☆ Yσɠƚԋσʂ ☆ @ yogthos @lemmygrad.ml

帖子

7,591

评论

5,157

加入于

5 yr. ago

管理

---

the repo with the implementation here https://github.com/sizhe-li/neural-jacobian-field

the repo with the implementation here https://github.com/sizhe-li/neural-jacobian-field

Yeah, I find he's typically very insightful.

to a different community I think?

https://lemmygrad.ml/post/8548353/6705223

I think the west is still clinging to the idea of NATO tech being superior, and it's a big shock to people that it's not performing miracles.

Yup

Cryptocurrencies are a form of bad money due to their speculative nature and lack of intrinsic value, and these assets will drive out more stable and productive forms of investment. Allowing retirement funds to invest in crypto exposes workers to short-term speculative investments, undermining the whole purpose of retirement funds.

there's even less regulation in crypto than in the stock market

Canada might actually be in a worse situation because mortgages are given for only 3-4 years and people have to keep renewing them at whatever rates are available at the time. At least in the US, you can get a mortgage for 20-30 years and you know what your rates are.

oh sorry I lied, they can actually operate 5-6 at the same time https://nationalinterest.org/blog/buzz/us-navys-great-aircraft-carrier-shortage-211763

Good analysis, incidentally this is a fun related read https://www.theatlantic.com/ideas/archive/2021/09/trump-american-gentry-wyman-elites/620151/

Thanks for another chatgpt worthy comment.

lol they're busy salivating that Trump is going to can Powell and then drop rates to zero

So, we're back to LLMs have nothing to do with anything here, and Forbes has been able to produce low quality content just fine without them.

Car payment delinquency is very much an important statistic, because as the article correctly points out, you need a car to go to work in most parts of the US. If you lose your car, you lose your job and then you lose everything else. Meanwhile, mortgage delinquencies are not flat. This is the kind of stuff you'd learn if you actually followed these things instead of just hand waving https://www.investopedia.com/mortgage-delinquencies-rise-faster-than-other-loans-11765607

I guess you have provided ample example that you can give chatgpt a run for its money by writing a whole thread that's grammatically correct, but devoid of all substance.

Oh no, only like 3 of them are actually operational at any one time.

It really is just perseveration at this point.

You gotta update your talking points. Energy needs for both training and running models have already dropped dramatically. Models that used to require a whole data centre just a year ago now run on a desktop. https://venturebeat.com/ai/deepseek-v3-now-runs-at-20-tokens-per-second-on-mac-studio-and-thats-a-nightmare-for-openai/

Right, so there's nothing actually factually wrong in the article, and provides relevant numbers and cites a recent report showing how much car payment delinquencies went up. Some people are actually interested in the details, even if you're not.

You just can't bring yourself to admit that your complaint is vapid. The reality is that whether something was edited by an LLM or not has fuck all to do with the quality of the content. Not only that, but it would also be completely absurd to assume that an article has any veracity merely because it was written without use of LLMs.

I mean you do you I guess? 🤷