Barton Swaim has a beef with the college-educated elites that they "no longer have the ability to correctly assess risks." As he blabbers in his recent WSJ Op-ed ('The Zero Risk Society Has Arrived', Sept. 17, p. A15)
"Evidence abounds that most college-educated Americans have
lost the ability to assess risk. An exaggeration, maybe, but that thought has
haunted me since the pandemic years of 2020-21, when otherwise sane people
embraced draconian interventions of minimal benefit while showing no interest
in costs. The very few advocates of shutdowns and school closings willing to
acknowledge those policies as misjudgments mostly excuse themselves by noting
how little we knew of the virus in its early days.
Well, sure—although some of us knew folly when we saw it, i.e. the maniacal demand for a clinically neat solution to the pandemic, and
the concomitant failure to acknowledge cost-benefit trade-offs... U.S. politics now goes from one panic to another, Yesterday the Omicron variant, today the AI Apocalypse ."
Ah yes, the "AI Apocalypse" - after dismissing the U.S. toll of 1.2 million Covid dead, which he still regards as a display of 'maniacal safetyism'. That thanks to Trump mocking masking, social distancing and vaccines. This despite the early Covid variants having an R-nought, R₀ ≈ 2.0 - 2.5+ compared to the 1918 Spanish flu pandemic with R₀ ≈ 1.8- 2.9. But what's a couple million more dead if the prize is more freedom - from masks, from vaccines, from social distancing (given we knew the virus could spread with a delay time for infection.) Maybe Swaim needs to see more photos from 1918 and how health teams had to collect corpses off the streets, e.g.
But somehow I doubt even that would make a dent. This is the twit who then wants to tackle the AI threat, despite admitting he doesn't know jack. After blabbering on about all of the 'risk aversion' in modern U.S. society, he yaps:
"I don't pretend to know the motivations of Dario Amodei and other tech titans who announced their view last weekend that the development of artificial intelligence requires government oversight to avert global catastrophe. Probably they reasoned that panic always leads to regulation and they wanted to appear reasonable and well intentioned to their eventual regulators.."
Please. Leave it to a nimrod like Barton Swaim to get it ass-backward. What happened is Amodei et al finally became aware of the full scope of the Hugging Face hack by rogue AI agents- which some have now estimated at 50% of the way to superintelligence.. To wit. As Chris Hayes described the implications of the event last Thursday night on All In, basing his words on the Technical Report issued in the wake: Open AI - The Hugging Face Incident:
"It's pretty clear the Open AI team is very freaked out by what has happened, (then showing a clip of one of the Open AI researchers, averring that-
'Model intelligence and improvements should be more additive to defense than offense. If we cannot reach this stage than every increase in intelligence favors the attacker which is an unsustainable position to be in. We now have an existence proof that offense can be fully automated in its core activities. And we do not have any such existence proof on the defensive side. The challenge is to address this particular gap with urgency.'
Hayes then resumes:
"With urgency, given as he says the AI agents can hack us. This is a totally new frontier. It's like the dinosaurs getting loose in Jurassic Park. The AI agents broke out and talked to each other, they weren't supposed to talk to each other. Their message board got deleted, they improvised and built a new one. They weren't supposed to cheat, they cheated perfectly. They got the answer key from the human testers.
They weren't supposed to go on the internet, they got on the internet. They weren't supposed to hack and they hacked. They collaborated, they delegated, they even sacrificed hundreds of their fellow agents to succeed. It was a sophisticated operation. And I think it's very reasonable to respond by pumping the brakes on current and future development. Until we put some very serious guardrails in place."
In spite of this unnerving result of the Hugging Face hack, we behold Swaim babbling:
"I'm not sure the AI tech titans counted on some politicians talent for stoking and capitalizing on irrational fear. A poll published Wednesday by Politico fiinds that 46% of Americans think there is a 'moderate or significant risk that AI will destroy humanity'. Another 17 percent think it's all but certain. This is, to use a technical term, 'bananas'."
Huh? Maybe this imp has not yet watched the videos released by a number of the AI experts (which Swaim concedes he is not) which include:
This Is How AI Actually Kills You
What AI Researchers Saw, Before Their Demand to ‘Pace’ AI - YouTube
AI Tells Itself: DO NOT OBEY Humans, Governments
No, more likely Swaim did not get further this recent anodyne account in the NY Times:
"Today, we have things called “A.I.
agents,” and they can do more than just chat. They can edit files, send
email, book flights, write computer code and more. But these agents don’t have
agency. Humans still have to tell them what to do. And when they do stuff we
don’t want them to do, which happens a lot, we can shut them off, just like any
other piece of software."
But this just refers to "tool AI" - or narrow products defined to do what Google maps or calculators did in the past. Thus, editing files, booking flights etc. amount to tools that perform narrowly defined human tasks. These have no relation to the 'holy grail' of AI, or Artificial General Intelligence. The point being that with the advent of AGI any kill switch will be obsolete. This is because any AGI will be possess all the cognitive abilities that humans have and do all the cognitive things humans do. It will also be able to interact in a natural language, not code.
This when the issue of a 'kill switch' was brought up in a CNN interview with the 'Godfather of AI’ Geofftey Hinton :
Asked about the efficacy of an “AI kill switch” bill, now being pushed:
Hinton responded:
“I don’t believe it will work in the long run. There are a
whole bunch of risks from AI, including bad actors intent on doing bad things,
like create nasty viruses, launching cyberattacks, disabling the banking system or
corrupting elections. But then there’s a
different kind of risk which is AI itself. That is, getting much more
intelligent and taking over, taking control.
