Probably raises $9M to build a more reliable kind of AI

Editor's take

Probably has secured $9 million in seed funding to develop AI models designed to significantly reduce factual inaccuracies and hallucinations. This initiative directly addresses a core limitation of current generative AI, aiming for a reliability that rivals traditional deterministic computing.

The pursuit of factual accuracy is critical for the widespread adoption of AI in sensitive sectors like healthcare, finance, and education, where incorrect outputs can have severe consequences. Probably's focus on deterministic-like accuracy positions it against other startups and large AI labs, such as Google's Gemini or OpenAI's GPT-4, which are also striving for more dependable AI.

Future developments to monitor include Probably's specific technical approaches to achieving this enhanced reliability and its performance benchmarks against established models on real-world tasks. Success will hinge on demonstrating a quantifiable reduction in errors without sacrificing the generative capabilities users expect.