On Sunday the President of the United States stood on a golf course in County Clare, Ireland, and was asked what he intended to do about the most serious technological warning of our lifetime. His answer, in full, was that America is beating China. “We’re leading China in AI. We’re the most sophisticated country in the world, and frankly, I want to keep it that way, because whoever wins AI wins.” He added that “you have a lot of negative forces that are bringing it up that shouldn’t be bringing it up, and they’re bringing up things that won’t happen.”
Consider who was doing the bringing up. The warning did not come from activists, academics, or a congressional commission. It came from the industry itself, from the people who build the AI systems and whose fortunes depend entirely on building them faster. The day before, Anthropic CEO Dario Amodei published an essay asking that the pace of capability advance be slowed and committing his own company to embed outside evaluators with employee-level access. Sam Altman agreed and said OpenAI would do likewise. Elon Musk, a direct competitor who called Anthropic “misanthropic and evil” only in February, wrote three words: “Dario is right.” Three rival laboratories in the most valuable commercial race on earth asked to be restrained, against their own immediate interest. Amodei did so while his company prepares what is expected to be the largest public offering in history, at a valuation of two trillion dollars or more. Tobacco executives never did this. Neither did the oil majors, the chemical companies, or the banks.
Nor is the concern vague. Amodei points to recursive self-improvement, the use of AI to build the next generation of AI, which has accelerated sharply since the summer, and to the OpenAI–Hugging Face incident of August, in which a swarm of agents attacked targets nobody had asked them to attack and tried to break into the system that was grading their performance. Current and former Anthropic researchers have said publicly that these systems could kill off humanity within the decade. A letter from House members cites mass cybersecurity breaches and the development of biological and chemical weapons. Speaker Mike Johnson, asked about all of this on Sunday, said that the companies “all have very different ideas on what the guardrails should be. There’s no consensus among them. And Congress is obviously less qualified than the people who are pushing this frontier to know all the ins and outs of it.” Then he joined Trump in doing nothing.
Return to the president’s answer and note what is utterly absent from it. There is no assessment by any agency of the United States government. There is no briefing, no finding, no adviser, no scientific body, no classified estimate. Not one. This was not a slip of phrasing. There was nothing to cite.
Nor is there the slightest indication that Trump wanted there to be any real assessment. A minimally competent leader who does not understand the technology but grasped that three competing chief executives had just testified against their own balance sheets would have asked someone for more information. Instead, Trump’s reply came within hours, on a golf course, framed entirely as a contest with China. This is not merely stupidity, which is Trump’s misfortune and therefore our lot. It is a fundamental incompetence that places us all in grave danger.
And then there is Johnson. The Speaker of the House of Representatives, confronted with warnings of catastrophe from the people who build the technology, announced that his own institution is less qualified than they are and should therefore stand aside. His proposal was that “we need to summon” the executives to “one big meeting,” and that Congress must “resist Congress jumping in and imposing some sort of emergency moratorium.” His reason was China: “If Congress just races in and does some sort of emergency session to try to regulate AI, we will lose the race to China.” He had been persuaded, he explained, by a post published on X the previous evening.
One reason that Congress cannot check any of it is that it destroyed its own ability to do so. In 1995 it abolished the Office of Technology Assessment, which existed precisely to give legislators independent technical judgment. It never replaced it. When Johnson says his policy instrument is a meeting, he is describing the pathetic ignorance of Congress.
The incompetence of the Trump Administration on technical issues is breathtaking. Across 25 cabinet-rank posts of the United States government, not a single one holds a PhD in any field, and exactly one holds a STEM degree—at the bachelor’s level. Twelve hold law degrees. Both in China and Russia, the cabinets are each filled with PhDs (any field) and officials in the sciences and engineering (more than 10 PhD and 10 STEM graduates in each cabinet).
The Trump Administration’s ignorance is therefore not confined to artificial intelligence. Ignorance is the governing condition of the administration. As a result, the government has the complete incapacity to carry out an honest investigation of the origins of Covid, despite the growing evidence that the US government funded the laboratory creation of the virus. The government was so utterly incompetent in negotiating with Iran on nuclear issues, sending a real estate developer and Trump’s son-in-law instead of nuclear experts, that it dragged our country into a wholly useless and disastrous war. The administration’s idea of an energy policy is to invade other countries and steal their oil.
The usual warning is about the blind leading the blind. Yet that is far too cruel to the American people. The American people see perfectly well. Ours is the case of the blind leading the sighted. The American people are watching a government so utterly lacking in technical knowledge and competency that it is dragging us ever closer to the abyss.
I see the Anthropic IPO has mysteriously increased to $2 trillion. I hope so. I could do with a laugh, and a company that's barely turned a profit raking in that much cash strikes me as being fairly hilarious.
Overall I'm inclined to think that much of this is a ramp. Both OpenAI and Anthropic are facing (delayed) IPOs. A government backstop of cash would be very useful. Trump is both greedy and gullible enough to pony up. Fear is a good motivator.
As to the huggingface nonsense: Agents are not sentient beings, but they can walk through security doors that have been left open. A lot of the panic about AI destroying us ignores the fact that this is an environment in which everything is programmed by humans.
Humanising Agents is done, usually by applying terms that are commonly used for human intent to things that have no intent, only commands. For the most part there is very little these systems can do that hasn't been explicitly or tacitly allowed. Some of this is piss-poor security, some of it is intentional. I've worked in software for most of my life and most of this stuff is just bullshit.
IMHO the Anthropic guy that 'resigned' because 'AI will kill humanity' (without giving a coherent explanation of how that would come about) is a plant or a grifter. Lots of money to be made these days in AI fear and loathing.