The RTX/DGX Spark, Mac Ultras with 128GB unified ram are all ~$5k. Its still an expensive toy for rich people, it might as well be an H100 for 99.9% of the population (not devs with high paying jobs, of course).
the value of local models is allowing normal people to access AI without needing to subscribe to cloud services. this is esp imp for the rest of the world where even a 12GB gpu is extremely expensive.
there is no real viable local option that will come even close to Sonnet/Gemini Flash or the cheaper chinese models. Even if your pc costs <$2k you are never going to recoup the hw costs, and the results will be far worse.
qwen IMO is far better for coding, esp agentic coding when combined with something like Pi, it comes probably close enough to Sonnet for a lot of use cases.
Gemma family is better for almost all other tasks you'd use a local llm for.
Meta illegally scraped 80TB of data from Anna's archive, Libgen, Zlib etc. I'm sure other tech giants did too. Without paying them a cent, costing these projects $$$ in bandwidth/hosting etc.
when I hear people complain about these projects it just sounds like hypocrisy.
I've always liked powerlevel10k, or its equivalent tide for fish shell, which I much prefer over bash/zsh. Its fast, async, has everything you need, and is much easier to configure.
I've always wondered why someone doesn't just bundle a nice looking shell prompt with common nerd fonts and make it the default in a single package you can install.
Most websites send massive amounts of telemetry data which is never mentioned but somehow MS doing this for Windows/Edge etc, which they fully disclose, always is.
the value of local models is allowing normal people to access AI without needing to subscribe to cloud services. this is esp imp for the rest of the world where even a 12GB gpu is extremely expensive.
there is no real viable local option that will come even close to Sonnet/Gemini Flash or the cheaper chinese models. Even if your pc costs <$2k you are never going to recoup the hw costs, and the results will be far worse.