Jev founder Diogo Almeida argues decision models come after RLHF
In a 17-minute talk clip, Jev founder Diogo Almeida claims Jev is 200x faster, 400x cheaper and has zero hallucination, and argues the field should move from human preference alignment to calibrated decisions.
Jev Founder, Diogo Almeida (ex-OpenAI): "Claude Code is not the next era. It is still the assistance era - Jev is what comes next Jev is 200x faster, 400x cheaper and has 0 hallucination. This is the step after RLHF" In 17 minutes, he shows why every model you use is built to please you Human preference → Correct answers → Calibrated decisions → Real automation Your model is not lying, it is doing exactly what it was rewarded for Worth more than most $500 Claude Code courses Bookmark this and watch it today
This post quotes another post: x.com/distortgeekin/status/2099117449232699548