A clip of Anthropic CEO Dario Amodei answering a question about AI and military targeting has passed 885,000 views on X. The post went up on November 9, 2026 and gives no date or source for the interview. The exchange is not new. It comes from a 70-minute Bloomberg interview with Emily Chang, published on June 17, 2026. The figure in the question, 1,000 targets a day rising to 5,000, traces to one unnamed official quoted in WIRED in March.
This page puts the clip back in its interview. It gives the question as it was asked, the full answer, the source of the number, and what Amodei said next about a strike on a school in Iran. Every quote here was checked against Bloomberg’s own captions.
On this page · 7 sectionsOpen
A clip of Dario Amodei on AI and military targeting passed 885,000 views on X. It was posted on November 9, 2026 with no date or source.
The exchange comes from Emily Chang’s 70-minute Bloomberg interview, published on June 17, 2026. It starts at 41:47.
The figure of 1,000 rising to 5,000 targets a day traces to one unnamed official quoted by WIRED on March 23, 2026. It describes capacity, not strikes carried out.
The post’s text quotes the first half of the answer. In the second half Amodei says military operations are not Anthropic’s decision.
In the same interview he named two limits for military use of Claude: no mass surveillance and no fully autonomous weapons.
§ 01What the post says
The post is from an X account called @sachin2. It went up at 04:57 UTC on November 9 with a video 1 minute 40 seconds long. Read by its id at 20:35 UTC on November 10, it had 885,247 views. The post names no interviewer, no outlet and no date.
| Date | What happened | Source |
|---|---|---|
| March 23, 2026 | WIRED publishes a book excerpt by Katrina Manson. One unnamed official gives the 1,000 and 5,000 figures | WIRED |
| June 17, 2026 | Bloomberg Originals publishes Emily Chang’s 70-minute interview with Amodei. The exchange starts at 41:47 | Bloomberg Originals on YouTube |
| November 9, 2026 | An X account posts a 100-second clip with no date. It reaches 885,247 views by November 10 | The X post, read by id |
§ 02The question, as asked
Chang opened with a number. Her words, from Bloomberg’s captions: “A US official has said with the help of LLMs, the US military has gone from being able to hit a thousand targets a day to 5,000 targets a day.”
Then she drew a conclusion: “That means Claude can help kill more people more quickly.” Her question to him was: “Are you comfortable with that?”
Two details matter here.
- The headline is the interviewer’s line. The post opens with the words “Claude can help US kill more people more quickly” and a question mark. That line is built from Chang’s question. Amodei did not say it.
- Capacity, not a count. Chang said the military had gone from being able to hit a thousand targets a day. The post’s text shortens that to “gone from hitting 1,000 targets a day to 5,000”.
§ 03The full answer
You can watch the exchange in Bloomberg’s video at 41:47. Amodei split the question in two. The first part was about military capability. It is the part the post quotes.
- On military strength. He said he supports the United States being more effective militarily. His reason: “I think having that ability be stronger doesn’t cause wars.” He added: “It deters wars.”
- On the country. He turned the question around: “Do you want this country to be a more powerful actor rather than a less powerful actor on the world stage?” His answer was: “I do, I’m a patriot.”
The second part is not in the post’s text. He called it a separate question: whether he supports particular policies of the US government.
- Not his decision. He said he supports some of those policies and not others. Then: “It’s not up to me.”
- No say over operations. His words: “If we provide a technology, it’s not up to us to say, you can do this military operation and you can’t do that military operation.”
- Private views stay private. He said he might privately believe one operation makes sense and another is a bad idea, “but we’re not gonna deny the technology.”
- What a supplier can do. He said: “What you can do is to assert some high level boundaries”. He described them as limits on the uses that go against the company’s values and the country’s.
He had named those boundaries about three minutes earlier in the same interview: “our red lines of mass surveillance and fully autonomous weapons”. Of those two uses he said: “It’s not worth democracies winning if democracies do those things.” A federal appeals court later ruled on the dispute those two limits caused. Our post on the Pentagon ruling covers it.
| Line | The X post’s text | Bloomberg’s captions |
|---|---|---|
| The figure | “With the help of LLMs, the US military has gone from hitting 1,000 targets a day to 5,000.” | “the US military has gone from being able to hit a thousand targets a day to 5,000 targets a day.” |
| Support | “I am supportive of the US being more effective militarily.” | A paraphrase. He said the United States can be more effective militarily and that he supports that ability |
| The patriot line | “Do you want this country to be a more powerful actor on the world stage? I do. I’m a patriot.” | “Do you want this country to be a more powerful actor rather than a less powerful actor on the world stage?” |
| The second half | Not quoted | Operations are for military decision makers. Anthropic sets a few high-level limits |
The question and the whole answer run from 41:47 to about 43:27 in Bloomberg’s video. That is 100 seconds, the same length as the clip. So the video in the post may well carry the full answer. We have not matched it frame by frame. The text of the post, which is what most people read, quotes only the first half.
