If your CPU isn’t ancient, it’s mostly about memory speed. VRAM is very fast, DDR5 RAM is reasonably fast, swap is slow even on a modern SSD.

8x7B is mixtral, yeah.

Mostly via terminal, yeah. It’s convenient when you’re used to it - I am.

Let’s see, my inference speed now is:

• ~60-65 tok/s for a 8B model in Q_5_K/Q6_K (entirely in VRAM);
• ~36 tok/s for a 14B model in Q6_K (entirely in VRAM);
• ~4.5 tok/s for a 35B model in Q5_K_M (16/41 layers in VRAM);
• ~12.5 tok/s for a 8x7B model in Q4_K_M (18/33 layers in VRAM);
• ~4.5 tok/s for a 70B model in Q2_K (44/81 layers in VRAM);
• ~2.5 tok/s for a 70B model in Q3_K_L (28/81 layers in VRAM).

As of quality, I try to avoid quantisation below Q5 or at least Q4. I also don’t see any point in using Q8/f16/f32 - the difference with Q6 is minimal. Other than that, it really depends on the model - for instance, llama-3 8B is smarter than many older 30B+ models.

Have been using llama.cpp, whisper.cpp, Stable Diffusion for a long while (most often the first one). My “hub” is a collection of bash scripts and a ssh server running.

I typically use LLMs for translation, interactive technical troubleshooting, advice on obscure topics, sometimes coding, sometimes mathematics (though local models are mostly terrible for this), sometimes just talking. Also music generation with ChatMusician.

I use the hardware I already have - a 16GB AMD card (using ROCm) and some DDR5 RAM. ROCm might be tricky to set up for various libraries and inference engines, but then it just works. I don’t rent hardware - don’t want any data to leave my machine.

My use isn’t intensive enough to warrant measuring energy costs.

The article isn’t about automatic proofs, but it’d be interesting to see a LLM that can write formal proofs in Coq/Lean/whatever and call external computer algebra systems like SageMath or Mathematica.

it cuts out the middle man of having to find facts on your own

Nope.

Even without corporate tuning or filtering.

A language model is useful when you know what to expect from it, but it’s just another kind of secondary information source, not an oracle. In some sense it draws random narratives from the noosphere.

And if you give it search results as part of input in hope of increasing its reliability, how will you know they haven’t been manipulated by SEO? Search engines are slowly failing these days. A language model won’t recognise new kinds of bullshit as readily as you.

Education is still important.

Sure, they’re just typically from the same industry with similar perspectives, similar blind spots and similar affinity for rants on topics X, Y and Z. Some get annoyed by this after a while so ignoring comments is a valid choice if you feel like that.

Probably. Still, it might be useful for blog discovery. One doesn’t have to read the comments.

Many such articles are reposted on HackerNews occasionally: https://news.ycombinator.com/

Alas, their IP requirements are too much for me.

Anyway, is there much of xenharmonic sheet music there? Take, for example, Easley Blackwood (but not his books).

I have a MediaWiki instance on my laptop (I’ve found the features of all other wikis/mindmaps/knowledge databases decisively insufficient after having a taste of MW templates, Semantic MediaWiki and Scribunto).

Also some smaller things like pihole-standalone, Jellyfin and dictd.