Hacker Newsnew | past | comments | ask | show | jobs | submit | seizethecheese's commentslogin

> genocide datacenters

AWS is one of the biggest military contractors, they're a valid target


Moral condemnation has little coercive force unless it is backed by material power.

It’d like to hear more

Just to be sure that I understand, you're saying that your model "can't hallucinate" because it only outputs a single thing, right? In this way, an LLM can't hallucinate either if I prompt it to do a classification task with a discrete set of possible outputs, right? (Assuming I reject non-conforming output. Actually, maybe what you're saying is that your system can't output non-conforming output?)

Yeah that's precisely correct.

For e.g. classification tasks, even in 2026 people are doing things like hallucinating deliberately, and then matching the hallucinations to embeddings -

https://softwaredoug.com/blog/2026/08/10/hypothetical-classi...

With TypeSafe it just picks the class (actually probabilities across classes), reliably every single time.


404?


You do realize the limited liability corporation was a key innovation that unlocked the Industrial Revolution, right?

Companies definitely respond to fines or liability. They just need to be big enough.

For example, I recently heard an interview from an environmentalist who expected to be outraged touring a Chevron drilling location but was surprised by how much precaution is taken these days. Basically, liability for oil spills is massive. We could just make data leak liability massive too.


That works for Chevron because they have enormous assets to lose.

In the ID company case, there simply aren't enormous assets available, despite enormous damage being possible. As such, we really need to re-think just how much we limit liability.

Perhaps it's time to stop allowing degenerate gamblers to freeroll their risks... perhaps it's time to start zeroing out investors, so that they have to start behaving responsibly.


Quality already is the norm for many products. For example, I own Sennheiser headphones, an iPhone and a MacBook. (Sure maybe not the absolute top of the line, but definitely more expensive than commodity products.)

I find great utility in cheap low cost products for things that I don’t know I’ll use more than once. For example, I’m glad I bought the cheapest roller blades a few years ago because I used them twice. I also buy a lot of cheap toys for my kids, they may use them only a few times and also quantity really has a certain quality (novelty is good for kids).

I find the moralizing and lecturing around quality to be distasteful. These holier than thou types really think they know better than the buyer.


> For example, I own Sennheiser headphones, an iPhone and a MacBook. (Sure maybe not the absolute top of the line, but definitely more expensive than commodity products.)

If anything, all of these are probably in the sweetspot for value. I have a hard time calling them the norm, at least not unless you're in the know and are knowingly looking for good value. I had to go through 3 sets of headphones before I got a Sennheiser set I was happy with and lasted 10 years. There are countless shitty Android phones and Windows laptops out there, and people do buy them.


In what case is capex (for any company) not required for the end product? Are you arguing capex is disingenuos as a concept?

In your car example the training is much like setting up the manufacturing line.

I think the issue here is that the capex depreciates super fast since the models obsolete really fast.


If my company sells widgets and dongles, capex for the dongle line does not figure into widget COGS

Both world wars dramatically sped up technological progress, what are you talking about?

Which is it? Would the regulations slow down competitors or let them catch up, or are you contending it would let American competitors catch up but Chinese ones not?

That’s a charitable way to interpret what they’re trying to do in this letter. I would say they are trying to maintain the status quo to protect their status system.

I don’t think they make a coherent argument here? This seems to hinge on some argument that because AI doesn’t properly explain its breakthroughs, therefore the breakthroughs are less fertile for human understanding? This makes no sense. Why wouldn’t these under-explained breakthroughs be extremely fertile soil for explanations?

Imagine time traveling back in time and offering Leibniz a packet of proofs from the intervening years, but with the caveat that there would be no explanations. Would he say no?


Because the fertile soil comes from the process of reaching the breakthrough. This should be fairly obvious.

I don’t think it’s obvious. In my Leibniz example, mathematical understanding surely would have accelerated right?

> Would he say no?

And that’s the issue the article is trying to explain.


My thought experiment was trying to extract the issue from the zeitgeist around AI, so not really

There’s answer for you there if you’re willing to see it, whether you agree with it or not. So there’s that.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: