@thomasfuchs I would love to read the training data needed to make it do that.
my favorite is when you try to guess the help command and the utility says “that’s not a command I recognize, you fool. here’s the help listing, you pathetic lost soul” https://fedi.astrid.tech/objects/eac10929-34bd-4d6e-9ca5-22291158fda3
@darkrat OMG I have given up and just clear everything past a few months (media storage wise). It's a leaky engine and needs constant tape but it pretty much gets the job done
I know I upgraded recently but my computing power feels so low compared to what folks use these days. Honestly I could probably cope with 8GB RAM even but having even 32 feels like such luxurious excess. It does mean a lot of chromium based react apps can siphon more resources tho. 😝
For real even a few GBs of VRAM extra would make some stuff so much easier. I hate that LLMs make me want a beefier computer.
@jalcine Like LLM models? It has been a small hobby of mine to tinker with them if you're into chatting about em. I like the 3B ones that can run fast on my shitty computer
@erlend Yeah, I used to bring my cardboard around to conferences and show off my web based dev environment. These days I'm working towards using "steam deck" form factor computers with consumer head mounted displays like the Rokid Max and am working towards using @stardustxr for the spatial computing part
CS Researchers: Here're some programming languages mathematically proven to make the computer do exactly what you say.
Programmers: «create vast complexity that no one person could comprehend in order to render a web page»
Also Programmers: What if we made the programming language English, required that the compiler run on a monstrously huge array of GPUs, and set it up so it only did what we ask sometimes, and other times just do nonsense, but we don't know which is which?
Hacker News: Neat!
One thing I find sad is that a lot if folms treated VR as just another tech fad and ditched it for crypto when that became the fad. Now a bunch of the same folks are following AI. Especially sad when fad chading meant ditching code and community for the next best thing only to abandon it again. It feels like the fad chasing means the fad in question doesn't get a fair chance to actually accomplish it's potential.
Slack is now using all content, including DMs, to train LLMs.
They offer a thing they're calling an "opt-out."
The opt-out (a) is only available to companies who are slack customers, not end users, and (b) doesn't actually opt-out.
When a company account holder tries to opt-out, Slack says their data will still be used to train LLMs, but the results won't be shared with other companies.
LOL no. That's not an opt-out. The way to opt-out is to stop using Slack.
https://slack.com/intl/en-gb/trust/data-management/privacy-principles
@ww Yeah that's totally understandable. Having structs documented and their flow through the code is super handy. Lately I've been using classes a lot in vanilla JS and the clarity in structure has been nice. Not super OOP tho
@ww Yeah I feel you, had the same situation in the past. Tbh tests have been good enough for catching most issues for me and when I really want types I can go use a typed language that was actually made for it 😅 TS is neat but it ends up limiting you when what you really want is JS
shitpost
carpenter who's so fed up with the constantly changing wood specifications, and the way the 1 year old wood saw is not compatible with the latest wood version, and the people who were gifted a small table for no reason suddenly showing up with a huge questionnaire and wanting the table to be certified for potentially safety-critical industrial use free of charge just because, the carpenter just leaves all of that behind and starts a new career with computers.
Occult Enby that's making local-first software with peer to peer protocols, mesh networks, and the web.
Exploring what a local-first cyberspace might look like in my spare time.