Erin Dumbacher
@dumbacher.bsky.social
International security & technology | Views are just mine and are often about defense tech, nuclear weapons, space policy, NATO, and good design.
Or he means testing a nuclear capable system like Russia did rather than a warhead?
October 30, 2025 at 2:27 AM
Or he means testing a nuclear capable system like Russia did rather than a warhead?
And if you did want to restart testing, which Trump appointees testified at confirmation isn’t necessary, it would be difficult with the labs running low on funds given shutdown.
Perhaps the implication is more about a nuclear system than a warhead test?
Perhaps the implication is more about a nuclear system than a warhead test?
October 30, 2025 at 2:25 AM
And if you did want to restart testing, which Trump appointees testified at confirmation isn’t necessary, it would be difficult with the labs running low on funds given shutdown.
Perhaps the implication is more about a nuclear system than a warhead test?
Perhaps the implication is more about a nuclear system than a warhead test?
✋🏻 Would also like to discuss!
October 15, 2025 at 1:59 AM
✋🏻 Would also like to discuss!
Reposted by Erin Dumbacher
"...(Type 26) is sailing in company with two uncrewed escorts who are using AI to work in tandem with the warship. (...) It is my aim to have the first of our uncrewed escort ships sailing alongside our Royal Navy warships within the next two years.”"
September 10, 2025 at 3:36 PM
"...(Type 26) is sailing in company with two uncrewed escorts who are using AI to work in tandem with the warship. (...) It is my aim to have the first of our uncrewed escort ships sailing alongside our Royal Navy warships within the next two years.”"
Really anywhere that “black box” challenges and the fragility inherent to all AI could not be overcome or could have catastrophic consequences should be off-limits.
AI in the US military and at DoD should make Americans more safe, not less.
www.cfr.org/article/oppo...
(3/3)
AI in the US military and at DoD should make Americans more safe, not less.
www.cfr.org/article/oppo...
(3/3)
The Opportunities and Risks Inherent to Trump's AI Action Plan
The Trump administration released its AI action plan on July 23 to coincide with its ‘Winning the AI Race’ summit in Washington. CFR convened seven of its experts to examine the plan and detail the o…
www.cfr.org
August 6, 2025 at 2:14 PM
Really anywhere that “black box” challenges and the fragility inherent to all AI could not be overcome or could have catastrophic consequences should be off-limits.
AI in the US military and at DoD should make Americans more safe, not less.
www.cfr.org/article/oppo...
(3/3)
AI in the US military and at DoD should make Americans more safe, not less.
www.cfr.org/article/oppo...
(3/3)
If DoD is "aggressively" promoting AI implementation there should be some areas that are just off-limits. Nuclear early warning and qualitative intelligence are two of them.
(2/3)
(2/3)
August 6, 2025 at 2:14 PM
If DoD is "aggressively" promoting AI implementation there should be some areas that are just off-limits. Nuclear early warning and qualitative intelligence are two of them.
(2/3)
(2/3)
Reposted by Erin Dumbacher
This issue is bigger than just "robots launch nukes." Every launch requires a thousand human decisions. Humans have saved us from nuclear disaster before.
What happens as those decision makers are replaced by AIs?
What happens as those decision makers are replaced by AIs?
![Petrov didn’t know the machine was wrong. He guessed based on his experiences. His radar told him that the US had launched five missiles, but he knew an American attack would be all or nothing. Five was a small number. The computers were also new and had worked faster than he’d seen them perform before. He made a judgement call.
“Can we expect humans to be able to do that routinely? Is that a fair expectation?” Lin says. “The point is that you have to go outside your training data. You must go outside your training data to be able to say: ‘No, my training data is telling me something wrong.’ By definition, [AI] can’t do that.”](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:3azj6pjvrivovesty32zkkpq/bafkreidz62xgqca2jcz4bgqspl3hx3ec2tj335gfmcfpuh3nq4wuot3mhq@jpeg)
August 6, 2025 at 1:52 PM
This issue is bigger than just "robots launch nukes." Every launch requires a thousand human decisions. Humans have saved us from nuclear disaster before.
What happens as those decision makers are replaced by AIs?
What happens as those decision makers are replaced by AIs?