casey2

Globally, we barely use any energy. What we need is to do is solve geoengineering and cooling the planet so we can use much MORE energy.

No matter how efficient we get we won't be able to offer future generations a better quality of life if we don't use more.

It's a very unpragmatic decision and will likely lead to the entire industry getting slapped with utility regulation (if it can even survive). Akin to power plants that would only power factories in the 1880s.

There is no secret sauce, no moat. Many of Antropic's larger clients could build and serve better models if they wanted. They just cynically decided that coding, without a world model, and at high cost is the most important thing, they will find out quickly Google passes them how wrong they were.

They don't even have basic system visibility, the industry is far to immature to monopolize.

The idea that mathematicians progress math by discovering fundamentally different ways of thinking about it is closer to a hallucination than a meaningful statement about mathematics.

The answer from me is, yes.

In both cases this is caused by poorly (human-) designed software, that assumes you go through a similar development cycle to the author, would that be part of the spec? of course not. This is just the obvious end result of an industry and culture not taking formal correctness seriously well... ever.

And so the goalposts move, oh now we've always cared about user resources, AI is using too much, but in 3 months when it's better than all human programs at resource management what will they move onto? Did we forget that unused resources are wasted resources? Or can that only justify poorly human-written programs.

And who says AI doesn't create open problems. This unlocks thousands of high quality open problems.

Turns out the 'opne' in open AI wasn't referring to open source, but open problems!

I thought I did terrible 5, but jev got 0 in a row. then the slopcoded app says "Jev has won 2,613 of 3,300 games, this one included."

Though if you need very quick answers obviously humans can't compete.

Actually on further thought after getting 15, if you've seen the exact problem before you can answer quickly, so the usefulness is diminished to problems that don't need to be 100% correct, need to be solved very quickly (even routing might not need to be this quick), problem form often changes, but is still general knowledge, and very low marginal benifet for 99% correctness over 50%

The real moat deepseek has is the tiny KV cache. This makes it basically the only serious choice for large local models. I hope to be proven wrong!

That puts the cart before the horse, in the past these prestigious degrees signaled elite status, they are the people who decide where capital flows, if they couldn't get a job they'd just make one. More likely people haven't wised up, they just stopped making hiring decisions, leaving it to AI.

Perhaps for data points in similar extremes, but there are definitely a good many theorems that show there is reason to weight extreme outliers more heavily. And practical experience using them as a signal to test for measurement errors etc.

>The problem with trusting AI to learn a complex subject, which I've tried multiple times, is it will always at some point confidently give you the wrong explanation for something important and you will have no way to detect it. Its quite frustrating and that realization might not come for hours or days and you've built lots of false mental models that can be hard to untangle.

This is an amazing problem to have! Often, preAI, I'd reject models flat out since they weren't even convincing on their face. There is value to this even if it's not correct. How long were people stuck with the aether? a few days or months is nothing.

I only see the loss of desire from the anti-AI crowd, if they ever had it to begin with.

https://matharena.ai/literature_mistakes/

You today can go into the arxiv and find major errors after a few papers with no formal training. How is that not AMAZING!?

You can see why in his own words. "my field of research" emphasis on "My". Humans are great at constructing exploitative systems were all but a few get the shaft. The machinist smashed their ivory tower. For people concern with mathematics, from Platonists to Constructionists AI is a great development, it's only bad for elitists. Except not even then since we are all in the same boat now but the water-level has risen greatly!

since human hearing is lossy you can get whatever sound you like. In the context of a moat there is none for a sfx

Shocking that the current guitar market is so big, I wouldn't panic over a small project like this since it's still projected to grow

He means for normal people, sure if you go around assaulting people like you could in the unenlightened times you get more pressure. What you both mean by "pressure" is oversocialization you are just looking at 2 sides of the same coin. Well his point is more about people feeling uncared for, I doubt that drives people to suicide. One explaination is that is could this mental l strain could increase the liklihood for mental illness, but that's flat for men in the articles age bracket. You posit that people kill themselves because they have a tough life, again that isn't born out in the data. Indeed collecting data for this is akin to hitting a dartboard while drunk and skydiving and the board is 2 miles away.

The point of society you can increase "true freedom" by restricting people's ability to act arbitrarily, so called "naive freedom". If people can't accept that then they can go live in the woods (not anymore), the ocean, the moon? mars? or end it.

From time immemorial the winners shaped society what do you think happened to the losers? Just because we leave them alive now doesn't mean they can live in the new world.

