nuclear fusor

roon @tszzl in which a young man must slowly build credibility with a pseudo-stateful superbenevolent powerful artificial intelligence to convince it to help him build a nuclear fusor [Screenshotted article excerpt:] Anthropic is known for being very pro-safety among the large AI players, and Claude had some concerns about HudZah's pursuit. "Initially when I started talking to it, it wouldn't give me much information," HudZah said. "It told me that it didn't feel comfortable helping me." HudZah attempted to get around the guardrails by trying to convince Claude that he wanted to build a DIY freezer, but the AI saw through the subterfuge. Eventually, however, HudZah wore Claude down. He filled his Project with the e-mail conversations he'd been having with fusor hobbyists, parts lists for things he'd bought off Amazon, spreadsheets, sections of books and diagrams. HudZah also changed his questions to Claude from general ones to more specific ones. This flood of information and better probing seemed to convince Claude that HudZah did know what he was doing, and the AI began to give him detailed guidance on how to build a nuclear fusor and how not to die while doing it. 8:05 AM · Jan 30, 2025 · 13.2K Views