skwirl

Making no effort to obtain a grasp on the basic facts of the situation doesn’t seem to stop people here from posting embarrassing rants.

Altman was on the board. He was not “just another employee.” Brockman was also on the board, and was removed. It was a 4 on 2 power play and the 2 removed were ambushed.

You also don’t seem to realize that this is happening in the nonprofit entity and there are no shareholders to fire the board. I thought OpenAI’s weird structure was famous (infamous?) in tech, how did you miss it?

The architect of the coup (Ilya) is strongly opposed to open-sourcing OpenAI's models due to safety concerns. This will not - and would not - be any different without Sam. The decision to close the models was made over 2 years before the release of ChatGPT and long before anyone really suspected this would be an insanely valuable company, so I do believe that safety actually was the initial reason for this change.

I'm not sure what you mean by your second paragraph.

We have diversity in the space, and OpenAI just happens to be the leader and they are putting tremendous pressure on everyone else to deliver. If Sam leaves and starts an OpenAI competitor I think it would take quite some time for such a company to deliver a model with GPT-4 parity given the immense amount of data that would need to be re-collected and the immense amount of training time. Meanwhile OpenAI would be intentionally decelerating as that seems to be Ilya's goal.

For those of us trying to build stuff that only GPT-4 (or better) can enable, and hoping to build stuff that can leverage even more powerful models in the near future, Sam coming back would be ideal. I'm kind of worried that the new OpenAI direction would turn off API access entirely.

>I don't believe professionals who work in biggest AI labs are spooked by GPT.

Then you haven't been paying any attention to them.

> Suburbia may have little traffic, but they are also devoid of "third places" where teens can be teens.

Could you unpack this? I grew up in a small town, moved to a very walker/public transit friendly city for 10 years, and now live in a close suburb. I didn’t really see any extra third spaces geared towards teens in the city. In fact, I rarely saw kids out on the street or congregating anywhere. In the suburb I am in now, they are everywhere, biking to each others houses and hanging out in the town center.

This is a pretty unimpressive list for 9 years of progress, and is merely a small subset of features Kotlin was already providing.

I left the Java world for the C# world around 2008 and then returned to the Java world last year. C#/.NET changed much more significantly much faster. Almost all of the significant differences in day to day life with respect to the Java language and standard libraries that occurred between 2008 and 2022 are due to changes introduced with Java 8.

I’ve seen this sentiment a lot and I don’t understand it. I actually feel like very little has been added between Java 8 and Java 21. Virtual threads is probably the biggest thing, and it looks very promising, but the language itself is still a chore to use, and Kotlin does a good job at cleaning a lot of that up.

I think Kotlin is at risk, though, of having the JVM force them to support two competing models to accomplish the same thing - coroutines and virtual threads is a good example (although Java still doesn’t have language level structured concurrency). This could really make things messy.

Do you have to baseline on Java 21 if you want to add support for virtual threads? Couldn't you continue using heavyweight threads on older versions of Java? My understanding is that both use the same Thread abstraction.

Have you ever actually written software professionally? I get strong teenage vibes from your comment history.

C# is one of the best designed languages that I've had the pleasure to use.

This hasn't been true for many years now, but that legacy persists in the popular consciousness.

How are "the elite" forcing companies that do not own real estate and who rent their office space to mandate RTO? How are they forcing these companies to renew their leases when they expire?

It's very hard to define what is and is not "related to language" and this is kind of a fundamental question that seemed to get a lot of attention in the 20th century. Maybe these language models can help shine some light on that.

According to OpenAI, GPT-4 scores 4 on AP Calculus BC, 5 on AP Statistics, 4 on AP Chemistry, 4 on AP Physics 2. But is mathematical/logical reasoning largely a language task? I don't really know. I feel pretty confident saying that riding a bike is not a language task, but logical reasoning, I'm not so sure.

LLMs are good at a lot of things we don't have a good reason to expect them to be good at. It's very hard to come up with "theoretical reasons" it should be good at things, in "theory" they should not be nearly as capable as they are. Even NLP researchers have been shocked at how well this has worked.

They will have to fix Boston’s horribly broken subway system first.

I’m guessing you aren’t from the US because in the US there is not much prestige in the title of engineer, but it seems to be a title people in Europe get weirdly hung up on. If anything, software engineering positions are more prestigious than most traditional engineering positions here.

Whooosh

You could have just read the article to see what the author meant by AI Engineers. You also would have seen that the entire article was addressing a point you tried to make, and you could have responded to the arguments made in the article instead of, ya know, just sounding off based on having read nothing but the headline

Where was the lie? It’s pretty crappy to baselessly accuse someone of lying.

Carmack was on the Lex Fridman podcast last year and doubled down on these comments and actually went even further. He has not changed his mind at all.

