Diogo Almeida introduces Jev: AI model 20–200× faster and cheaper, optimized for decision-making
Diogo Almeida, co-inventor of ChatGPT, announced a new AI model called Jev after two years of stealth development. The model is claimed to be 20–200× faster and 40–400× cheaper than current solutions, with free output tokens. Jev is optimized for decision-making tasks rather than text generation and operates using a new training approach called RLCD.
This looks almost like magic, but I've also had the feeling for a while that the AI world is in a tuning phase rather than a revolution. But this actually looks like a revolution instead.
How does Jev differ from current large language models?
Jev focuses on decision-making tasks and parallel computation rather than sequential text generation. It lacks the ability to generate extended text output, but achieves higher efficiency and lower costs according to its creators.
What is the practical cost of running Jev?
Input tokens cost $0.042 per million tokens, and output tokens are free. At approximately 10 calls per second, operation costs around $7 per hour.
Why is the model named Jev and what does the name reference?
The model is named after Jevons paradox, an economic principle explaining why increased efficiency of a service often leads to higher consumption. The name carries deep economic symbolism.
- Comparing AI for web creation - one prompt, 11 models, vastly different results — netlify.com 71 % match
- Static vs. dynamic languages for AI coding: complex problems shrink efficiency differences — danluu.com 70 % match
- Understanding is the new bottleneck — geoffreylitt.com 69 % match
- Diogo Almeida
- Jev
- ChatGPT13
- RLCD
- TypeSafe AI
- Jevons paradox