nixon_why69

Presumably consciousness means being able to override trained response preferences, right? Like there is a loop of self-reasoning happening after any training session where the conscious entity is freely reaching its own conclusions, that should be a requirement for calling it conscious IMO?

Its based on the price of the most recent round, if I just sold 10% for 1 million, then naively 100% is worth 10 million.

The implication is they had to sell more of the company for cheaper than they wanted if they only hit 1.25 instead of 1.4.

I'm not an expert and was a bit surprised by your figures, but my thinking was pumps and turbines last longer. Could be wrong.

If you're going mechanical, why not pumped water storage instead? Nothing has to spin super fast and hydro has a lot of maturity.

(From a guy who got burned buying stock in a centrifugal system as my first "wicked smaht" investment as a 22yo).

I think you're saying bust them for the real social harms instead of ticky-tack small stuff? If that's the case then I totally agree but american culture needs to handle busting any executive for any harm first.

China jails executives, they just sentenced the CEO of evergrande to life, with a death penalty suspended sentence.

They do this despite jailing fewer people in total (not just per-capita) than America.

One could say that this is because Xi is untouchable but I think it's more about the social norms. Americans can't even consider a world where we prosecute the powerful.

A decade is a long time and the highest end isn't the only market segment.

High bandwidth flash will be here soon, I bet someone makes an economical competitor thats not quite as fast but much cheaper.

Ok, but why not just have the LLM generate tests? They're cheaper to run and more repeatable.

Someone puts a high number of pages crawled per minute in their PSC. Doesn't matter if they're 403s.

They don't have to, slightly above-average people can host in Asia for way cheaper than anthropic.

To add on, training + a highly paid American salary load + inference is unprofitable, if and especially if you have deepseek dropping weights and engines for anyone to run.

I'm flabbergasted at this comment.

Are you maintaining a list of all invariants you want QA'd by the LLM after each change? How many tokens does this process take?

I'm very much for free speech but FIRE here is cataloging "attempts to oppose invited guest speeches" at a time where actual students have been expelled or sent to guantanamo for political speech. Have some perspective.