krona

That battle was fought in 2012-2015 and except for a few exceptions, IBM lost. The only real advantage IBM had was in mainframes with 1-10TB RAM. For example, customers who like doing database joins on a time-slice of countrywide telecoms data. Only a few customers have that use case, unsurprisingly.

abstract inductive reasoning requires dynamic learning in unfamiliar contexts. All the benchmarks which measure this specifically (in particular, task composition) see LLMs fail catastrophically. There is quite a lot of research published on these limitations now.

LLMs don't have the executive control to dynamically manage and compose mental sub-routines, when it has never seen those sub-routines in its training corpus. That requires fluid intelligence. ARAOC benchmark measures this specifically.

LLMs have almost no fluid intelligence or capacity for abstract inductive reasoning, relying on crystallized intelligence for basically everything.

Interpolation vs extrapolation? It's very simple distinction and fundamental. Look at ARC-AGI-3 for empirical evidence, if that's important to you. My point however is purely an a priori one.

That's a common meme but it's the opposite of true.

But you're restating what I just wrote - We're training a status quo machine and the probability of anything outside that distribution rapidly drops to zero.

so the proof to the unit-distance problem was on the manifold, given it was outputed by a LLM?

By design this must be the case, even when you account for stochastic sampling i.e. 'temperature'. All it's outputs are a highly-dimensional combinatorial interpolation (I'm talking about GPTs here)

That's probably why Claude is very good at producing plausible nonsense rather than the often correct response of 'I don't know'.

They construct space between the points and extrapolate outside of the points.

They don't. They interpolate between the points on a manifold.

The output of a GPT is an interpolation (an estimation of new data points inside the range of known data) rather than extrapolation (estimations outside that range).

99% of the time we don't need a true intellectual breakthrough to get the job done, and often 'new ideas' are simply riffs on or blends of old ones, like fashion or music genres.

The worry to me, however, is that if society comes to rely on this form of 'AI' then eventually the model collapse bleeds into academia (e.g. grant proposals reviewed by AI?) causing a kind of incremental sociocognitive atrophy. Everything becomes a reaffirmation of the status quo.

That being said I think people said something similar about electronic calculators (that if you couldn't do long division by hand then you'd be too incompetent for higher-level calculus.)

How do you know they're not also swallowing a portion of the supplier costs?

Nazism is as eschatological as Marxism (class vs racial struggle) and many other secular/religious ideologies with utopian endpoints.

Depending on your point of view there's no difference (because the end justifies any means), but I think the True Believers of these ideologies would beg to differ.

literal, unashamed Nazi

He's a Christian Zionist which is the belief in the fulfilment of old testament biblical prophecy. This seems diametrically opposed to Nazism to me, ideologically. I don't know how you'd square the two.

The typing was never the bottleneck.

I'm actually talking about both. WSJ publishes Anthropic artificial profitability. Days later the reason for the profitability appears in SpaceX S-1; it's compute costs were artificially suppressed. Both are going public. It's a quid pro quo.

They need to financially engineer a good looking quarter beforehand.

Perhaps Larry Ellison can cut them a nice quid pro quo for a few months to make OpenAI look profitable (like the SpaceX/Anthropic deal), although that's probably unlikely given the debt Oracle is taking on to build it's infra.

One reason is that blaming Israel for the Iran FUBAR situation is very convenient, especially for keeping MAGA on board.

Polling shows support for Israel is far greater among Trump loyalist voters than non-loyalist Republicans, so this is surely false.

Perhaps you're confusing "MAGA" with actual American nationalists, who are statistically irrelevant.

The actual cost is going to drop 99%

Do you mean the marginal cost by the producer, or the cost on the consumer? I can't see the price of electricity falling much, and the demand curve is apparently exponential if the hype is to be believed.

I like the alliteration but the article is trying to pitch a dichotomy where one need not exist. China is a good example of classical fascism with 21st century characteristics and according to the article it's working great for them.

Apparently both dev's and the AI are vulnerable to the Dunning Kruger effect.

About 5,000 counter-protesters from trade unions and anti-fascist groups marched in a separate route in central London.

There's your answer.

The Metropolitan Police are (justifiably) expecting this protest to turn into a violent riot

Robinson has organised 4 London rallies in recent years and this is the second Unite The Kingdom rally. So what makes you think this will be the one which turns violent?

It's basically families listening to speakers on a stage.

I'm not sure what your argument is since the police enforce the law as it is, not as it should be. "Without fear or favour."

The ISIS-linked kid that bombed Manchester Arena was known to every intelligence agency and was even physically stopped by venue security before being released due to concerns about racism in enforcement.

The bureaucratic solution to situations like the Arena bombing is to remove human judgment and replace it with 4k video analytics. The technology already exists. I don't like it either but if there is ever a way to remove decision making power from a person by means of technology or process, the bureaucracy will gleefully use it.

Perhaps it will be the first protest where FR is used, but the first pilot (which ended in March) just put 2 FR cameras on a street in Croydon and they arrested "170 wanted criminals" in 6 months.

https://news.met.police.uk/news/met-makes-one-arrest-every-3...

I can't tell if you're trolling but `unsafe { crash() }` is safe from the compiler's perspective. Otherwise you wouldn't be able to achieve anything in 'safe' rust, even print to stdout.

Go's Webassembly story is a joke at the moment, so no.

Two main reasons: Foreign direct investment, averaging ~5% GDP/year, largely to build and fully integrate Polish industrial base in to Germany. Secondly: an education system designed to create an economy on advanced manufacturing.

The same has been happening in Slovakia; GDP growth per annum very comparable to Poland since 1995.

As a typical example my very German car has many components with "made in <Poland/Slovakia/Hungary>" on the side.

FWIW The lowest hazard ratio they came up with in both their models was 5+ eggs/wk + 1 portion (14g) nuts/seeds per day.

The study wasn't about veganism so it isn't as useful as other finer-grained dietary patterns or intake variables which they account for.

...adjusting for other dietary factors, demographic variables, lifestyle behaviors, and comorbidities.

They give the data for the specific factors. What is the missing variable which explains their result?

Daily hours of sunlight? Average number of sudoku completions per day?

True, but it conveniently hides the fact that voltage control becomes more of a problem in a low inductance environment.

I already said REE (the operator) was partly to blame, but they were reacting to something their grid was ill equipped to deal with.

Perhaps you should understand the difference between distal and proximal causes of events? Both are important. Voltage oscillation was the proximal cause. Where did this come from? It's explained in the operators own report:

During the incident analysis, it was determined that the oscillation was not natural to the system but rather forced. This oscillation is observed with significant amplitude at a Photovoltaic Plant located in province of Badajoz (PV Plant A). At the time of the oscillations, the plant was generating approximately 250 MW. Since the oscillation was forced, it ceased once the plant stabilizes it.