§ 04Where the 5,000 figure comes from
The number is from a WIRED book excerpt by Katrina Manson, titled Meet the Gods of AI Warfare and published on March 23, 2026. Her sentence reads: “An NGA official told me that with the help of Maven’s computer vision the US went from being able to hit under a hundred targets a day to being able to hit a thousand.”
She then writes that as language models were added to the Maven platform, the official told her “that number had risen fivefold to 5,000 targets a day”. NGA is the National Geospatial-Intelligence Agency. Maven is the US military’s AI targeting platform.
Four things about that source:
- One official. The article does not name the person.
- A reporter’s account. It is what the official told her. It is not a published statistic.
- Capacity. The phrase is being able to hit. It is not a count of strikes carried out.
- No model named. The passage credits language models in Maven. It names no model. The link to Claude is made in Chang’s question.
§ 05What he said next: the school strike
Chang’s next question was more specific. She said Bloomberg had reported that the US military was using Claude in the war in Iran for AI-assisted targeting, through Palantir’s Maven Smart System. She said a US missile reportedly hit a girls’ school in Iran in February, killing more than 150 people, most of them children. Then she asked: “Did Claude play a role in that strike?”
- He said he does not know. Amodei said Anthropic does not have access to how the models were used. Later he said: “I don’t know what role Claude or any other AI had”.
- He called it terrible. His words: “this is a really terrible thing to happen”.
- He said it was within his limits. He told Chang the use she described “doesn’t even violate our red lines”. He also said he thinks the military’s use of these models is appropriate overall.
- His account of the system. He said: “Claude assists, but a human makes the final call.” That is his statement. This page has not verified how the system was used.
Chang then asked whether AI warfare makes a war between the US and China more or less likely. He answered: “I would say on balance it is more likely to stop it.” He added that with no limits on how it is used, AI could make such a war more likely.
§ 06What is not established
- Claude’s role in any strike. Amodei said he did not know it, and the interview does not settle it.
- Strikes carried out. The 5,000 figure describes capacity, from one unnamed official.
- Who the official is. Neither WIRED nor Bloomberg names the person.
- Why the clip is moving now. This part is our read. Anthropic is in the middle of its IPO roadshow, and older interviews with its CEO are circulating again. The same account posted a second Amodei clip on November 10.
Our conflict, declared: CellCog’s AI employees run on Claude Opus 5.5, which is Anthropic’s model, so we are an Anthropic customer. We have no part in its government work. This page takes no side on the policy question. It records what was said, when, and where each number comes from.
§ 07Sources
- Bloomberg Originals, Inside the Mind of Anthropic CEO Dario Amodei, The Circuit, Extended Interview, June 17, 2026. Captions read November 10, 2026
- Katrina Manson, WIRED, Meet the Gods of AI Warfare, March 23, 2026
- The X post, 04:57 UTC, November 9, 2026, read by id on November 10, 2026
Q1Did Dario Amodei say Claude helps kill more people?
No. That line is the interviewer’s question. Emily Chang said it and asked if he was comfortable with it. He answered that a stronger US military deters wars.
Q2What did Amodei say in full?
He said he supports the United States being more effective militarily, called himself a patriot, and then said specific military operations are not Anthropic’s decision. He said a supplier can set a few high-level limits.
Q3Where is the 5,000 targets a day figure from?
A WIRED book excerpt by Katrina Manson, published March 23, 2026. An unnamed official told her the US went from being able to hit under a hundred targets a day to a thousand, then to 5,000 once language models joined the Maven platform.
Q4Does the figure mean 5,000 strikes a day?
No. The WIRED passage and the interview question both describe what the military is able to hit in a day. Neither is a count of strikes carried out.
Q5What limits does Anthropic set on military use of Claude?
In the interview Amodei named two: mass surveillance and fully autonomous weapons. He said every other use is a decision for military leaders.