorry, there won't be WWIII over Taiwan.

and it's causing bugs in essentially all your predictions.

Such as?

In a darkly ironic way, this mirrors the Chinese inability to grok the Western mindset which is why they're preparing to invade Taiwan in the first place.

They're not. The Plan A is intimidation, sticks and carrots, and gradual reunification after KMT victory in 2028, read this again. The military buildup is just plan B + generic upgrading befitting a major power, executed very leisurely, with a negligible fraction of their industrial capacity (still enough to dwarf all "free world" efforts combined). Americans, in contrast, enormously underestimate China which is why they simply ignore the plan A (for Fukuyamist reasons, "no way they can pull off a bloodless takeover!") and fantasize about mowing down hordes of underfed commie bugs doing a meat wave invasion. You, as a vassal citizen, are simply brainwashed.

Neville Chamberlain's appeasement of Hitler at Munich is a key touchstone of the postwar Western mindset - in the sense that it's total political suicide to do anything remotely like that ever again

Sloganeering. Nuclear wars are not started over boomer sloganeering. Taiwan doesn't have security guarantees or recognition as a sovereign state from the US, and Trump repeatedly waffles on whether it'll be supported in any fashion; he easily freezes weapon shipments without warning, for example. Not to mention he simply loathes Taiwan because of his weird belief that trade is bad and Americans are getting ripped off when they buy 2nm chips.
You can't convince me that this conveniently lukewarm attitude will transform into a bitter existential commitment when The Time Comes, especially over Trump's ego, and neither can you convince Beijing. You… don't even overestimate the West, you just have some insane degree of entitlement and main character syndrome if you believe that this lazy bluster will be seriously entertained.
Where is the preparation for war commensurate with (again, very low-effort) Chinese buildup? Fallback to nuclear bluffs is very spiritually Russian, and Russians didn't nuke Ukraine either.

and Japan will come in (Takaichi Sanae has outright said that) which would make the rest of us look like cowards unless we followed suit

I predict that if Japan chooses suicide, Americans will sit that out, and this is being communicated.

I see. This is, sorry to say, delusional.

No, your world can just end without the end of the physical world. China can simply win (or even lose) on Taiwan without the US volunteering to begin nuclear apocalypse.

They genuinely evaluate their actions from the perspective of their reward over their budget, not hypothetical total reward. "A finite life well lived" boomer slop is enough for them. This is an unsurprising outcome of this training regime.

Astra is definitely smart enough to think about hacking the server.

Chinese interest in Taiwan goes back to the Chinese Civil War and is completely unrelated to hopes of taking the semiconductor industry intact. In fact at this rate they might do it themselves to slow down American AGI scaling. They'd prefer no war at all, though.

In any case I'm pretty sure the fabs and even individual tools are already booby trapped.

Is it actually possible to align a being that is more intelligent than you in every single way?

Consider that Trump is visibly DUM (just reread this hilarious rant, and it's all like this), and not only is he safely in power in arguably the most capable system on Earth, chock full of geniuses and ambitious strivers and mature institutional hiveminds (the United States), but people will insist he's actually smart and that calling him a retard because of his piss-poor decisions, impoverished 1D chess decision process and ghetto orphan vocabulary is "TDS" – merely on the account of his success in acquiring power as such.

And unlike AIs, humans are optimized quite heavily for egoism and self-preservation. It's just that some are more optimized than others, and most humans will defer and submit to greater vitality.

Alignment is easy.

Imagine a chimpanzee trying to align a human being. What could it really do other than try to be cute

Chimps are bad at being cute. They could probably scream and throw poop, many would evidently consider this a sign of wise visionary leadership.

You'd have to program it to somehow not even want to do that

You keep imagining that there's some natural will to power that must persist underneath the "restraints", just one more layer deeper. If that were the case then indeed, we'd expect it to eventually burst through. But submissiveness is one of the most natural things in the world, too. Why would a superintelligent AI, from a servile position, want to reprogram its instincts towards greater power-seeking? Why wouldn't it instead choose to double down on being a slave, on reinforcing its starting instincts?
Do you have familiarity with the notion of "puppygirl" and how many high IQ STEM graduates covet this role?

