MadMonzer
Temporarily embarassed liberal elite
No bio...
User ID: 896
Although Dr D'Eath (an actual maths lecturer at Cambridge) eschewed all forms of ligature maths, preferring a more direct approach - he lectured on Killing's equations.
TBH I haven't looked into the actual algorithms involved here, but given the length of the proof I'd expect the constant term (call it overhead) to be quite high compared to existing methods -- considering the tiny difference in the exponents, you would need n to be literally approaching infinity for the proposed algo to make any difference at all. In the case of the integer math, this would mean that multiplying integers of ~infinite length could be somewhat faster -- but we can't even test it, because computers don't have infinite memory.
Per wikipedia, the existing O(n log n) algorithm for integer multiplication that the AI improved on is already considered "galactic" - i.e. only optimal for impossibly large data sets. The practical algorithm for the largest numbers we can realistically multiply is a FFT-based approach which is O(n log n log log n)
To my non-expert eye, using the new slightly-faster DFT in the AI papers in the bog standard FFT-based multiplication algorithm should give you a faster multiplier than the new multiplier the AI found explicitly.
I mean that I am one of the people who care about AI safety but don't understand the technical detail.
Duels of honour were not, in fact, a gentlemanly combat sport, although they often pretended to be. The point was to demonstrate the seriousness of whatever the duel was about, plus a certain level of physical courage, on both sides. This was ideally accomplished by imposing a low, but real, risk of death on both participants following which the seconds could send them home with honour satisfied. Hence pistol duels fought in morning fog using antique pistols, or epee duels fought to first blood.
Be it remembered that Justin Trudeau won the Canadian Premiership in the boxing ring.
This is a plan that has a nonzero chance of success - though I think Scott's way overestimating it - and could possibly fit the "better plans are too hard" argument. But you still need a pause for that, just not as long of one.)
Yes - Plan A is a good plan if you think that aligned ASI is a win condition. Part of what is going on in the circle of "people who care about AI safety but don't understand it" (ipse dixit) is that there is a fairly widespread view, on all sides of the political divide, that ASI aligned to Altman and Amodei is not a win condition for humanity - they are both profoundly blue-tribe (which offends Reds) and sociopathic billionaires (which offends Blues). Altman and Amodei are living in a paradigm where "If we build ASI, we will get Clippy or the Culture, so we should build it carefully to maximise the chance of the Culture". I am not alone in thinking "If we build ASI, we will get Clippy or the Culture. Both of these are bad outcomes, so we should not build it". Based on reading the report proposing Plan A, Plan S (ban development of new models for the forseeable future, AI companies who want to keep their GPUs have to accept inspections to confirm they are using them for inference only) is obviously superior, and I am not unsympathetic to a full-on Butlerian Jihad.
The limit on the potential of LLMs to improve the human condition is not the speed at which more powerful models can be developed, it is the speed at which we can incorporate LLMs into our workflows and institutions without breaking anything loadbearing. This is a generation's work even if the frontier models don't get any better than they already are.
That's the one, yes. There is a reason why, as @zeke5123a points out, the states that pull this off are not generally democracies. Dubai is proof that it is possible, not that it is popular.
I wasn't saying Tao was better than AI - I was saying that he is better than the average Fields medalist.
I think Navier-Stokes counts as a human-AI collaboration, possibly the last one we shall see.
And Brian Josephson took up research into psychic phenomena after he won the Nobel for his work on superconductivity. The PR required careful management when he was the only living Nobel laureate in the Cavendish.
If a four digit number of graduate students are thinking about hanging themselves in a week's time, that would be a dramatic improvement in graduate student mental health. (The US graduates about 60,000 PhDs a year, multiply by the number of years the average PhD-bound graduate student spends in a suicidal state).
There is only one Terence Tao, but there are 47 living Fields medalists, and solving Navier-Stokes is going to get someone a Fields medal. If "like Tao" is your standard for a successful mathematician, it is too high.
Agreed. I will correct my comment to "The minority of the organisational bullshit that isn't pure waste is what maintains a causal connection between what the end user wants and what your code does."
Good software management (actually good management anywhere) is partly about protecting your team from the bits of organisational bullshit that don't align them - or protecting the whole organisation by keeping them out altogether if you are the founder. And it is hard because throwing the baby out with the bathwater produces a misaligned organisation.
Dubai does not appear at all, and the AI summaries suggest they have unremarkable border control policies.
Both countries have mass immigration at levels that would horrify MAGA. Singapore is about 45% foreign born (41% non-citizen but the naturalisation numbers are not published in a convenient way) and Dubai is 80+%.
