LittleTimothy

It's also a head-ache for options traders because some options models (black scholes) have log-normal pricing baked in which don't actually allow for the underlying asset to go negative. So nevermind worrying about taking delivery, your HFT options desk just had their algo blow up.

I find the Libertarian obsession with Argentina quite curious. There seems to be this view that Argentina is massively over-regulated with massive government and that if you strip that all away Argentina will be a power house of productivity and wealth. The problem I have with that is... that's not libertarianism. Thinking that Argentina should be regulated and taxed a bit more in line with a modern liberal European democracy is not a demonstration of the successes of Libertarians. I would think that the libertarian would be far better pointing at a low regulation, low tax economy and demonstrating how reduced regulations would help there. Maybe for example, take the US banking regulations and strip them away - if the libertarians are right, we'll see a boom lifting all boats. If they're wrong we'll see a string of notorious and massive scams ripping of the average American. Unfortunately that'll never happen so I guess we'll never know.

I think this is a simplistic take. In companies where there are clear management structures there are clear and obvious ways for managers to fuck around and play politics. When there aren't clear management chains, people with probably similar characteristics fuck around in different ways - it's just less obvious to some people.

Management is a tool used by people with their own motivations to acheive their goals. But a lack of management lets those same people acheive those same goals in different ways. Whether that's starting up duplicate projects and products, causing chaos and confusion by inserting themselves into topics that don't concern them, or simply picking fights. The same people get along in any organisation, the tool of management is just the easiest to spot from below.

That's not really possible. If you're writing a dissertation and find that some of the important foundational work you're building upon is wrong you can't just ignore it. If it's wrong, you need to say it's wrong in order to justify why you've chosen not to consider it in your dissertation. If you can't claim it's wrong, then you have to use it in your dissertation otherwise you're going to leave yourself open to criticsm that your dissertation ignores important prior work in the field. So the only choice you have left is to write your dissertation built upon work that you know is trash, and then if you do later choose to publish criticism of that work you're essentially trashing your own dissertation. And of course, none of that matters, because you're still going to face professional consequences when you choose to write the separate paper disputing the famous paper.

Non mainstream are they guys who are accidentally being directly paid by Russia...

Sorry but I don't think you understand the newspaper business. Bezos bought the entire Washington Post for $250m. Amazon has a market cap of $2T of which Bezos owns ~9%. The capitalist incentives are very clear, the market dictates that Bezos should do practically anything to WashPo to help Amazon. Let's say that WashPo drop to 0 value, that would be a real shame! And I'm sure Jeff would sigh really quite loudly while sailing his megayatch over to Blue Origin where they're working on that $3.4B contract for Nasa that Trump could cancel any minute. That megayatch? It cost him 2 Washington Posts to build.

I don't know what you think market forces are going to do here?

I have some sympathy for the view that we've just handed back some of the gains, but I don't understand the attitude about sanctions. Russia is a small economy. A small economy who primarily exports gas & oil to eastern europe. If there really is a peace deal and sanctions are lifted, I don't see how that positively impacts the US? It means cheaper oil for eastern europe (pricing out some of the US nat gas exports), almost certainly higher deficit defence spending across Europe (on local heroes, no one is trusting the US for security). Other than "good news markets go up" is the logic behind thinking a peace deal would drive US markets? A laughably theoretical deal for minerals?

I wonder how much stock people put into people like Andrej's opinion on an Elon Musk project? I would imagine the overwhelming thing hanging over this is "If I say something that annoys that man, he is going to call me a pedophile, direct millions of anonymous people to attack me and more than likely will attempt to fuck with my job via my bosses".

Let's say the model is mediocre. Do you think Karpathy could come out on X and say "this model sucks"? Or do you think that even if it sucks people are going to come out and say positive things because they don't want the blow back?

And one of the strongest militaries in Eastern Europe with a whopping 4.7% of GDP providing an independent bulwark against Russian aggression - fulfilling the exact role that Trump claims to want in Nato.

If you're getting so many applications that you have to apply such a harsh screen that you're likely losing most of your good candidates via false negatives then you shouldn't be soliciting more applicants to apply. This is what this thread is about - if you're saying these guys are getting so many applications they have to start just brutally cutting CVs almost arbitrarily then they definitely shouldn't be posting on HN about their vacancies. Not least because they're poisoning the well.

This is a real issue - I once got approached by a recruiter for a company, it was a good fit, I think I would've walked the interview and been a great hire - I'd heard of them before. The founder had acted like a dick head to one of my friends, I just immediately turned it down. There is a cost to very publicly treating people poorly. People don't seem to understand that these things that big companies might get away with due to scale, smaller companies cannot. People talk.

It makes total sense for a startup to be highly selective. But being overly selective at the CV/application stage is dumb. If they really do have some really highly specialized requirement that should be on the advert. If they don't then being having a high rejection rate at the CV screen stage is going to be easy - it's easy to reject people, but you're overwhelmingly likely to screen out the few candidates that are actually a good fit. So sure, expect a low success rate but a low reply rate is an indicator the company isn't serious about hiring.

The excitement isn't the capabilities of the model, it's how efficiently it was created. One of the major lessons in AI in the last couple of years was that scale mattered - you would want to throw more and more compute at a problem and that has turned into incredible share prices for Nvidia and incredible investments in data centre and energy generation. If it turns out that actually we didn't need quite such incredible scale to get these results and actually we were just missing some really quite basic efficiency optimizations then the entire investment cycle into Nvidia, data centres and energy generation is going to whipsaw in an incredible way.

I think about this quite often. What I'd really like to study at some point is: How much more does the receptionist at JP Morgan's head quarters make than the receptionist at Walmart's headquarters?

Because fundamentally I think there is an effect where the people in proximity to lots of money earn more. Obviously the Walmart receptionist and the JP Morgan receptionist are doing basically the same job. But the JP Morgan receptionist is surrounded by people who wouldn't think twice about doubling the receptionists pay and I would imagine that has a significant effect.

Does that matter? Waymo solved FSD. The technical detail of the implementation is moot.