sm0ss117

I'm legit freaked the fuck out by this, it feels like a flashing red warning signal that the alignment problem is wholly unsolved and OAI isn't taking it seriously.

The disconnection between pelican quality and overall model quality is interesting. I initially assumed that since pre-training is when a model gets its general skill that it happened around when RL started to really differentiate models. That is higher quality pre-trains result in higher quality pelicans, but RL is unlikely to touch pelican quality. However the fact that GLM 5.2 beats GPT 5.6 and Claude Fable puts a damper on that idea.

My only guess is that GLM 5.2 was specifically RLed for SVG generation and that resulted in superior performance.

1. The four color theorem is a useful case study, for which the original proof was validated and 400 pages long. My prediction is that the first couple waves of proofs will be hard enough that a layman couldn't produce them, but simple enough that experts can verify them. Over time the most advanced proofs will get more and more complicated until humans can no longer verify them, this process could happen over the course of a few month or could take literally hundreds of years.

2. Especially early on the overwhelming majority of the proofs are likely to be uninteresting and more novel just because actually producing them would take expert time that's better spent elsewhere. That being said, as above over time I expect the interestingness of proofs to go up until they eventually regularly produce interesting proofs. The vast majority of proofs are likely to maintain their position as of no interest to humans for the simple reason that the vast majority of proofs are of no interest to humans.

In neither case will I make any particular guesses about a timeline beyond it seems like the way things will go.

Mathematics seems like the ideal candidate for AIs to achieve absurd results. It's a purely abstract grammar with true auto-verifiability. Even SWE has the requirement of interacting with real physical things. In math there's no external feedback required, you're solely bounded by the rate and quality of token generation.

The evidence says that during disasters ppl come together and help each other during disaster while state and business institution act in an at best obstructionist and at worst predatory way. So why do ppl in power believe the exact opposite? The simple explanations are that it's a delusion they actually hold, or a lie they project and either way it serves as a justification for the systems they benefit from.

https://www.epa.gov/radiation/tenorm-oil-and-gas-production-... Basically there are radioactive rocks in the deposits that natural gas is stored in. Additionally there is some radioactive gas in the Gas deposits as a result of radioactive decay.

Look, in order to prevent scary radiation from entering the environment through nuclear we have to use clean and even more radioactive natural gas.

Time to do that most desperate of things, actually read past the abstract.

First, this study was only in academia, and not industry so the results aren't necessarily generalizable. They probably do have some correlation, but they're very different environments.

Second, the very first paragraph after the abstract goes into why actually getting hired is just a tiny part of why women don't go in to stem, "...including inadequate mentoring and networking (1); a chilly social climate (2); downgrading of work products such as manuscripts (3), grant proposals (4), and lectures (5); and gender bias in interviewing and hiring (6–9)."

Third, this was rating an applicant for a third party, not actually hiring an applicant for them self, ppl act different when they have no skin in the game. While experiment 5 did try to account for this by removing the direct competition aspect so they only rated a single candidate rather than selecting a better choice from 2 options, however again there's no consequence for the choice except having your response presented in a paper.

What this study does show is that academics think that other academics view hiring equally women over men as socially desirable. This is not the same as actually doing it when given the opportunity, for an example of this see https://www.jstor.org/stable/3711747?seq=1 and the general trend of Americans saying they go to church even when they don't.

There's a lot going on but the basic reasons are: Evangelical white Christians think the existence of Israel is necessary to bring about the Apocalypse, a thing they want because they go to heaven.

The Israeli right wing wants criticism of Israel to be inseparable from criticism of Jewish people. This means you can't criticize Israel without being painted as anti-Semitic, and to be fair there are a lot ppl who criticize Israel for deeply anti-Semitic reasons.

Quick rule of thumb, the reason we associate it with cancer is because the only ppl we systemically tested for it had cancer. If you tested everybody you'd probably find it in everybody since it was in a mandatory vaccine

Parts of India might get uninhabitable in the summer in some of the worse case scenarios. Given the population there and the fact they're a nuclear armed state that's a geopolitical powder keg.

If my choice is between definitely dying of boring old hypoxia and maybe dying in a cool explosion I know which one I'm picking.

But to be serious it's risk management and if the patient is definitely going to die of low oxygen then the patient having a risk of dying another way is still a lower risk.

This is a nightmare. Getting spammed by ads on my devices is bad enough. Now when I go out in public I get shamed for all to see. Targeted automated advertising has got to be banned outright.

Yea, if something is a Natural monopoly, which search almost certainly is, then it should be a publicly owned utility.

This is morally and politically a non-starter. Electricity improves people's lives way too much to be time restricted.

Just another step on the way to the US becoming a pariah state. I'm really not looking forward to the next few years here. While electing a Democratic government might be able to reverse some of the policies, check out what the senate is doing to the judicial branch to see why that might not be enough, it won't return the trust or reputation of the US.

The nativist, anti-intellectual, and outright fascist elements of the US have gone from a vocal minority to a genuine force in American politics thanks to the Senate. While they still are a minority of the population, probably 20% if you cast the absolute widest net possible, due to gerrymandering, 2 Senators per state, lifetime judicial appointments, the electoral college, and extreme voter suppression they are in charge of huge swaths of the government. It's gonna be a wild ride.

You don't work at a startup to make more money, you work at a startup for non-monetary benefits and if you're extremely lucky you might get rich quick. Average case is you make less, good case is you make even if your company is highly successful.

This study gets absolutely demolished by the fact that all the major old world democracies were trade/citizen empires in the tropics and sub-tropics; Athens, Rome, Carthage, the Malaccan states, Phoenician states, Swahili city states, etc. While Northern Europe and China were the kinds of totalitarian states that the DPRK would envy.

The yugoslav wars had a total death rate of < 1%, and saw most of the cities still standing. 3-4 Million ppl were displaced, ~15-20% of pre war population and there are 5-6 countries where there used to be 1. It doesn't take much to change a lot.

They have a good point about missing data, but they're glossing over the fact the fatality rate goes up drastically if the medical system fails.

If the medical system starts to fail it's not just COVID-19 you get to worry about, if the incapacity rate is high enough you get to start worrying about things like: how long will it take for my electricity to get fixed, how long will my heat be out for, where's the food for the local supermarket, when will my pharmacy get my prescriptions in stock, etc. Second order effects start getting noticeable quick in a society as connected and just in timed as ours.

Try public transit in literally any country besides the US, the US is currently terrible at public utilities for ideological reasons. Basically the line of thinking is the government doesn't work, so let us in power to prove it to you

This gets to one of the fundamental limits of for profit operation of businesses, different scales of firm have different incentives. For American Airlines it made logical sense to do stock buybacks, but if the whole of society was in control of American Airlines they would've prioritized service and just paid to keep the operation on life support during the downturn.

It also should allow the government to direct the business for the public good.

Gonna be that guy and say that this map almost certainly has inaccuracies at least in southeast and south asia. If individual owner/operator businesses count I'm sure you can find some hole in the wall restaurant in Thailand dating before 1878. Also the fact that still operating is a requirement obviously imposes a massive colonialism penalty.