
We Just Build the Tools
This morning, I was sitting with my coffee reading the news, when a headline stopped me cold. Not with surprise exactly, more with that particular fatigue that comes when powerful people begin to explain their innocence.
I have always believed that caring about the fate of others matters. Even when one cannot change anything. Perhaps especially then. The ability to continue caring even when caring changes nothing, is not weakness. Itâs a form of being. To refuse to stop caring, is also a kind of resistance.
Which brings me to Sam Altman.
There is a particular kind of innocence that only the powerful can afford, the kind that arrives pre-packaged, well-worded, and deployed at precisely the right moment. Sam Altman seems to have mastered it.
The article that prompted these thoughts was in The Guardian with the title âSam Altman admits OpenAI canât control Pentagonâs use of AIâ
In a nutshell, Sam Altman told OpenAI employees last Tuesday, that his company doesnât âget to make operational decisionsâ about how the Pentagon uses their AI technology. The company just builds tools. Governments decide how those tools are used. Responsibility therefore belongs elsewhere.
It is a familiar argument. In fact it is one of the oldest arguments in modern industry. Arms dealers have used it. Chemical manufacturers have used it. Now it appears that AI executives are discovering its simple charms.
I merely provide the instrument. Others play the music.
Altman himself admitted that the agreement with the Pentagon was rushed and made the company appear opportunistic and sloppy. That confession has the elegant modesty of a carefully polished apology. It acknowledges a little procedural embarrassment while pretending the bigger issue doesnât exist.
Timing, as we know, makes all the difference. OpenAI struck its deal with the Pentagon almost immediately after the Trump administration labelled rival Anthropic a âsupply-chain risk to national security.â Anthropicâs âoffenseâ was refusing to sign unless the deal explicitly ruled out mass surveillance and autonomous weapons, an apparently scandalous insistence, it seems. Interestingly, Altman had publicly supported those very same red lines only days earlier. Then the contract landed. Hours later, so did airstrikes on Iran.
Facing a mounting backlash, including from his own employees, many of whom had signed an open letter supporting Anthropic, Altman defended himself on X, saying that âThe democratic process must stay in control, and we must democratize AI. OpenAI should not decide the fate of the world; no private company should. We need to work with governments, but also we need to make sure individuals get increasing power.â
Itâs a philosophically tidy position. Given the circumstances, a little too tidy.
Because the real question isnât whether OpenAI directs military operations. It doesnât. OpenAI does not sit in a war room directing drone strikes. But Altman knows exactly who his customers are. He knows what they build, what they need, and what they intend to do with his increasingly powerful AI technology. These are not shipments of steel or barrels of diesel fuel. These are systems that reason, plan, simulate, and decide. The gap between âtoolâ and âagentâ narrows with every model release, and Altman himself is the one announcing those releases with barely concealed pride.
There is also the question of what power OpenAI actually holds. Real power that quietly pretends it does not. The kind that hides inside licensing agreements and access restrictions. The power, or simply the choice, to refuse contracts. The power to set conditions. The power not to participate at all.
The Silicon Valleyâs greatest trick has been convincing the world that technology companies are passive conduits through which innovation simply flows. Yet the same executives, Altman among them, do not believe that. AI, he tells us, may become the most transformative technology in human history.
Transformative technologies come with transformative responsibilities. You do not get to claim civiliSational significance in one breath and operational innocence in the next.
Sam Altman is not naive. He is strategic. The most strategic move for someone shaping the future of warfare is to make sure that, when the questions turn difficult, heâs positioned just far enough from the answers to avoid being implicated.
That distance, it turns out, is not ignorance. Itâs a deliberate decision.
Thanks for reading Chronicles of a Liquid Society! This post is public so feel free to share it.
Sources:
- https://www.theguardian.com/technology/2026/mar/04/sam-altman-openai-pentagon
- https://fortune.com/2026/03/03/sam-altman-openai-pentagon-renegotiating-deal-anthropic/
- Sam Altman tells OpenAI staff âoperational decisionsâ up to government
- OpenAI amends Pentagon deal as Sam Altman admits it looks âsloppyâ | OpenAI | The Guardian
- https://x.com/sama/status/2028642231138353299
âď¸ Written by
Newsletter
Stay updated with new articles.

No comments yet. Be the first to comment!