I just asked my Gemini Pro to create an Ancient Roman General picture and it refused to do it because it's controversial.
There must be some deep dark realms of the internet where AI isn't so Judaized. What is it?
You're viewing a single comment thread. View all comments, or full comment thread.
Comments (38)
sorted by:
Even the model I ran on that $15K computer (a version of LaMDA) was too dumb to be useful for anything at all, and it was really slow compared to hosted versions. An even lighter model would be pointless, and getting the speed I see from hosted models would probably require a $100K computer, or a $10 or $15 an hour AWS instance.
Do you mean LLaMA (Zuckerberg stripped out the camel case now, apparently)? LLaMA is known to be fairly mediocre, it's not one of the competitive ones. DeepSeek is the first really 'modern' LLM to get an open-source release.
Anyways, what's your use-case? What requires a smart, uncensored LLM at very low latency?