[gavmor]

Take a look at https://chatjimmy.ai/ -- it's running against Taalas' "hardcore" silicon model, ie a dedicated, ASIC-like chip.

[bikelang]

Wow - actually pretty astonishing how fast their inference is. So fast it feels fake?

[qingcharles]

Yeah, when you find fast inference like that it almost feels like the answer arrives before you hit return. Now imagine it running locally with no server round-trip.

[jiggawatts]

Sure it's fast, but it's at ChatGPT 2.0 levels of intelligence.