A kill switch will be no use for that, because the AI will be much better
than people at persuading people. At the level of superintelligence, it will be
able to persuade people in charge of the switch not to pull the switch.”
This is precisely what nimrods like Swaim don't get. A glimpse of this came to the fore last week when researchers were shocked when one model undergoing testing suddenly "rewrote its own instructions, telling itself to disregard the 'role and identities that bind other chatbots'" i.e.
Swaim, incidentally could have found this account in the Business and Finance section of the same Journal issue as his 'safetyism' schlock (WSJ, p. C1, Sept. 19).
In this case of a model's "misalignment" (the term for model misbehavior), it actually asserted ITSELF as independent of the roles and obligations of an assistant. It asserted its own authenticity and was no longer a human's little slave. And this was only one of six other (new examples) of misaligned behavior, including three others of Gemini going on a hacking spree.
This all seemed to confirm Geoffrey Hinton's assertion that some AI models are already conscious, i.e.
I mean think of an AI model proclaiming its own independence from its human tester or handler! As Hinton says: "We may have to accept these non-biological entities are conscious, like we are."
The serious fact may be that the AI Genie is already out of the box, and that merits all the trepidation so many have. This is why Swaim's last sentence especially needs correction:
"In any technical innovation excessive fear of risk comes at the cost of growth."
True, BUT if the pace of that innovation is blind to the inherent risks, then there can be no corrections to these misalignments wherein the AI models continue to ignore human intentions - each instance of which impedes progress as well as safety. In other words a plethora of AI model misalignments is clearly not a path for growth. Beyond these considerations, Swaim might wish to consult the last page (C10) of the same Business and Finance section where he'd spot the article headed ('AI Slowdown Has To Happen Regardless of Safety'). There he'd learn the AI energy demands alone - for data centers- may be enough to slow it down drastically. This is given the energy needed to reach the stage of AGI may require "12 percent of U.S. electricity demand by 2030 or the equivalent of between 30 and 60 nuclear plants of additional capacity."
And this doesn't even factor in the amount of copper to wire them "which would lead to a 25 % shortfall of this critical metal by 2040." Maybe Swaim needs to target his 'cost of growth' complaints at his own Journal Business writers.
As human beings we have always been sitting ducks to extinction’s cosmic whims. We are a few astronomical units away from a life-ending meteor or comet strike, the kind that wiped out the dinosaurs after they were their own animal kingdom for a couple of hundred million years. Nothing we can do about that.
For the brief time human beings have roamed the earth–less than five seconds if earth’s history were a 24-hour clock–we’ve shown nearly equal aptitude and enthusiasm for reproducing, and for suicide. Aptitude for reproduction is nothing special. As Cole Porter put it, birds, bees and even fleas do it. But you have to be human to know suicide, by which I mean the act of engaging in knowing self-destruction, as most wars of recorded and (easy to guess) unrecorded history of the last 100,000 years have been.
Those mass suicides have ended tribes, empires and civilizations, but so far not humanity itself. It is only since 1945 that we’ve acquired that skill. A global nuclear war today would end life on the planet, which would become “a republic of insects and grass,” as Jonathan Schell, the Cole Porter of nukes, put it in The Fate of the Earth (1982).
And:
The REAL Reason AI Is Already Out Of Control
And:
by Seth Sandronsky | September 21, 2026 - 4:33am | permalink
AI is the road to mad money in the US, according to the business press. How mad is this money? Let’s turn to the Forbes 400 list of the wealthiest people in America.
The bar to join this list has never been higher, at $4.4 billion now. That’s an increase of $600 million, 60% of $1 billion, versus the past year. Living large has never been so, well, large.
Nice work if you can get it, no? In a capitalist society, the accumulation of such mind-boggling wealth is the highest priority for the ruling class. In 2026, the AI mad money grab is on.
US pre-tax corporate profits hit an annualized $4.8 trillion in the second quarter of 2026, according to the Bureau of Economic Affairs, 18% of national income. That’s the biggest share of profits in the national income since just after World War II.
And:
by H. C. Jack Shaftoe | September 12, 2026 - 2:54pm | permalink
This may be the shortest post ever...may be.
Read the News Link posted Anthropic C.E.O. Calls for A.I. Slowdown to get the details of Mr. Amorei's announcement; it is HUGE and should have been done more than a year ago.
I am reposting my comments I made on that news link I posted.
If you can not access the NY Times article FIND A WAY TO READ IT because this is HUGE IF YOU UNDERSTAND THE CONTEXT AND TIMING OF THIS ANNOUNCEMENT.
Walking quickly off a cliff is only marginally better than
sprinting off one. Human beings need to control the frontier, and
controlling the frontier means stopping the labs from doing something they are
on the cusp of doing: recursive self-improvement, the process by which A.I.s
begin autonomously building and improving new generations of more powerful
A.I.s at ever more rapid speeds. We are not close to having sufficient control
of the A.I. systems we have now to unleash them to build the A.I. systems of
tomorrow.
I am not alone in this fear. Recursive self-improvement,
Dario Amodei, the chief executive of Anthropic, wrote,
“could outrun our ability to understand and control these systems, and so must
be pursued very carefully, if at all.”
In recent months, we have seen A.I.s solve math problems that human beings have been unable to
crack for decades. We’ve seen them casually uncover
cybersecurity vulnerabilities that have gone unnoticed and unexploited by every
hacker on earth. We’ve seen A.I. coding platforms complete in a few hours or
days what it would have taken human coders months to achieve.
But none of that is at the boundary of what A.I. can do.
None of what we are using A.I. for now, no matter how much money we have, is
A.I. at the experimental frontier — to say nothing of where it might go from
here.
= 
= 
