capitalsigma

It seems unlikely to me that Anthropic employees really care about X-risk (to a degree that the CEO feels it necessary to make public statements that he otherwise would not). I'm sure the pay is great and the problems are interesting -- lots of smart people work on much more evil things than LLMs

Crypto was mostly grift targeting non-technical people

LLMs are a real, genuinely-useful technological innovation that happens to have arrived first in software engineering. Lots of SWEs spend their whole day talking with LLMs and feel a personal connection/expertise with the thing that's also at the top of the hype cycle for non-technical news

If the author has preferences on their "own writing" that conflict with Anthropic's, then they should actually write it themselves rather than paying Anthropic to do it. Private companies don't owe you anything, even less so when they're beholden to laws in foreign jurisdictions.

> I chose to depend on a private company to express my own thoughts and now I'm mad that I'm not in control of the output

Who could have seen this coming???

Tokens measure "how much work the model did" in the same way that step counts measure "how far the person went"

GPUs "generate tokens" in the same sense that human feet "generate steps"

You can't compare token counts across different providers to get an absolute measure of "total work done" for the same reason that you can't compare step counts across different people to get an absolute measure of "total distance traveled"

Supposedly it was named "caml" because the author was a smoker who enjoyed camel cigarettes, hence the joke that [later implementations](https://en.wikipedia.org/wiki/Caml#History) were named "Caml Light" and "Caml Special Light"

This is more or less my belief too. But I think there are probably domains/companies where the "promo video" is actually good enough to ship (or at least, where that's the quality that devs were shipping before)

People don't write blog posts about how they wake up at 3AM to assign new tasks to their intern, nor do they build "orchestration frameworks" that involve N layers of interns passing tasks down between eachother

Also the process sounds like a nightmare: "it broke and I asked 4 different LLMs to fix it; my `AGENTS.md` file contained hundreds of special cases; etc." I thought this article was intended to be a horror story, not an advertisement

I'm pretty sure that "lack of market opportunity" isn't the reason that nobody has yet managed to invent eternal youth

It seems like they're targeting AI chip start ups with the promise of a compiler framework that can integrate with a custom backend

"matchmaking"

It seems a little disingenuous to call something written in an obscure Haskell dialect a "high performance UDP packet parser"

Money allows for economic specialization, leading to greater productivity. You don't build an economy that produces cell phones when you're stuck bartering goats for wheat.

I thought that Fortran was traditionally faster than C++ for numerical code due to stricter aliasing rules in the language, which I wouldn't expect to carry over to an IR?

Most big tech firms give engineers RSUs for this reason

Ponzi scheme is when the price collapses for something that I don't like, and the more that I don't like it the Ponzi-er it is

SWEs at Apple are not "barely able to afford rent"

Techies (at least at meta) are owners in the sense that a large % of their comp is in the form of equity, so it is in their interest for the company to maximize stock price.

Would you prefer subsistence farming?

There are overheads in server workloads that scale with the number of machines (network traffic, serializing/deserializing requests). There are also fixed costs per server that don't scale with core count, or at least scale sublinearly (storage, physical data center space, motherboard, ease of maintenance). So running 10 machines with 100 cores can be cheaper and more performant than running 1,000 machines with 1 core even if $/core is higher. And of course individual cores can be beefier: wider SIMD units, application-specific extensions like bfloat support for ML workloads, etc.

Of course Moore's law is slowing down, but cores/$ is an extremely silly metric to use

We're not talking here about being overly critical in a code review, or even criticizing someone who pushed bugs to prod. This is someone who intentionally violated company policy in a way that harms everyone who works there. It is not business as usual. Getting mad about this sort of thing is like getting mad about somebody keying your car or spreading false rumors about you: entirely reasonable and justified and not an interaction where professionalism is expected.

Increasing salaries will increase the cost of goods and services, meaning that more savings are required. It doesn't necessarily balance.

It's unclear to me how making these divisions dramatically increase their infrastructure costs (by either pulling them in-house or paying a premium to the original parent company) will "help them compete"

These models are very impressive, but the issue (imo) is that lay people without an ML background see how plausibly-human the output is and infer that there must be some plausibly-human intelligence behind it that has some plausibly-human learning mechanism -- if your new hire at work made the kinds of mistakes that ChatGPT does, you'd expect them to be up to speed in a couple of weeks. The issue is that ChatGPT really isn't human-like, and removing inaccurate output isn't just a question of correcting it a few times -- it's learning process is truly different and it doesn't understand things how we do.

The point is that ICs are judged based on their individual contributions, i.e. their own code, design docs, etc, whereas managers are judged based on their team's contributions

Stock traders already use ML models. "Replacing traders with ML models" means "making the job 'trader' into a job that develops ML models, rather than more traditional things like doing research on companies (or whatever)." My understanding is that this transition basically already happened over the course of the last two decades or so.

Obviously no, they shouldn't. You're not going to be able to keep that 6+ figure salary in a world where layoffs are illegal.

I'm personally happy to accept this volatility in return for higher comp and the opportunity to be part of a leaner, higher performing team. I don't want to work with people who are here to coast.

Clang support for AVX-512 intrinsics is pretty broken