Aletheia Research, a low-profile San Francisco lab founded by former OpenAI policy lead Tasha Reinhardt, has hired six more former OpenAI researchers in the past quarter, bringing its share of OpenAI's defunct superalignment team to 14 of the original 27 members.
The lab, which has raised approximately $340m across two funding rounds, is unusual in the current landscape for being focused exclusively on what it calls "deployment-time alignment" — the question of how to keep AI systems behaving as intended after they leave the training environment.
A different bet
Where most well-funded labs of Aletheia's vintage have pivoted toward building competitive frontier models, Aletheia has explicitly disavowed that path. "We're not in the model race," Reinhardt said in a rare on-the-record exchange last month. "We're in the deployment race. They are not the same race."