So you voted for Kodos instead of Kang [1]. Local models are the way out of the rug pulling

[1] https://m.youtube.com/watch?v=BUAnyVAanac&ra=m

I don't actually disagree, and advocate that for now the thing to do is use both.

You need to use frontier models to understand where the puck is going, but also to use local ones for anything remotely sensitive.

A lot of that is the lack of SRAM scaling. There’s still quite a bit of improvements in the back end (metal traces). That aside, if foundries adopted say a 1T-1C SRAM we’d see a pretty rapid density doubling

“Payments continued in large part due to the circulation of cheques based on mutual trust, effectively substituting them for cash”

I doubt this would be possible in today’s lower trust societies

Not a paleontologist but my 2 sons went through a dinosaur phase not long ago. It seems like the vast majority of known dinosaur samples come from the northern rockies.

Do you support the status quo of millions of H1Bs? Those people I doubt can or will take such a principled stance, making collective action by tech workers impossible.

This is why H1B needs to be a relatively fast path to citizenship. Without that, the bosses will continue to behave badly.

Why do people always assume that people on H1Bs _want_ a citizenship?

There's tons of things you can do to make people on visas more empowered to push back at work without them become citizens; and in many cases cutting ties to their original countries.

When you're drafting policy, you get to be choosy about who you want in your country. I think that's more or less standard practice.

We should want people who want to stay here long term, because then they have a stake, an incentive to stand up to bosses that want to wreck the place.

As sympathetic as I am to the desires of somebody who wants to come for a while and then leave, I think a better policy would deprioritize such candidates.

Wouldn't this be an argument for people who can take principled stances to do so?

that they'll be replaced by H1Bs?

Absolutely, there’s a reason as well why the VC’s funding these defense tech startups are so pro H1B. It guarantees them a compliant workforce who can’t easily quit, ask for higher wages, or complain about things like genocide.

That's exactly why they are imported in such huge numbers.

You can comfortably go back to your home country. It might feel that you wasted your time in the US, but it was hard to foresee Trump ruining the US so fast and there's little to do about it as an immigrant.

Apple has most likely stress tested its engineers to stress test this phone

I can’t read the article but in my group, which is not terribly AI exposed due to its hands-on nature, we are in fact hiring extra people into next year to handle additional AI initiatives that I am confident will succeed. In subsequent years, the plan as I have heard it is to let headcount go down by attrition.

I would not take extra hiring now as evidence that AI is preserving job numbers.

The article mostly cites the huge growth in blue-collar jobs caused by the data center buildout. It’s very handy-wavy about white collar jobs, “yes lots of jobs have been lost, but also the tech sector always has a lot of year-to-year churn, maybe this is just churn”.

Not a super insightful article. The AI jobs apocalypse may or may not be real, but this article doesn’t provide any compelling evidence for its thesis.

It's handy-wavy because nobody knows how it will turn out and which new jobs will be created. In the same way as every predictions about AI bust, job lost, moat (or lack of it) is. If you feel confident about your predictions, don't.

This is indeed HN, so comments are of course skewed to a particular upper echelon of knowledge work, but I'll ask anyways:

When you say (and perhaps you might speculate across other businesses) "hiring extra people... To handle AI initiatives" does that mean you're hiring Very niche and specialized AI people to handle these initiatives? I'm curious because I have seen some versions of this sentiment on HN and it sort of feels like reading "at my company we're hiring more quantum physicists because of these initiatives".

Basically, how much of the general populace is actually going to have a shot at nabbing a slice of this meager pie?

No we are not hiring AI experts, just people we think can be trained to think AI-natively. It’s a different mindset that presumes whatever work you’re doing now is ultimately going to be used to train AI to do it later.

ok well I went from not knowing what a tensor was to having my pi coding agent polish up my orchestration layer in about ~60 days, wheres the ai dreg hiring at?

Ahh I see. Any spots for philosopher / educators?

What type of activities your group performs?

I unironically paraphrase this at work all the time when talking about ops concerns. That and “the absence of evidence is not evidence of absence”

> the absence of evidence is not evidence of absence

takes notes. love this!

Guess it’s that time in my lifecycle, but when I was a kid it was possible, nay probable, to accidentally stumble on some really horrific things on the internet, particularly on p2p sharing, and I’m fine (married, 3 children, no medications). I don’t see why kids need special protection from positive feedback loops. Parents need to be involved to turn this exposure from an affliction into an inoculation

AlexNet was 2012 and they explicitly called out the use of NVidia GPUs