1ba9115454

Thanks. I'll seek an alternative

I'm running a Ryzen 7 8745HS with Radeon 780M (27GB RAM), using Ollama's Vulkan backend.

I get ~45 tokens/s generation and ~680 tokens/s prompt processing.

The problem is ollama is not caching the prompt, I get -> forcing full prompt re-processing due to lack of cache data

So this wouldn't work in its current state in an agentic loop.

Nitter is/was a free, open-source alternative front-end for X (formerly Twitter) that lets users view public posts without an account, ads, or tracking scripts.

I can't imagine this setup will get more than 1 token per second.

I would love to see Deepseek running on premise with a decent TPS.

A tenary is all you need.

Why did you say you didn't play an instrument when you did?

What would be the outcome of this? Are you trying to reduce the gap between the poor and the wealthy?

I seriously doubt that YC discriminates on age deliberately.

My understanding is that you're expected to do 3 months in San Francisco and that would stop most people with a family. Which is like most people over 30.

I didn't look too deep at this as soon as it said JSON. I was gone.

I've been using dbmate which uses SQL and works really well.

The vendors performance metrics rate their own product the highest marks.

That's really cool.