@nt
okay a few hours later and I have performance comparable to that of Vapi/Elevenlabs agent SDK, albeit along the green path - I'm sure there are hundreds of edge cases that these companies spend much time figuring out, not to mention making their offerings flexible/observable etc
but it's crazy how quickly you can get to fairly performant, somewhat deep tech from the comfort of your living room with a ~$100 in LLM spend
17 replies
17 recasts
171 reactions