Not that it’s a good thing but don’t either of these companies have enough leverage to refuse with practically zero consequences?

4chan didn’t care because of the absurd logic and they still haven’t been blocked because they know it would be stupid to do so.

The ones being built or planned to be built will have to be sold off.

Probably not abadndoned, but land owner changing several hands in what is essentially someone buying a vacant lot. Maybe a handful will reuse any existing building already made, but most will be used for something new.

The existing ones will start their fight to the lowest market price which will probably cause many to either sell to a different cloud provider or change hardware to something more profitable.

None of the electronic hardware will go to waste because there will always be a demand for it, even if they sell it off at a loss. But any permanent fixtures like industrial cabling, ventilation, piping, etc probably will be wasted.

There’s still no guarantee on whether it’ll implode or just slowly deflate though. The natural assumption would be that the post IPO pricing will completely wreck sales and value, but they could easily move to an enterprise vendor lock scenario where they can charge whatever they want due to tech stack deals like how MSFT runs teams.

Remember, they’ll do anything in their power to keep the pumping going, even if it involves tax bailouts.

An algo/ai would only be profitable if you fed it insider information giving it enough time to make an actual informed decisions and not “fair” play.

Which obviously you don’t need an AI for.

It’s why the suggested strategy for us plebs is into throw money into something stable and come back to it years later with just enough ROI to stay ahead of inflation.

Haven’t touched 3D printing in a long time but about the only thing I remember related to the printer itself was doing a liquid cooling mod for the head which gave a more consistent structure output.

Otherwise everything revovled around slicing techniques & settings and snapoff structure points in the correct spots.

It’s not reliable enough to actually automate anything

This is half true. It’s not reliable enough to automate an entire job but it is reliable enough to automate tasks that would otherwise take a lot of time, usually related to sifting or searching data.

If I need to look through a massive set of data like Google for something thst I can only describe with an explanation, the LLM will do a much faster job actually finding what you need rather thsn spending an hour manually sorting through SEO slop.

You don’t even need the cloud models for this, you can slap SearXNG onto a local model at home.

It’s basically just an autocomplete search on steroids which is its biggest advantage. Any documentation you need is immediately accessible, which is especially useful if you have zero experience with something niche or new.

Now actually getting the LLM to consistently generate output is a completely different story lol.

We call that vibe coding.

Thst branding reminds me of the extensive complaints of Anubis beeing too weaboo for production applications lol.

The only thing I’m gonna try investing in from this AI shitshow is China’s CXMT RAM since they have a good chance of shanking both Nvidia and the RAM thug monopoly lol.

https://lemmy.world/post/39025760

wtf? Half the post is nuked even after being locked. I don’t even see how such a small community can be so stuck up about relevancy and purity washing selfhosted as if we all own our own DNS registrars and can do outbound SMTP.

I upgraded to an RTX Quadro 4000 on my media server (from a 750ti) and it still doesn’t support AV1 lmao.

Even last time I was on PC partpicker, the top beefy 10k USD media chonker machine was targeting a specific CPU for cheap AVX512 support because apparently it was required for heavy AV1 work, which I assume meant the GPUs couldn’t keep up.

Yeah they use kbdx so it works out of box for both. I think it even technically works for the original Keepass.

Haven’t used KeyPassDX much yet outside of some quick tests, but the autofill is really good because it shows up as a slim popup tab option on your keyboard where the typeing suggestions usually go.

Gonna do this soon as well. Used to use syncthing for auto photo transfers and I’m tired of Bitwarden’s crappy UI and terrible interop with autofill/autogenerate.

It’ll probably never happen due to the nature of KBDX, but I would kill to make a resilient native sync feature so that orgs wouldn’t be locked into proprietary vaults which drags you into vendor lock in when one of them starts to tank.

The two greatest problems in handheld assistants is speech recognition and command integration.

Gemini solves neither of those, and I was already complaining about Google’s crappy assistant long before the LLM era lol.

I could be wrong, but why hasn’t someone just bothered gluing openai whisper to something like voice attack from elite dangerous?

It’s been downloaded over 20 billion times. It supports 25+ protocols. It’s in cars, refrigerators, TV sets, routers, printers, phones, and every goddamn server on the planet.

Everything except my random podman container I need to test something on, but for some reason will have wget lmao.

Also a good time to mention you can use Ctrl+x Ctrl+e to edit your multi line commands in your default terminal editor so you can keep a clean, line separated command which is easy to read and follow.

I wonder how long they can keep the circus running though. They were obviously trying to vendor lock enterprise size businesses in before they eventually nuke the cost model, and Deepseek just preemptively slashed their cloud pricing probably knowing that everyone else will just migrate or even run on older cloud hardware which providers can sell by watt (and demand).

China might be behind on hardware, but their biggest wrench in the gears will probably just be popularizing their published LLMs which don’t require multiple nuclear powerplants to run lol.

Domain for $8 a year and 300Mbps fiber for $45 a month which snake ass AT&T keeps increasing in 5 dollar increments, so thank you for reminding me to call Spectrum for a quote so I can then call AT&T and harass them into giving me the correct price for another year.

Linus Torvalds: Nvidia, Fuck You!

We fixed the “glitch”

Hard disagree lol, the American OEM standard is a bar so far down you can see the sparks of hell. The improvement was just their initial attempt to catch up before they gave up.

They nuked the EPA regulations which is why everything in the US is an SUV now and they bypassed competition with Japenese OEMs by lobbying congress to make anti import laws (exactly like what they are doing right now for Chinese EVs) which is how we got all these hodpe podge 90s era hybrid deal brands like diamond star or mazda & ford.

By the time those brands finally entered the US market with local production in full, they had already learned the gg ez system from their American counterparts and began to follow the same crappy practices of reducing cost and quality on every possible corner.

I wouldn’t buy a Ford vehicle of this decade even if it ends up being cheaper because the thing is made of ABS plastic and Chinese aluminum glued together with the freshly harvested tears of their yearly department layoffs.

I’m kind of annoyed that it’s still used as a benchmark performance basis despite the game coming out in a buggy and subpar visual state lol.

Batman Arkham Knight being peak visual fidelity a decade later is really not so funny anymore when no one seems to use actual high fidelity games to compare instead of the latest EA or Ubisoft slop.

In a way it has actually.

Deepseek was big because not only did they publish the full model for everyone to use, but the MoE structure significantly brought down the hardware requirements in terms of processing power. As long as you have enough VRAM, you can run it on older hardware with no need for the latest Nvidia stuff.

Now they got v4 which many have found to be within a 10% margin of Claude and ChatGPT.

On top of that, China has cheapo VRAM GPUs available or soon to be released, like the MTT S80. Yeah it sucks as a Graphics card because the chip is behind, but you get 16Gb of GDDR6 for much cheaper than anything else.

But its not a conspiracy to fight China. The infinite scaling was just Nvidia solidifying themselves as the monopoly because they want all AI infrastructure to be dependent on them, which is why they still illegally export to China, despite an export ban attempting to reduce their potential competition.

Moore Threads (MTT) already has their own CUDA like system called MUSA, and I’m sure they’ll be happy to put in proper hardware support for new stuff like Bf16 and FP8/4. It’ll take a few years, but eventually China will catch up to the point where Nvidia gets shanked by cheaper hardware.