this post was submitted on 06 Jan 2025
775 points (99.9% liked)
TechTakes
1533 readers
96 users here now
Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.
This is not debate club. Unless it’s amusing debate.
For actually-good tech, you want our NotAwfulTech community
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I was about to say, a selfhosted LLM means I'm not competing with every market analysis tool, customer service replacement, and 10 y/o kid bombarding the service with junk. It doesn't need to be ultra fast if I'm the only one using the hardware.
and who’ll supply the model and training and updates and data curation, dom? is it as manna from heaven? do you merely step upon the path and receive the divine wisdom of fresh llm updates?
fucking hell
Honestly, the data used to create these models was ripped from the public and I think that they are owed back to the public. OpenAI started as a non profit, and I think it should stay that way.
The FOSS model works well enough for other projects and I think that corporate AI will be exactly the same as the industrial revolution, progress at the cost of humanity. This isn't a problem to solve, it's a solution looking for problems.
Base open source model.
Topic expert models.
Community lora.
Program extensions.
Look what comfy UI + Stable Diffusion can achieve.
Base open source model just means some company commanding a great deal of capital and compute made the weights public to fuck with LLMaaS providers it can't directly compete with yet, it's not some guy in a garage training and RLFH them for months on end just to hand the result over to you to fine tune for writing caiaphas cain fanfiction.
And with the pruned llama models, it runs really quickly on a 2070.