ai_ja_nai

(64GB not VRAM, I meant) I also see people claiming fast performance on a 128GB machine, which is not exactly consumer hardware)

I am not getting it: I see a fp2 quantized model going on a 5090 with 64GB of RAM at 90 tops with -10% accuracy over original model. How is this supportive of the claims?

I don't understand who should be this article for. Europeans, that are doing everything according to the plan to reduce emissions? Chinese, who are going to save our neck by converting an entire nation that is 25% of world population to electrical/renewable/nuclear? Americans? It's the wrong news outlet, then.

what will happen when in 10 years they will want to discontinue those games? will they be hosting them forever? how are we going to preserve all the videogames production from 2028 on?

And I thought that it was related to Doom. Disappointing. (tongue in cheek)

Crazy difficult, but I guess a RL AI could learn how to beat it

Current LLMs are in the Pentium IV Netburst stage: hypertrophy of components on old archs have gotten to the limits of what is physically feasible. Get ready for the next phase of leaner models.

I remember how it was in 2008 on Linux desktops/laptops: AMD laughed straight in the face of their Linux customers. NVidia did support them, although at its own rules (closed source). When the need to run scientific computation over GPUs arouse, NVidia was there, ready. AMD got what it deserved for ignoring a market of scientists and researchers

Vim is a very good editor in the simple task of editing text files of every size. Wrote master thesis with it

I don't get this: the Transformer dates back to 2017, the patent is from 2018. This should be invalid as prior art by default.

Fabien Sanglard is an ace, his 3dfx posts are legendary. Very enjoyable reading

But this is not the point. The point is not to engage in meaningless or easy stuff only. The point is to recognize that you should not dedicate more than 8h per day because for many jobs (including software engineering) there are no deadlines or most are made by people that are not involved in the actual process so they are artificial limits imposed to make you fret.

I think the author is not stating "don't bear any responsibilities", but rather "work is infinite, do not try to finish it by dedicating 10 hours per day", which is a fallacy many people fall in (myself included)

> Why should that be so? I'd expect the all-in cost of a cloud machine to be less than my own hardware

Because cloud hardware doesn't have all the burdens of physically managing a real server. Replacing SSDs. Upgrading RAM. Logging to a iDRAC to restart a crashed server. All those things don't exist in the cloud and make you loose so much operational time. That's why clouds will ALWAYS cost more than bare metal. The cons is that with cloud you keep paying for the same servers: there are no assets anymore, only costs.