AI Researchers Leave OpenAI, DeepMind and Meta to Found Neolabs
Researchers who worked on current AI models are leaving major labs and founding their own companies, which raise billions before they have a product.
Researchers who contributed to today's AI models are leaving the major labs in large numbers and founding their own companies. According to the report, departures are coming from OpenAI, Google DeepMind and Meta. The newly created firms, referred to in the industry as "Neolabs," are in some cases raising billions of dollars even though they do not yet have finished products.
The report covers 45 such Neolabs and sorts them according to their respective research bets. Most of them share the assumption that the existing recipe of ever-larger models and more computing power is not enough on its own. Instead, the new labs aim to overcome specific, named limitations of today's AI.
The development stands in tension with the industry's public statements. OpenAI already speaks publicly of an era of artificial general intelligence (AGI). Nvidia CEO Jensen Huang claims that AGI has already arrived. At the same time, co-developers of the current models are leaving their employers. The question of why people are leaving when they are supposedly so close to the goal therefore remains at the center of the debate.
Whether the alternative research paths will succeed is open. It also remains disputed whether AGI is actually near. Nvidia is seen as a reliable beneficiary of this development, since the large funding rounds for startups before the product stage also benefit demand for computing power. The report suggests that the movement is less a departure from AI research than a race toward different solutions.
Source: the-decoder.com
8
