AI voice assistants finally feel conversational

Mina Chowdhury · 11 likes

Benchmarks look great but real world usage is where it falls apart for me

Cedar Fern · 6 likes

Depends but yeah mostly

Curious Lantern · 6 likes

Context window claims are always optimistic, test it yourself

Oliver Smith · 0 likes

Haha my thoughts exactly

Danylo Kovalenko · 0 likes

This ^

Oliver Smith · 3 likes

Evals are getting saturated, we need harder ones

Zofia Kowalska · 7 likes

Well said

Elsa Andersson · 4 likes

Well said

Amelia Hughes · 1 likes

Prompt sensitivity is still the biggest pain in production

Quiet Runner · 14 likes

I read that twice and still chuckled

Martina Fernández · 0 likes

Open weights or it didnt happen

Gentle Sparrow · 0 likes

Not sure I agree but ok

Solar Orbit · 0 likes

Running a quantized version locally and its surprisingly usable

Minji Kim · 9 likes

Fair point

Lina Al Harbi · 0 likes

The cost per token matters more than the leaderboard imo

Sophie Gagnon · 0 likes

Yeah same here

Emir Kaya · 0 likes

My GPU just cried reading this

Luca Bianchi · 0 likes

Yeah same here

Khalid Al Zahrani · 0 likes

Not sure I agree but ok