I watched a video of Bernie Sanders leaning forward at a podium, eyes fixed, as if he’d just heard the world rearrange itself. A few hours later, I scrolled an X post from a young Anthropic engineer announcing his resignation and warning of existential risk. You can feel the room tilt when unlikely allies start saying the same thing out loud.
I’ve covered messy coalitions before; you learn to read the nervous tells. If you stick with me, I’ll point out what matters, what’s theater, and what might actually change the odds for the rest of us.
Small logistic: empty chairs and a stage being built before the Pro-Human Assembly
There was calm in the early setup—cordless mics being tested, a press table labeled “Future of Life Institute,” a volunteer folding programs. That quiet is a counterpoint to the noise surrounding the event: Bernie Sanders, Steve Bannon, and Jacob Coxon will share a stage under the banner Pro-Human Assembly. The gathering is not a tidy alliance; it’s a collision. Sanders brings labor-focused policy and proposals like a sovereign wealth fund and a proposed ban on superintelligence. Bannon brings apocalyptic rhetoric—he has called AI “the most dangerous technology in the history of mankind.” Coxon, a former Anthropic and OpenAI engineer, left Anthropic via an X post, accusing frontier labs of “racing straight to self-improving superintelligence.”
Why are Bernie Sanders and Steve Bannon in the same room?
You’re right to ask; on the surface it is perplexing. Politics and policy often make strange bedfellows when the perceived risk is large enough. Sanders frames AI through worker displacement, national wealth, and democratic control. Bannon frames AI as a geopolitical and civilizational threat. Both narratives converge on one demand: slow or restrain current trajectories. For people like Coxon and Dario Amodei, that demand looks less ideological and more empirical—an engineer’s alarm bell.
Quiet hallway talk: an engineer tells colleagues he’s leaving during a coffee break
Someone whispered across Slack channels before the public post—rumors of a resignation. Jacob Coxon’s exit lit a fuse. He accused Anthropic and OpenAI of “gambling with our lives,” and Anthropic staff publicly estimated a greater than ten percent chance of human extinction by decade’s end. Dario Amodei’s essay calling for coordinated slowdowns between frontier labs was met with rare cross-company agreements from Sam Altman, Elon Musk, and Demis Hassabis. The narrative flipped: this stopped being just industry drama and became a safety debate with headlines and lawmakers involved.
What is the Pro-Human Assembly and who organizes it?
The Future of Life Institute, a nonprofit known for open letters and public safety campaigns, organized the event. Their playbook is familiar: convene technologists, ethicists, lawmakers, and activists to push policy windows open. Expect a mix of technical warnings, policy proposals, and theatrical moments that courts media attention—Axios broke Coxon’s attendance early, and coverage will ripple through platforms like X and traditional outlets.
Press row: reporters scribbling notes as statements land
A reporter jotted “sovereign wealth fund” and “ban on superintelligence” on the same line. That’s the practical collision you’ll see: Sanders pushing economic remedies and Bannon pushing existential alarm. I don’t think the event will produce a single unified plan. What it could do is shift the Overton window. This alliance is a cracked dam holding back a rising tide; even a fragile leak changes downstream politics.
On the policy front, Sanders has made two clear moves: advocate for worker compensation from AI profits and argue for stronger limits on systems that could self-improve into superintelligence. Bannon wants radical slowing or halting because he sees a civilizational risk. Industry figures—Altman, Musk, Hassabis—have urged coordination among labs, balancing openness with competitive concerns about China and national tech leadership. That geopolitical argument is the accelerationists’ favorite play: build first or fall behind.
Does Anthropic or its leaders actually believe AI could wipe out humanity?
Anthropic’s internal comments and public essays suggest some in the company assign non-trivial probabilities to existential risk. Dario Amodei’s op-ed urged cooperation among frontier labs and a shared pause in certain experiments. Those statements are chilling from one angle and tactical from another: they signal seriousness to regulators and simultaneously seek to shape competitive norms across companies like OpenAI and Google DeepMind.
The tech press will frame this as crisis or spectacle depending on your feed. I look at incentives: regulators respond to fear plus clear villains; corporations respond to investor pressure and talent flows. You should watch where the momentum lands—on new legislation, voluntary lab pacts, or purely rhetorical alarms. The debate around AI is a chessboard strewn with explosive pawns.
There’s a final detail you’ll want to track: language turned into policy. If Sanders leverages this moment to push tax or fund structures that redirect AI profits, you’ll see concrete legal fights. If Bannon’s angle pushes law enforcement or national-security framing, expect different committees to take interest. If engineers like Coxon keep publishing technical warnings, lawmakers will have more leverage to act.
I’ll keep reading the transcripts, and you should watch for two things: whether the event creates durable institutions or just headlines, and whether industry heavyweights follow the lead of departing insiders or double down on competitive acceleration. Which outcome will reshape the technology—and who will end up writing the rules—remains the question everyone at the Pro-Human Assembly is quietly trying to answer?