w4der

If you'd never heard of WebOS (which sounds a bit like "eggs" in spanish), the story of how it was developed by Palm, then killed by HP and re-released as a TV OS after being acquired by LG is fascinating. It could do real multitasking on a tablet in 2010, years before either iPads or Android tablets could.

Look, that may be true, but why would the dispatcher then say "Yes, proceed normally"?, as if he'd been asked a question?

Also the train driver could have pushed back, in the radio recording with the dispatcher, he tells the driver to "proceed normally", the driver asks "proceed normally?" and the dispatcher confirms it. Here's were I would expect that better training or more availability of information to the driver would've made him push back on that decision.

Trains in spain have been running at reduced speed, delayed or even cancelled for a while due to aging infrastructure, which might've actually caused another deadly accident back in winter, so this behavior does not surprise me.

I'd like to see an LLM-aided implementation of row hammer being used for something like this.

Many times issues with clinical trials is not a technical one, it's a human one, practitioners unwilling to test, patients that stop showing up, red tape, committees, denial of consent, etc. I do not see how AI as it stands currently can solve those issues.

Wasn't there a comment a while ago when all major LLM providers went down almost at the same time that it was caused by a sort of virtual "Room 641A" thing?

But the tooling required for guns and rifling in particular is pretty hard to come by, whereas the tooling required to assemble a DIY drone is basically indistinguishable from R/C hobby equipment. You can (try) to regulate where/who/how/when can fly a drone, but regulating drone building seems almost impossible to me.

Neither gold nor pickaxes did, but both business went back to reasonable levels and valuations.

That was a scam or at least overblown in scope, the "control room" you see in pictures was basically a movie prop, and their idea was basically to connect that room to every factory/port/mine with teletypes, it's similar in spirit to what a modern stock should do.

The average person may not even own a personal computer, has a company-owned notebook with barely 16 GB of RAM and maybe a TPU, and here you are saying that owning a prosumer GPU and more RAM than what most people will buy across their entire lifetimes is common.

And further down the rung, the electric Twingo, Panda and such go toe for toe with the cheaper Chinese EVs.

I'm gonna plug a recommendation here based on your comment: if anyone wants to marvel or horrify at the US' water development history, Cadillac Desert is an excellent read, https://en.wikipedia.org/wiki/Cadillac_Desert .

I think it is well known by this point that hardware-backed security is good until an unfixable hardware bug is found, this just reads to me as Nvidia saying "please don't regulate open models out of existence, look, I have a solution to appease the regulators, please let me keep selling accelerators"

If this comes through, there's gonna be a grey market for "unlocked" GPUs, were the watchdog is disabled either from firmware, or physically replaced if it's not embedded into the die.

You mean mirror.vim? Or something else?

Where did you get the 10L/100km figure from? I saw closer to 7L/100km when driving it, and the 2.2 diesel on the older X-Trails has about 100 less horsepower.

Why is EV insurance that more expensive than an ICE vehicles'?

I think these kinds of vehicles are fine for countries were fast charging networks are still not widespread. I drove a rental X-Trail e-Power AWD a while ago, and fuel economy, while not spectacular, was still great given the quality of the powertrain. Of course a low displacement manual can get better fuel economy, and an electric will be cheaper, but if you are driving 800km on a country with little charging infrastructure, powertrains like this one or Toyota's eCVTs make sense.

Chrysler built one in the 60s: https://en.wikipedia.org/wiki/Chrysler_Turbine_Car

The one this that is better than just SSH+Tmux+Vim, is that if your latency is higher than 30-50ms, since VSCode's SSH agent streams the files to your computer, the typing experience feels snappier. When you work half a continent away from where the servers are, it makes life nicer.

I am not debating IA's capabilities at coming up with proofs, I'm debating whether such systems will be able to figure out new mathematical fields or not. If they can't figure out some field from what was known at the time, what guarantees are there that AI-mathematics won't just bog down at its current state? There must be some evidence that AI can come up with novel mathematics before we decide to throw academics out the window. The risk of creating a sort of mathematical dark age just based on the current situation is far too great to stop funding mathematics research.

The hell is that reply? If we're obsoleting human mathematicians, which have fundamentally been the same thing brains-wise since forever, why shouldn't an AI system be able to replace a 1870s mathematician in the same way you are proposing it replace a 2020s mathematician?

Google trends shows it spiking last year: https://trends.google.com/explore?q=frontier%20mathematics&d...

You are making your argument on the basis that LLMs and future AI system will be just as good or better as current mathematicians at all the relevant processes involved in academic math and mathematics in general. I'd like to see an experiment where an AI system was trained as a 1870s mathematician and see if it can come up with set theory, if that is not possible, then current AI systems are not fit to replace mathematicians.

Isn't $1000 a bit low for compliance consulting fees?

> I brought this up to someone else and was told that airgapping is apparently much more expensive than I'd naively think.

These labs are one of the most valuable and heavily funded enterprises in the whole world, that they can't properly air-gap their systems to me reads as if they "agents" and LLMs are not as good as they say they are, because if they were, why would it be hard/expensive to air gap a system? They already scraped most if not all of the internet, where did that data go?

> Agents could also hack Anthropic/OpenAI and make it appear that API access has been turned off, when in reality it hasn't.

You know cables, modems, RF equipment and optical transducers can all be unplugged right?

As a kid I don't remember, or have pictures of, an animal-looking piñata, they were just colorful boxes full of candy and things.

For some reason my iphone sporadically won't go back to "music quality" mode after cutting a call on a wireless speaker or car play, and it's incredible how bad music sounds with telephone quality.

And what if you take a picture across a IR-filtered window?