I essentially ran out of patience and tried another approach. It involved an LLM running C code so I could check the library output compared to my implementation to make sure it was byte-for-byte.
The C will never ship. I don't have practice writing C so I am very inefficient at it. I read it okay. LLMs are pretty decent help for this type of scrap code.
Should they add Voice Activity Detection? Are these separate filters or just making the whisper filter more fancy?
However BEAM is not the only factor in this process. the entire hardware platform as well.
This is after all a lot about that nice and huge cpu.
I mean when you have all 5000 started why not let the do some work? Stress test it with a few real life scenarios for 48h and let us see some number.
But if you are looking at a hosted erlang VM for a capex of one dollar then these folks are onto something
Cores really are the only way to escape the broken moores law - and this does look like a real step in the important direction. Less LLMs more tiny cores
But I will certainly try to leverage my telco-connection to get to play with more of their kit if I can.
Can someone explain this ideological whiplash?
Intel's death would be very embarassing to that whole effort. So Intel has incentives (survival) and the administration has incentives (jobs in the US). The method is "whatever can be claimed as a win".
No US party seems particularly capable or keen to hold an ideological line but especially not the GOP from what I've seen. Not saying other countries have particularly impressive parties either. I'm less than thrilled with ours over here.