This video is a technical proof of concept showing an LLM running client-side in a browser using Llama.cpp compiled to WASM with WebGPU. Is this a dumb idea or do you think this idea can be improved to make client-side LLMs practical?
Over the weekend I built a technical proof of concept showing an LLM running client-side in a browser using Llama.cpp compiled to WASM with WebGPU.
I was bored this morning
In my latest video I try and build Ai using PHP instead of python
Nobody will get it, but it felt right-
Clients ask me to build very very small PHP APIs all the time, I don't want to use a full Laravel instance, so I hacked this together over the past few days. Is this a dumb idea? What might go wrong?
I've been slowly refining this codebase over the past 8 years. I started this in 2018 with JS for building micro-services when I was working at LegalZoom, since then I have implemented it at many different tech start-ups and corporates.