ChatGPT broke Diogo Almeida’s heart, according to him. Almeida, an OpenAI researcher, helped build the chatbot and invent reinforcement learning from human feedback (RLHF), a model-training technique responsible for the current age of AI. Despite its capabilities, he was disappointed, saying “We have lightning in a bottle, and yet it is not useful.”
Almeida told TechCrunch, “We have lightning in a bottle, and yet it is not useful,” and “I’ve been battling that problem since then. It took me a while to come to the conclusion: The problem is we are optimizing for human language … We have been super good at human language for four years, but it’s not useful for automation because computers speak a different language.” Two years ago, he left OpenAI to start TypeSafe AI, a startup trying to fix that problem.
This week, the company released a new transformer-based model, Jev, that is not a large language model (LLM). It doesn’t output text, but instead produces probabilities, or what the company calls “calibrated decisions.”
A New Approach
By eschewing language, the model becomes incredibly cheap and fast, and because users define the outputs in advance, it cannot hallucinate. Its output tokens are free, and input tokens are metered by the billion, not the million.
Developers are taking a great interest in the product; the company briefly lost the ability to serve users from its API because demand was so high. Jev appears most useful for software automation.
For example, Pranit Sharma, a software engineer at Vercel, said his company had used OpenAI’s ChatGPT Luna 5.6 to run a classifier to review commands for safety. When Vercel replaced OpenAI’s Luna with Jev, it got results five to 18 times more quickly and with greater accuracy.
Use Cases
Another developer, Nikhil Mudholkar, tested Jev against Gemini for classifying business emails. In his test, Gemini was slightly more accurate, but 10 to 20 times more expensive. More interesting to Mudholkar were Jev’s confidence scores — “it is the only one that hands back a real probability which makes it ideal for automating workflows!!”
Besides replacing LLMs in certain use cases, the new model can also augment them, acting as a smart check on misbehavior. Using agents to monitor agents can quickly become expensive, but using Jev to do so, Almeida argues, makes sense.
Armin Ronacher, the CTO of Earendil, which builds the open source model harness Pi, explained that Jev delegates the hallucination problem to the user. The user has to decide whether to disregard a result with low probability or act on one with high probability.
Future Possibilities
Another potential use for Jev is model routing, Ronacher said. Predicting whether a given workload requires a specific model would be useful, but using an LLM for the job would be expensive. Jev’s low cost and speed make that kind of real-time sorting possible.
Almeida hopes that the falling cost of intelligence will lead to its widespread deployment. He envisions smart software all over the place, in a way that’s emergent and distributed, much like the early internet.
Almeida is tight-lipped about the model’s architecture, which outside observers suspect is built on top of an open-weight LLM. The company refers to Jev as a “System One model,” focused on intuition rather than reasoning, and specifically focused on the right task.
Almeida said that his company is not a lab in the sense of betting on infinite wealth or building a religion, but rather focused on producing intelligence. He believes that the main product of his company should be intelligence, not fear or hype.