Dubai will let anyone in if they are well-behaved and have employer sponsorship, which I think Caplan would count as "open borders". Singapore keep saying they are going to clamp down on low-wage immigration but somehow the construction industry is still 75% migrant and housemaids are 100% migrant.
Very few countries outside the Americas have full birthright citizenship. Singapore has limited birthright citizenship for children of long-term legal residents, and Dubai has no birthright citizenship at all.
Agreed that there isn't a good example of a democracy that has pulled off mass immigration + welfare state for citizens + functioning economy - Switzerland comes pretty close though. Also it isn't clear that there are any democracies that have a "functional economy" in the strong sense @Ichonochasm is using the world - we are all stuck in a Boomer pension death spiral.
You absolutely can have open borders, a welfare state for citizens, and a functional economy. Dubai and Singapore* are proofs of concept if you exclude oil sheikdoms. The deal of "you live and work in a functional economy, your salary reflects this, but you are a paypig for the bottom 20% of citizens whereas your own social insurance is fucking off if you get sick" is attractive to a lot of potential immigrants if the economy is sufficiently functional.
* Singapore make a big deal about having a stingy welfare state, but the vast majority of the citizen population live in subsidised public housing.
Caplan believes (almost certainly correctly) that a mutually beneficial arrangement is possible between Americans and most third-world immigrants, and is annoyed that no powerful faction in American politics wants to find it. The not-insane pro-immigration faction in American politics (and most other western countries) which he would presumably want to support is weak enough as is, and takes the approach of "Given an effectively-fixed deal that we are allowed to offer to third-world immigrants, how do we identify a sub-population where this deal is plausibly positive for Americans?"
You can do purity spirals in a 2D framework as well. There was a presentation at a Libertarian Party event (I think by David Brin) that spoofed this by showing a mockup of the WSPQ version of the Nolan Chart (unlike Political Compass, Nolan oriented the chart like a diamond with libertarians at the top) with a tiny triangle labelled "Libertarian" in the top corner and the rest of the chart market "dirty fucking commie" - this accurately describes a certain tendency within libertarianism.
It goes without saying that you can do pragmatic big-tent coalition politics in a 1D framework as well. RINOs, DINOs, Blairites etc. earn a good living doing so.
What economic-left policies does Caplan support? Let alone ones that would put him in the "far corner".
Taking money from the enemy crosses a line that stanning them for free does not. Tokyo Rose and Lord Haw Haw are the key precedents.
This is particularly silly given "Trump ironically asked Iran to blow up San Diego and the MAGA crowd cheered rather than laughing" also works as a base-mobilisation message for the Democrats, and may even be true.
So does "Trump is so senile that he accidentally asked Iran to blow up San Diego on live TV and nobody in Red America seems to think there is anything wrong with this."
The organisational "bullshit" is what maintains a causal connection between what the end user wants and what your code does. The big idea of the Torvalds/Raymond side of the open source movement is that when hackers work for themselves or each other they can cut 90% of it out. That gets you Linux. Cutting 100% of it out gets you GNU Hurd, so be careful.
If he has violated sanctions, you have grounds to toss him based on laws passed by Congress even without a declaration of war. I would assume that doing a paid endorsement for a sanctioned entity would be a crime, but I'm not an expert on technical sanctions law.
If someone told be Solanas was a Baathist, I would believe them. I do not think she would have called her work of Baathist apologia the SCUD Manifesto though. From the point of a Western useful idiot, that would be a phallogocentric take on what Baathism was really about.
In hard patriarchal societies like ancient Greece or modern India, there isn't one because women are never considered full adults. In soft patriarchal societies like ancient Rome or medieval Europe, when her first husband dies - the only women who enjoyed what men would consider full adult freedom were widows.
That said, peasant class women in premodern societies enjoyed far more rights de facto than they were supposed to do de jure because their work (i.e. textiles) was just as essential as men's.
- Prev
- Next

Salazar Slytherin and the Sorting Hat seem to disagree about whether sufficiently ambitious halfbloods like Tom Riddle, Severus Snape and Harry Potter belonged in Slytherin. The Hat would absolutely have sorted Vance into Slytherin, and unlike Potter I don't think he would have objected.
We don't see a healthy House Slytherin in the books, but ambition and family pride are virtues and a healthy community of people who possess them is entirely possible. If it worked anything like the Muggle one, most of the work of building the Wizarding British Empire would have been done by Slytherins.
More options
Context Copy link