In addition about how technically unfeasible this is for various reasons mentioned above: by now we have learned how evolution works. They limited advanced chips and DeepSeek come with advanced methods. They tried to limit Huawei and it is stronger than even. You limit the chips and Chinese will learn how to make good enough replacements. And in 10 years their replacements will be competitive, while if you just let them buy the chips they would remain dependent on your tech.

You can go to "Customize GPT" and explain under traits and it should be polite and patient, but not sycopanthic, to never congratulate me on my question and to not ask follow-up questions. Then it becomes saner.

This could genuinely be useful. I often find myself paying $5 more for the same product on amazon vs retyping my address and CC number of umpteenth time. If an AI agent could do it for me, I would pick the cheaper option. (only as long as MC is willing to pick up the tab, when AI screws up)

If would be very cunning for EU to offer to host it for N years, fully fund its base operations and *shield it from any political interference*, i.e. own it but give it absolute and unrevokable editoral freedom.

I have never ever seen anybody exiting Performance Improvement Plan (and I have seen a few) and survive with a job. The sole purpose of PIP is to shield the company / institution from being sued for wrongful dismissal.
The moment you are on PIP you should be looking for a new job. No point even in trying, just cash in the last few salaries.

I haven't completely deactivated mine as I have a club page that I occasionally need to admin on - but I've stopped using it for myself for at least the past 18 months or so. The thing was that I wasn't missing "updates" from friends or family or anything - because almost no one actually posts updates about themselves or what's going on in their lives anymore (at least not in my circles).

Almost everything I was seeing was just sharing of generic memes and stories from the alphabet soup of "funny" Facebook pages. I don't need FB to see memes. Or to see anything funny. There was just no value in anything in my feed anymore. Just that and the ever-increasing number of ads, "do you know these people" blocks, irrelevant stories, suggestions, etc. So why bother? All of that on top of the things you mention like mental health, tech bros, and so on.

I work for a big national lab and I can make a few observations:

a) Postdocs that have already accepted a job are having second thoughts. Statistically speaking, the likelihood of being detained by the ICE with a valid visa is probably about the same as the airplane suffering a catastrophic issue, but for human perception this does not matter. US is just not attractive any more. Nobody wants to enter a country where you are looked up with suspicion. The fact that the salary just dropped 10% in EUR is also not promising.

b) You would be shocked just how much US research institutions rely on foreigners. It seems immigrants support both ends of the economic chain: slaughterhouses and agri jobs on one end and high tech jobs at the other with natives filling in the middle. There are grants for which you need to be a citizen and they are so much easier to get simply because there are so few of them.

c) The mood is totally depressed. We had cuts this year despite CR with draconiam further cuts next year. And I'm talking about physics, not sociology of woke people. Everybody is sniffing for an exit.

It is not more accurate than GPS. It is more accurate than " traditional GPS backup systems (like Inertial Navigation Systems or INS)" as per f'in summary on slashdot. INS basically means take accelerometers and integrate in time twice.
So, no, I'm not going out of bed for this.

I was decent coder before entering university. And I could probably pick up some skills and some maths along the way while doing stuff.

But to really understand how Fourier transforms work you need to go to college. And spend real time thinking about it vs picking something as you go along.

So, no, I would not recommend this to anyone who wants to be something more than a coding monkey.

In a round-about way, you're not completely wrong. Many of those "best company for _____ of YEAR" awards are based on self-nominations from the companies. You'd have to read the details of some particular award, but for a lot of them, your company is doing the legwork of applying based on whatever criteria the award has defined. Doesn't mean they haven't earned it or are evil because of it or anything... but it also doesn't always mean they were hand-picked out of the blue by some external entity because of how awesome they are.

At 20 tokes / second, you do 630M tokens / year, which at 630*15 has the value of $9450 which just about the value of the desktop computer you need to run this.

And while it is true that o1 is better than deepseek, it is also true that $15 is a heavily subsidised price. I'm sure it costs OpenAI more than $10k to run o1 for a full year, not to mention electricity cost.
The point being that AI can be commodified in the sense of enabling small outfits buying a bunch of servers and starting to compete with big guys by selling those access to those free models.

So, I think you're right that as a person, you are better off using o1 at $15 per 1M tokens. But if you happen to be the OpenAI CEO you should be worried about your business model.

I thought the same, but I think it is a bit more complicated than this. If every parking spot would have a charger attached to it (which will undoubtedly happen in 40 years, but not there yes), then your argument hold and even better, overnight charging is fast enough. But if I go on a vacation and the closest charger is in some random mall that has nothing of interest to (something that has recently happened to me), then fast charging is indeed a very useful feature.

You may have already read it in a couple of the replies above, but this wasn't his work computer. It was a personal computer at home. So not really much "unauthorized" about that or "permission" required. They got in through his then compromised 1Password account where he had apparently stored Disney credentials. Of course, should he have been storing work account info in a (assumed) non-work password manager? Don't know what Disney's policy is there.

While your overall thinking is mostly correct... TFA says it was his home computer, not a work one.

But they were able to compromise his 1Password account on that computer, which had Disney credentials stored in it. And I don't know about you, but I'd suggest that many corporate password storage policies are not as clearcut as their software/download policies. Was it his personal 1Password account that had work credentials in it? Does Disney have recommended password storage guidelines or requirements? Was he following any of those?

I work as a Physicist in a National Lab and we were recently given fedramp compliant access to chatgpt including the thinking o1 model and I've been trying to use it. And, it is indeed getting useful.

First I gave it the full problem and it wasn't able to get very far, although it got the basic intuition about the problem right. But then I chopped the problem into small pieces -- I had a vague idea on how to do the calculation and was stuck at the first step. Admittedly, I wasn't really working very hard, but my PhD student spent one month looking at it and got sidetracked many times. O1 did manage to get a crucial insight. Surely, it is a standard math technique that I've seen applied to in many contexts and chatgpt must have seen it in its trainings but the fact is that it got me on the right track in 5 mins. I could probably find it in the right text book or eventually work it out myself, but it would definitely take me much longer. So the next step is to lead it through the rest of the problem step by step. Will see how it goes, but whoever claims these things are not intelligent in at least some sense of the word is a moron.