bird0861

7 ms·

31.
by bird0861 1y ago
https://ai.meta.com/research/publications/meta-large-languag...
32.
by bird0861 1y ago
Not true. Perhaps very few do, but some do in fact run them at 0. I've done it myself. There are many small models that will gladly perform well in QA with temp 0. Of course there are few situations where this is the recommended setup
33.
by bird0861 1y ago
Quite pleased you mentioned this. I would like to add transformer LLMs can be turing complete, see the work of Franz Nowak and his colleagues (I think there were at least one or two other papers by other teams but I read Nowak's the cl
34.
by bird0861 1y ago
Billboards shit in your eyes so this is a good comparison tbh.
35.
by bird0861 1y ago
https://www.youtube.com/watch?v=DiFtmrpuwNY https://mannlab.com/eyetap Dr Steve Mann who eventually led the Google Glass team (along with Dr Thad Starner, who like Mann was also a frequent poster in the MIT
36.
by bird0861 1y ago
You're really shrugging off an interactive device engineered by at least one (likely a ~dozen) person to be X percent more sycophantic and agreeable? Do you think the user experience is really so inscrutable? Guardrail models are a com
37.
by bird0861 1y ago
OpenAI should be held liable, they didn't accidentally create a model that glazes people, they went out of their way to do it. Many uncensored models are not as agreeable as ChatGPT tends to be.
38.
by bird0861 1y ago
Right, I can't sustain for a moment the idea that the guy who fumbled Recall like a stack of wet fish dipped in baby oil is actually a wise sage full of caution. I permit myself one foolish idea a day and that's not going to be th
39.
by bird0861 1y ago
Sounds good, doesn't work. Source: ask rail workers in the U.S. under Biden.
40.
by bird0861 1y ago
Please don't associate actual journalists GamersNexus with those hucksters.
41.
by bird0861 1y ago
Youtube will not win this battle.
42.
by bird0861 1y ago
This just looks like Android launchers of the past 10+ years. I'm remembering also Windows Longhorn leaks and Sun's Project Looking Glass. Ironic Apple gets good at hardware and then can't even build a UI or AI.
43.
by bird0861 1y ago
To be fair there are lots of Facebook, Instagram, and Youtube cargo cultists telling people to fine-tune on their documents for some reason. This got to be so common in 2024 that I think it was part of the pressure behind Gigabyte branding
44.
by bird0861 1y ago
Just to be clear, unless I'm misinterpreting this chain of comments, you do not want to fine-tune for information retrieval. FT is for skill enhancement. For information retrieval you want at least one of the over 100 implementations o
45.
by bird0861 1y ago
funny enough I am AuDHD but I just sm ke w*d (low basal dopamine gang gang) and no-life this stuff because I made the classic blunder of turning my hobby into my jobby
46.
by bird0861 1y ago
Aider, AIchat (both are TUIs), and crawl4ai are my daily essentials. I don't even bother with other coding agents or IDEs like Cursor although I did try it and others throughout 2024 but this year I'm just doing those. I would say
47.
by bird0861 1y ago
No. It's not microtubules. Enough with the g-darn microtubules already. https://www.biorxiv.org/content/10.1101/712794v1 "We used an antimicrotubular agent (parbendazole) and disrupted microtubular dyna
48.
by bird0861 1y ago
I think it's because they are betting on JEPA and they're trying to carve out brand identity by being special little Steve Jobs clones. They're taking "Think Different" to its neurotic limits. "See this is why
49.
by bird0861 1y ago
One would have to be absolutely cooked to consider using a censored model to translate or talk about anything a preschooler's ears can't hear. There are plenty of uncensored models that will run on less than 8GB of vram.
50.
by bird0861 1y ago
I would put myself in the 100x camp but my start in AI (and programming) was symbolic approaches in lisp/scheme, then I put that away for about a decade and got into NLP and then got into question answering and search/information