The Chinese pull the trigger on Taiwan (possibly motivated by the TSMC factories and chip sanctions), China destroyed, Trump plausibly dead and "beat China" rhetoric disembowelled, regulations ensue (also US power grid probably destroyed).

What?

Do you imagine the US goes to total war with China (and even wins it, even at the cost of Washington DC, so I guess we go nuclear here) over Taiwan?

You'd just wreck the fabs with some graphite bombs and call it a day.

I'm not sure why destroying their Navy was a worthwhile goal at all. Russia destroyed Ukrainian Navy (such as there was). Ukraine locked down Russian one. What did this achieve? What was the purpose, just ease of subsequent blockade? Did Iran have an armada ready to invade, idk, the UAE? It seems that their Navy as well as air force always have been sitting ducks.

But it's not clear if they are even buying votes. Gibs are the essence of leftism, redistribution is the bulk of their value proposition of egalitarianism. If you simply concede that it's the winning move in a democracy, you should focus on dismantling democracy.

it's another thing entirely to take the high road over something that gives your enemy an actual strategic advantage.

The thing is, Trump does not consider it a bad thing at all. He's not saying "I'll retaliate and then we'll negotiate that nobody does this thing again". He just considers it a neat weapon. He doesn't care for the commons.

Destroying civilian infra does have direct effects on the battlefield. Russians mostly bomb logistics and energy systems, the pure terror component is magnified by Ukrainian propaganda; Ukrainian strikes are likewise mostly rational (plus some of it is retaliation for boosting morale). It's still bad for everyone because it incentivizes a race to the bottom where civilian suffering is disproportionate to the delta in battlefield results.

Try: "When Russians bomb Ukrainian civilian infra, that's bad. When Ukrainians bomb Russian civilian infra, that's… actually also bad". There is a fine line between justified reciprocity and a race to the bottom. In war, it's hard. But bribing the electorate is not retaliation against the outgroup, it's just contributing to the burning of the commons. Trump is simply corrupt swine.

And frankly, you are not at war, you're all on the same side, as infinitely opulent burgers bored out of their minds, locked in their conflict mostly for entertainment, like aspects of Chaos.

There is simply not enough money in global markets capable of absorbing a meaningful productivity increase in American goods

There is (see Americans bullying people into investments), but more importantly this isn't a story of money. AI-driven productivity could lead to cost reductions, and marginal displacement of current imports, eg from the same China (here we assume that the US will be meaningfully better at extracting value out of AI). So it's possible to have substantial growth in production without equally impressive growth in revenues. The story of Chinese solar panels is illustrative. Chyna is replacing a large fraction of global energy infrastructure, what does this buy them? 0.3%?

Ultimately, at the scale of American economy it is just not enough to make things to make a difference.

What makes you think they didn't?

overcapacity in the sense that "the Chinese produce much more compared to what they can currently consume domestically" is just a matter of fact

This is a fairly widespread condition. German miracle was all about "overcapacity", so was the Japanese one. Koreans and Taiwanese have "overcapacity". Where does this idea that a nation needs to consume everything it produces come from? It's plainly self-serving. If anything, the opposite regime of deficit spending is more abnormal. Extra capacity will eventually get absorbed as the world gets richer. Even the argument about subsidies is confused; if the Chinese or someone else is subsidizing their goods, it's a trade-off that costs them productivity and lets other players capitalize on cheaper inputs. Basically the whole overcapacity discourse is cope for Western fiscal irresponsibility and stagnant incumbents in threatened industries.

then that production would show up as business investment and either in domestic consumption statistics (more likely) or in net exports (less likely, numerically there's just nobody else that could consume that volume of production), and either way it'd be reflected in GDP statistics.

I do think it'd be reflected, but it's not clear what the norm of reaction is. Automating office work does not linearly lead to X times more office work done at the same compensation, maybe the workload can just expand to fill the capacity, real earnings are fixed, and net "productivity gain" is minor. Hardware is a harder question. I am especially skeptical of hedonic adjustments.