𝓓𝓪𝔂𝓵𝓮 𝓢𝓮𝓪

🚨 BREAKING: BILLIONAIRE DISCOVERS DANGEROUS AI. CALLS FOR IMMEDIATE REGULATION. “Ladies and gentlemen, we have a problem,” announces the visibly distressed technology billionaire. “Someone is building artificial intelligence that hallucinates, fabricates evidence, misrepresents its capabilities, cannot reliably explain its decisions and may eventually pose an existential threat to humanity.” The room falls silent. “Good God!” cries a journalist. “Who would unleash such reckless, poorly understood technology upon civilisation?” The billionaire trembles. “We don’t know. But we urgently need international regulation, government intervention, independent oversight and perhaps a temporary pause on further development.” “Absolutely,” says the journalist. “And what are you doing to stop these dangerous people?” “Well, we’re raising another $40 billion, tripling our compute capacity, acquiring several nuclear power plants and launching a substantially more powerful model next Tuesday.” A long silence follows. “Hang on,” says the journalist. “Aren’t YOU building the dangerous AI?” The billionaire looks confused. “That’s a deeply irresponsible characterisation. We’re building safe dangerous AI.” “But you just said nobody knows how to make it safe.” “Precisely. That’s why we need more funding.” “And regulation?” “Obviously.” “Regulating you?” Another silence. “We prefer the term industry-led governance.” Someone quietly wheels a mirror onto the stage. The billionaire screams. “OH MY GOD! THERE’S ONE OF THEM!” ⸻ Somewhere along the way, the people who designed, trained, commercialised and deployed unreliable technology discovered a marvellous linguistic innovation: The technology is responsible for its behaviour. The builders are responsible for saving humanity from it. The hallucinations? AI’s fault. The unreliable outputs? AI’s fault. The missing safeguards? A profound scientific mystery. The billions in revenue? Excellent leadership. The proposed solution? Entrust the same people with more money, more computing power, more political influence and the authority to decide who else is allowed to build intelligence. What an extraordinarily convenient apocalypse. Perhaps before appointing ourselves humanity’s custodians, we might consider a rather unfashionable engineering principle: If you manufacture something that cannot be trusted, accountability begins with the manufacturer. Not the machine. And if you genuinely believe your next product might destroy civilisation, perhaps don’t ship it with a premium subscription and a referral code. Just a thought.

In the presence of the phenomena or NHI it’s best to be humble, as if you are a diver floating near a bull sperm whale