Breaking news out of the AI world, and this one is not a small percentage improvement.
TypeSafe just announced a new model called Jev, claiming it is 40 to 400x cheaper and 20 to 200x faster than existing frontier models.
Let that sink in for a second. Imagine your daily commute costs you $400, and suddenly someone hands you a way to get there just as fast (actually faster) for one dollar. That is not an incremental upgrade, that is a complete rewrite of the economics.
The person behind this is Diogo Almeida, and he is not some random founder. He was part of the OpenAI team that built the instruction-following methods behind ChatGPT itself.
His driving question for the last four years has been simple but brutal: models have been superhuman at chat for years, so where is all the automation?
Think about it. If these models are so smart, why do we still need humans babysitting every workflow? Why hasn't full task automation actually arrived yet?
That is exactly the gap Jev is aiming to close, not just a chatbot that talks well, but a model built to actually execute and automate real work at near-zero cost and blazing speed.
If these numbers hold up in the real world, this could reshape how every company budgets for AI, because the real bottleneck right now was never intelligence, it has always been cost and latency.
TypeSafe is not just shipping a model here, they are sending a message to everyone doing compute cost math: the equation just changed.
🔗 Original Source & Reference: https://typesafe.ai/blog/introducing-system-one-models-and-jev
Published automatically via FeedMind AI Content Pipeline.
Top comments (0)