Diogo Almeida, one of the pioneers behind ChatGPT, has launched a new low-cost artificial intelligence system named Jev that aims to challenge the dominance of established models like ChatGPT and Claude. Almeida, who helped develop the reinforcement learning technique used in ChatGPT, introduced Jev earlier this month through his company TypeSafe. The system has rapidly gained attention, sparking speculation about valuations reaching as high as $10 billion.
Almeida, a former researcher at Google Brain and OpenAI, has expressed disappointment with how current AI models have fallen short of delivering an economic transformation. Unlike text-based chatbots designed primarily for human interaction, Jev is built to serve machines, providing responses in numerical form rather than natural language. This distinction allows Jev to operate orders of magnitude faster and more cheaply than existing “frontier” models from OpenAI and Anthropic, companies currently valued in the trillions.
The use cases for AI are shifting from consumer-facing chat functions to machine-to-machine applications such as routing customer service emails, screening job applications, and detecting fraudulent financial transactions. These automated systems require near-instantaneous results, where delays of even a few seconds can be costly. Jev addresses this need with response times measured in nanoseconds, and its pricing reportedly makes it up to 44.5 times cheaper than rival offerings.
A notable demonstration showcased Jev’s speed in real-time control of the video game Doom, illustrating its potential for robotics and other applications demanding rapid computation. Jev’s developer claims the model is 19.4 times faster than current competitors, a combination that has driven unprecedented demand, leading the company to temporarily suspend new signups.
Industry observers have noted the rapid emergence of copycat models following Jev’s debut, indicating a new wave of budget AI systems that could reshape the market landscape. Analysts from HSBC warned such developments might trigger a market sell-off analogous to the DeepSeek incident in early 2023, which affected shares of companies like Nvidia and AMD due to concerns over reduced demand for costly computing resources. However, they also suggested such fears may be misplaced, as lower costs could fuel greater overall AI adoption.
Almeida draws inspiration for Jev’s name from the Jevons paradox, a theory in economics stating that increased efficiency in resource use can lead to higher overall consumption. He envisions Jev driving a significant expansion in AI utilization, potentially sustaining the industry’s growth even as high-priced models face disruption. While Jev’s long-term impact remains to be seen, it represents a shift toward faster, more affordable AI solutions focused on powering machine intelligence.