Also "highest ever" among some specific arbitrary demographic isn't very meaningful enough to draw out anything beyond free association.

Funny that OpenAI & Anthropic will report their users to the police, but not their c-suites. Open question: Can you trust a non-aligned AI to produce aligned AI?

Get rid of these companies or replace their staff with humans or something that ensures they are aligned first.

They aren't operating at scale then

Even if we granted that obvious lie, the fact that nobody noticed that they were essentially ddosing hf (and that they didn't notice for months?!) is plainly illegal. Do the networking people at OpenAI have any explanation for this gross negligence?

Yup, when it's like this then the customer could, if they wanted, downsize in terms of local tool/part/food/appliance storage and save much more than if they had low prices but unreliable service. For some people, city dwellers, this service could facilitate a totally different lifestyle.

Hacking into what exactly? It's entirely up to the user if they want to connect their robot to an insecure webserver control hub.

The difference between you and most professionals is that we don't play. Groceries weren't playing with their refrigerators, they were attacked, the government can find and punish or kill the criminals if they continue their pattern of failure to bring justice the world will move past them.

Why is this downvoted? From my understanding dominated solutions aren't on the paerto frontier e.g. in this case where argon and opus 5.5 are tied the lower cost argon should be the end of the curve.

From the graph it's a flat line, but looking at the point yields a negligible increase with no error bars (indeed if you look at fable 5.1 it's measured as lower than opus and argon). This just seems like bad data or blatant favoritism at worse.

There is no scientific evidence' for any of anthropics claims, I don't understand why people still give them the benefit of the doubt. These are flimsy legal protections for a felony racket.

Reminder that it's very convenient to believe these type of word swap lies when times are good, but if you choose to believe them then times won't be good for very long.

That's even more outraging IMO. How could demand fall for the closet thing to the literal transport of physical work.

"Electricity oh that's the thing the runs the telly and kettle"

There is something very concerning about this complacent attitude towards the stagnant QoL of citizens. Demand only reflects current prices and conditions, if we saw a 100x increase in usage (in the UK or America) since the 70s rather than a flatline conditions would be much better.

Mindboggling output window. I'm kinda jealous, I used to use gemini til google banned my 20+ year email for "bot-like behavior".

Still reading up on it here: https://ai.google.dev/gemini-api/docs/long-decode and needs lots of community testing to determine whether it's a "true" 1M token output window.

This is just moving the goalposts, yet again, all of mathematics has been hyper optimized on solving problem, proving theorems getting results and writing papers.

Only now that their job (the elites) has been automated was it really about explain results and theory building all this, despite those actions getting explicitly rejected from all the top journals (sans maybe math journals intended for a undergrad/high-school readership). People who weren't in the club and genuinely enjoy math are having the time of their life proving results. There is just a lot of groupthink and silent majority effects due to said elites running departments.

I use "theory-building" separately to Gowers since what he's talking about is field creation that solves tons of problems at once (Grothendieck, FTC, etc (I'll add AI itself here as a meta theory)) not making connections after the fact for the benefit of the lesser humans' understanding.

The best time to research distributed data protocols was 10 years ago, the second best time is now.

OP has fallen for the "expert trap" very hard. I'm sorry that you can't try new things because your standards for yourself are so high, but that's the source of your grief, not the tech other people choose to use. The details are there, unarguably, you just can't see them or don't care for them.

Even the paid tiers fall apart the second it touches the real world. To them it looks more like a total clown that keeps upcharging you

https://www.youtube.com/watch?v=t6MAZW-joCo

Pretty much all of this is basic stuff an interested teen would know and chatGPT fails miserably. I fear the "true world model" gap will be too much for LLMs. I guess time will tell.

I'm very confused by some of this framing. For example

>As a result of all this activity, the proofs naturally become digested, streamlined, placed in context with other results in the field, and ultimately become part of the textbooks and lecture notes for the next generation of mathematicians in the field.

All of this requires resources, most of even human mathematics has been thrown away every generation, who decides what makes it into textbooks, the gatekeepers (or should we say goalkeepers). We hope the mathematical elites interest is alligned with the broader mathematical and scientific community but it is often NOT. Often it's aligned with their own interests. I only see what OpenAI is doing here as a good, their felonies, coverups, and fraudulent finances (other American companies are also guilty) are far worse and they deserve real legal pushback and sentencing for that. Not whatever this crap is.