this post was submitted on 23 Mar 2024
200 points (96.3% liked)
Technology
59421 readers
3519 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Am I the only one who feels dotcom vibes around this field?
I expect this to proceed similarly: many companies and funding dollars will burn in flames and still, the world will be a different place in a decade thanks to this technology.
Why am I feeling it isn't going to be a repeat of the standards-driven co-operative development supported by open source software infrastructure that occurred during the decade and a half after the dotcom bubble.. I have a feeling it would resemble the pre mass computing world of AT&T, GE and IBM.
There are a lot of open source LLMs being developed, ones you can run at home on your own data.
i hope these take off too
What would be the threshold for them to "take off"? It's all already out, so already there no?
its been a while, but last i tried it wasnt as good as the proprietary models.
Which ones did you try?
i tried the llama model for text, and another one meant for images i cant quite remember the name but it was one of the main ones.
are they any good now? running an llm actually sounds mildly useful.
The Mixtral models are pretty good, although they require a LOT of memory to run at a decent pace.
Honestly i think speed is something I don't care too much about with models, because even things like ChatGPT will be slower than Google for most things, and if something is more complex and a good use case for an LLM it's unlikely to be the primary bottleneck.
My ~~gf~~ private chat bot right now is a combination of Mistral 7B with a custom finetune and ~~she~~ it directs some queries to ChatGPT if I ask (I got free tokens way back might as well burn through them).
How much of an improvement is Mixtral over Mistral in practice?
Sillytavern by any chance?
And I'd say the difference between mistral and mixtral is pretty big for general usage, feels like it's a next generation.