Emil Michael doesn’t forget slights. That much became clear in a recently released podcast where the Department of Defense’s senior technology official sat down with Kleiner Perkins partner Joubin Mirzadegan and, with barely concealed irritation, revisited his forced departure from Uber back in 2017. When asked point-blank if he’d been shown the door, Michael’s answer was stark: “Effectively.”
It’s been nearly a decade, but the wound apparently hasn’t healed. Michael didn’t mince words when Mirzadegan pressed further, asking if he was still “salty” about the whole thing. “I’ll never forget that, nor forgive,” he said.
The conversation reveals something deeper than personal grievance though. Michael and former Uber CEO Travis Kalanick had bet big on autonomous driving as the company’s future. When investors forced them out, the self-driving ambitions went with them. Uber eventually sold its autonomous unit to Aurora in 2020 for what many saw as a fire sale price. Now Waymo’s robotaxis are operating across 10 U.S. cities, expanding rapidly, while Uber is nowhere in that game.
Michael blamed the investors’ short-term thinking. “They wanted to preserve their embedded gains, rather than try to make this a trillion-dollar company,” he said during the interview.
The Ghost of Autonomous Futures
You have to wonder what keeps people up at night more: being wrong about something, or being right but silenced. Kalanick has stayed oddly active in the space. Just recently he unveiled Atoms, a robotics company he’d been developing quietly for years. He’s also the largest investor in Pronto, an autonomous vehicle startup, and is apparently moving toward acquiring it outright. The man clearly never let go of the dream, even if he had to build it elsewhere.
But Michael’s gotten pulled into different wars. The interview, recorded before things fully exploded, captures him defending the Department of Defense’s ongoing dispute with Anthropic over access to large language models. And this is where things get genuinely interesting from a technology policy perspective.
When the Pentagon Meets Silicon Valley Values
Michael frames the DoD as incredibly constrained already. The department operates under such a suffocating web of laws, regulations, and internal policies that “we almost choke on them,” he told Mirzadegan. Anthropic, by contrast, wants to layer its own corporate values and ethical guidelines on top of all that.
The tension is obvious. Michael argued you wouldn’t accept such restrictions in any other context. “If you buy the Microsoft Office Suite, they don’t tell you what you could write in a Word document, or what email you can send,” he said.
Then he went further, invoking something Anthropic had itself published: Chinese technology companies have been hammering Anthropic’s models with a technique called distillation, essentially reverse-engineering the behavior closely enough to replicate capabilities. Through China’s civil-military fusion laws, Michael argued, that would hand the People’s Liberation Army access to something functionally equivalent to an unrestricted version of Anthropic’s model. Meanwhile the DoD would be hobbled by Anthropic’s self-imposed restrictions.
“I’d be one-armed, tied behind my back against an Anthropic model that’s fully capable, by an adversary,” he said. “It’s totally Orwellian.”
From Boardroom Tensions to Courtroom Drama
The dispute has since escalated dramatically. Defense Secretary Pete Hegseth labeled Anthropic a “supply-chain risk” in late February, and the government filed a 40-page brief in U.S. District Court arguing that giving Anthropic access to DoD infrastructure would introduce unacceptable risk. The filing suggested Anthropic could theoretically disable or alter its technology to suit corporate interests rather than national security in a conflict.
Anthropic fired back Friday with sworn declarations claiming the government’s case rests on technical misunderstandings and allegations never raised during months of prior negotiations. Thiyagu Ramasamy, Anthropic’s head of public sector, directly challenged claims that the company could interfere with military operations by disabling or altering its models.
A hearing is scheduled for Tuesday in San Francisco. This isn’t just corporate theater anymore. This is business bleeding into geopolitical strategy, with real infrastructure and real military capability hanging in the balance.
What strikes you most about this situation isn’t any single fact but the collision of incompatible worldviews. Michael’s asking why a company wouldn’t want to help its country win wars. Anthropic’s asking why it should surrender control over its own technology. One side sees patriotic obligation. The other sees corporate autonomy. Both probably think they’re right.
The question isn’t who wins Tuesday’s hearing. It’s whether either side will ever convince the other that they understand the stakes, or if this is just the beginning of a much longer fight over who actually controls the tools that will shape the next decade of warfare and security.
