catigula

It's just a collation of sleep-end RHR/HRV. It's completely trivial and many manufacturers do this correctly; both are extremely reliable predictors of recovery metrics.

Garmin has absolutely shitty sleep tracking but it doesn't matter; any wearable more or less nowadays can get static heart rate correct, leading to more or less accurate recovery metrics.

"Collect every paper you can find on RHR/HRV/etc. recovery metrics. Develop an algorithm for predicting recovery score per day based on the research" as a prompt to a Codex or Claude agent can probably get you 98-100% of the way there.

Edit: As a proof of concept, I had ChatGPT Pro 5.6 write the algorithm with my Fitbit data. Looks great.

This notion that Dario cares only about the bottom line is simply misinformed, in the most charitable interpretation.

It doesn't track at all with any of his prior stated beliefs or past actions. It's an absurd claim. It's a baseless conspiracy theory, smuggling in traditional conspiracy mechanics for plausibility.

FYI combat vets are a counter example. They frequently kill a lot of people effectively. See: the DC sniper.

“We clearly can’t “trust” these companies not incentivized to randomly kill lots of people, so we might as well give everyone the ability to randomly kill lots of people”.

It’s interesting that most of the examples people making this argument give are high consequence, high difficulty (could you actually murder your coworker with a pen? I’m skeptical.) low kill count methodologies of violence.

We’re specifically talking about low difficulty, highly effective, high kill count methodologies of violence.

And yes, being able to kill a lot of people trivially, even with huge consequences is something people do. Most countries don’t let you run around with an armory as a result of that.

The ideas are dumb. They wouldn't kill anyone, the barrier to entry is too high and you'd trivially be caught and punished.

No, you don't. What stops people is actually access to knowledge. There are killers of varying levels of efficacy; making killing easier means more people die. It's 1-1.

You could literally develop a hyper-intelligent advisor on how to kill people or perform dangerous hacks using ablated 'open source' AI. You can do this right now, this very moment, and have an extremely adept advisor on how to do really, really bad things.

>According to the IAU working definition (from August 2018) an exoplanet can orbit a brown dwarf. It requires a mass below 13 MJ and a mass ratio of M/Mcentral<2/(25+√{621}), or roughly 1/25. This means that an object with a mass up to 3.2 MJ around a brown dwarf with a mass of 80 MJ is considered a planet. It also means that an object with a mass up to 0.52 MJ around a brown dwarf with a mass of 13 MJ is considered a planet

I do the same kind of thing. "You're smarter now, time to try to cut down dumbo ChatGPT".

High IQ bros.

I’m certain they did.

The problem is… what are you going to do about it?

This is obviously an idiotic and dangerous Cold War and has no happy ending.

Via sophisticated cyberattacks, which we know are now possible on an unprecedented scale without nation state resources or capabilities.

Have you… have you been following the news at all?

This isn’t science fiction. It’s happening right now. AI models can execute massive cyberattacks autonomously.

You could, in theory, use an unbounded GPT-6 level model to basically destroy the world economy for many years.

The problem is that it’s impossible to out think a robot you designed to be an expert at cybersecurity on the topic of cybersecurity. The alternative is not developing this and that’s not going to happen.

"We made 3.6/4 Pro, but it sucks, so this is the distilled model" vibes.

That's nice.

It's called Alpha Zero specifically because it was trained from scratch - zero, not on human data.

No, they are not. Specifically, this was a method based specifically on learning from scratch, like most modern AI models.

Why do you think it's called Alpha ZERO?

Chinese models, hurray!

I think you misunderstood. “Narrative” typically frames something as polemic or open to interpretation, which this is about equally so to the whole “spherical Earth” story.

Narrative?

Kind of like the sphere Earth narrative.

Damaging American AI companies just means that China has the lead with closed models.

China isn’t an altruistic state. They’re an aggressor in many fields, economic and otherwise, and this one of them.

Companies had Claude Mythos access in April. Low chance this is on that level.

The extinction threat from AI to all organic species is a little more comprehensive than a little bit of extra concrete, per the people making the AI themselves (I disavow this etc. etc).

The state intervened to stop the Chinese from sabotaging their furnaces. Seems like an open and shut case unless you have 5 eyes intelligence sources.

Unfortunately, his plans aren't very good.

Being good at developing AI and being good at AI safety are diametrically different skillsets with obvious conflicts of interest.

Classic

>racism

>racism, Japan

Type scenario.

Tariffs and import controls. Why do you think that BYD is banned from the US?

Which "frontier models" are you familiar with in a programming context?

I'm a staff engineer and the amount of work and even troubleshooting the latest crop of LLMs can do at my behest is jaw-dropping.

I perused your comment history, and you refer to models as 'ChatGPT', which doesn't really inspire much confidence that you know what you're talking about.

People keep telling you that you're behind and using old technology and that the most recent models are really good and a sea-change because, well - they are.

Any need to address the huge numbers of foreign workers in America putting direct, nonnegotiable pressure on those jobs?

It's time for the program to end.

Sorry, but you aren't familiar with frontier models if you think it's not "good enough". The latest models are quite a bit more capable than most people in many regards already.