Carmack is very much atypical, but I don’t think getting completely wiped after 7 hours of any focused work is normal, either. It’s certainly the case if you are doing 7 hours of something you hate or find tedious. I would probably be fried doing 7 hours of nothing but code reviews. But if I’m building something I feel excited about, I can work as many hours as my life obligations will allow and feel pretty good about it. Lots of people have no trouble playing 12 straight hours of civilization or other focused video games.

Looks like the owners of “kbin.social” are learning the hard way that hosting your own social media service isn’t easy. Just 25 minutes on HN in the early morning US time brought them down.

Principal Engineers are supposed to raise solutions to problems. This (3 year old) blog post raises complaints, but does not propose a solution. People who can raise complaints are a dime a dozen.

There seems to be some unfortunate social herding around AI beliefs in the 'rationalist' community that seems very un-rationalism-like to me. Some of these arguments are really poor.

The episodes are really hit or miss. I really enjoyed the interviews with John Carmack and with Guido van Rossum. I started several other episodes that I just bailed on, though, and it's a big time investment just to get to the point where you realize you are wasting your time. So at this point I won't go out of my way to listen to episodes but I will go listen to one when it's getting talked about elsewhere.

The less Lex is talking and the more the person he's interviewing is talking, the better.

BTW if anyone who enjoyed the van Rossum and Carmack interviews has any suggestions on other episodes I might like, I'd love to have them.

I think the fact that a bank called "Silicon Valley Bank" collapsed is really breaking a lot of people's brains. The big tech companies (and other big companies) have treasury departments that manage short and long term cash. The companies that are impacted by the current banking instability are largely small (and some medium) businesses that generally only manage cash reserves for short term expenses. The tech companies that were impacted are largely startups that are depositing their runway which they generally quickly burn through. Many of these startups are small enough to not even have a CFO, let alone a treasury.

But it's not a tech startup-specific problem - it's a small business problem. Signature Bank (collapsed) and First Republic Bank (barely hanging on) largely handle bread and butter small business accounts. These are companies that just need a place to store cash that they use for payroll and recurring business expenses. Many of SVB's customers were also non-tech small businesses.

One SVB customer interviewed here in Boston was an electrical contractor. I'm not sure that it's wise to expect electrical contractors to have CFOs and treasury departments to "mitigate counterparty risk" with their banks if you want to have functioning small businesses in your community.

Not sure of the relevance of the last line of your comment, but it's wrong. A bank run can either be the cause or the effect of bank insolvency. My point was that there's no real orderly way of noticing that a previously healthy bank is now unhealthy and transferring out your money. As soon as the bank is believed to be unhealthy you get a run and near-instantaneous collapse, as we have been seeing. Notice that I used the words "healthy" and "sound" and not "solvent." Any bank will become insolvent in a bank run.

I think there's a real misunderstanding here with the distinction between banks, depositors, and what protections apply to each. The scenario (betting on red) is illegal, but assuming it was not, it doesn't matter to the bank whether or not its depositors have their deposits protected.

I can kind of assume what your misunderstanding is, but it's not completely clear. I think you are assuming that if the deposits are protected, the bank gets to keep the deposits and continue running. This isn't how it works, though. As soon as the bank becomes insolvent (loses its bet on red), the bank is shut down and its shareholders are wiped out. The FDIC sets up a new, government run bank to hold and guarantee the deposits, and then tries to find another bank to sell the failed bank's deposits and loans. Right now there is no Silicon Valley Bank. If you had deposits there, they are now held by Silicon Valley Bridge Bank, N.A, which is a new bank operated by the FDIC.

I don't think it's a big change. I think very few people are qualified to do due diligence on their banks. It's a bit like expecting people to inspect bridges before they drive over them. It's a piece of financial infrastructure that we expect to just work. It would be a big change to not have that expectation, and would likely result in the collapse of the regional banking system as people flock to the Big Four "systemically important banks" which have stricter regulatory requirements and the implicit backing of the federal government. We are seeing a lot of that flocking despite the actions of treasury on Sunday.

Moody's gave SVB an A rating until they were already collapsing. The State of California said SVB was financially sound until March 8.

And the problem is, even if it were possible to know the health of a bank as a depositor, a sound bank that people come to believe may be unsound collapses in a bank run the instant that the realization occurs, so you would lose your money anyway even though the bank was sound when you first deposited unless you got in early on the run.

To prevent bank runs.

>Why would any bank look at SVB and NOT think "oh, time to take more risk for more profit; the government will prop up the FDIC limit if we fail anyway".

This makes zero difference to the bank. The bank doesn't get saved by the FDIC limit, as you know. What happens after the bank fails - whether the depositors are made whole or not - is immaterial to the people who owned the bank, who now see their asset (the bank) worth $0.

If you want to make a moral hazard argument with respect to the FDIC, you'd have to make it with respect to the actions of depositors.

Also, WRT the $250k limit, that's the minimum they will guarantee. They have always tried, and in recent decades always succeeded, in making depositors whole one way or another, usually without spending much (if anything) from their insurance fund. The $250k is the worst case scenario.