His hairless skull gives him the air of a bonze, his steel blue eyes intimidate, his goatee-shaped beard adds geometric straightness to his face. His surname rhymes with those of the mathematicians or divinities of ancient Greece from whom he draws inspiration to name his companies, all linked to artificial intelligence (AI). Jonas Andrulis, 41, has for several months been the magnetic entrepreneur we see on the covers of magazines, the one with whom politicians like to be photographed. He is the embodiment of new German tech, associated with the emerging revolution of generative AI. The one which, like OpenAI in the United States with ChatGPT, programs conversational robots which seem capable of imitating human intelligence.
The company he founded in 2019 in Heidelberg (Baden-Württemberg), Germany, is called Aleph Alpha. Alongside the French Mistral AI, the German start-up is considered one of the great hopes for building an AI made in Europe. On November 7, it closed a historic fundraising campaign: 466 million euros, record funding for a German start-up, in the presence of the German vice-chancellor, Robert Habeck.
Among the investors are some of the biggest names in German capitalism: the giants Bosch and SAP and the Schwarz Foundation, owner of the European discount leader Lidl. Like ChatGPT, Aleph Alpha publishes artificial intelligence software based on large language models, capable of answering questions by exploiting a wealth of data.
The difference is that the service is offered not to the general public, but to companies or administrations, with their own data. “We are the next step in an Internet search. I can use it anywhere there is human-machine interaction. It’s the equivalent of a good intern, from whom we could ask for infinite research, based on a company’s data., explains Jonas Andrulis on “OMR”, a famous podcast dedicated to the German technology scene. Aleph Alpha already sells its software as a service to several large groups. The model allows them, for example, to write financial reports or create conversational robots that are experts in the functioning of the company.
Reduce the risk of “hallucination”
The security approach is the start-up’s strongest argument. Aleph Alpha promises its customers the “sovereignty” on their data and “AI respectful of European values and the fundamental principles of the rule of law”. Unlike ChatGPT, the service allows you to understand the path taken by the AI to give a particular response, which reduces the risk of“hallucination”that is to say false information.
You have 75% of this article left to read. The rest is reserved for subscribers.