Harmonic, CEO of Robinhood Ai Math Startup, calls Chatbot, Amnesty International

Harmonic, a start -up company of artificial intelligence in which the CEO of Robinhood Vlad Tenev participated on Monday the Beta launch of the iOS and Android Chatbot application where users can access the AI, Aristotle.

With this launch, the company aims to expand the scope of access to Aristotle, which is called by the consensual to provide “hallucinogenic” answers to questions that include mathematical thinking-bold demands given reliable problems of artificial intelligence models today. Harmonic focuses on creating “sports sport” or MSI; Finally, you want to help users in all mathematics -based fields, including physics, statistics and computer science.

“(Aristotle) is the first available product for people who think and is officially achieved from the output,” said Harmonic CEO and co -founder Tudor Achim in an interview with Techcrunch. “Among the areas supported by Aristotle, which are the areas of quantitative thinking, we are already guaranteed that there is no hallucinations.”

In the end, Harmonic says she is also planning to issue an applications programming interface to allow institutions to reach Aristotle, as well as a web application for consumers.

The launch of Beta comes just a few weeks since Harmonic collected $ 100 million in a tour of the series B led by Kleiner Perkins, with a value of $ 875 million. Achim claims that Harmonic “is progressing very quickly” on his way to MSI, and that investors believe this was a fair assessment due to the start -up scope of startup.

Achim says Harmonic achieves its excessive excessive solutions by having Aristotle to produce responses in the open source programming language. Before Aristotle provides an answer to users, he says the model is verified that the solution is correct through an algorithm that does not include artificial intelligence. Harmonic CEO indicates that a similar technology is used to verify outputs in high risk fields such as medical devices and aviation.

Even in a narrow field, achieving hallucinogenic performance from the artificial intelligence model is a very difficult task. Studies have found it Even leading artificial intelligence models a lotAnd the problem does not seem to be improving. The latest thinking models of artificial intelligence from Openai are more hallucinations than their most older models.

Harmonik says that Aristotle achieved the performance of the gold medal at the 2025 International Mathematics Olympics through an official test (which means that the problems were translated into strength coordination. Artificial intelligence models that achieved the performance of the gold medal in the International Maritime Organization this year have been developedBut through informal tests that were conducted in the natural language.