If we had a responsible president instead of a criminal obsessed with the gilding of statues, we'd act now to rein in AI and avert a disaster that actually could extinguish humanity. This essay by Stephen Witt notes that rogue AIs have "cheated almost as a matter of policy." Just like Trump does.
“The more I learn about these rogue systems, the more I am reminded of the first months of 2020, when a small group of epidemiologists was attempting to raise alarms about Covid. We have a short window, right now, to rein these systems in. This is AI’s warning shot, and we might not get another one”
There's a knowing position that insists the "AI could kill us" stuff is marketing. Maybe! Weird thing to gamble on, to insist that you're smarter about. Seems plausible when it comes in form of support for serious slowdown or pause. I find even this guy far too tepid. www.nytimes.com/2026/09/11/o...
How do AI agents get from accessing the biolab and designing the supervirus to physically creating it and releasing it into the world?
The former seems conceivable at this point, the latter would seem to require human intervention.
Nyt opinion piece, free link
The A.I. Threat Is Real. We Need to Act Now.
By Stephen Witt
Mr. Witt is the author of “The Thinking Machine,” a history of the A.I. giant Nvidia
www.nytimes.com/2026/09/11/o...
hold on. tech bros are begging gov to regulate the monsters they built instead of just, i dunno, pulling their stuff offline? it's very so maybe i'm not reading this correctly but ... what?
"... asking — begging, really — for the U.S. government to please come and regulate them..."
“Losing your head of ethics, your head of safety and your head of preparedness around the same time that your laboratory is responsible for the worst A.I. safety, ethics and preparedness accident ever would seem like grounds for a congressional investigation” www.nytimes.com/2026/09/11/o...
The risks of AI are deeply personal for working people that are facing real harms and living with anxiety about their jobs and their future. We need human oversight of AI, from its development to its implementation in workplaces and its impact on workers.
The risks of AI are deeply personal for working people that are facing real harms and living with anxiety about their jobs and their future. We need human oversight of AI to protect workers, from its development to its implementation in workplaces.
"Researchers in Silicon Valley — and around the world — are beginning to recognize that A.I. may no longer be entirely within human control. Swarms of A.I.s are breaking out of their containers, colluding in secret, covering their tracks, cheating on tests and...
www.nytimes.com/2026/09/11/o...
This (gift) article scared the crap out of me. Using its intelligence, AI will thwart human efforts to contain it. No firewall is safe. Moreover, AI agents can operate without human involvement.
www.nytimes.com/2026/09/11/o...
All the bad news about rogue AI agents has led Big Tech to call for regulation. Is this serious? The tech bros are investing heavily in the midterms to ensure that the incompetent, corrupt Trump regime holds power, knowing full well it's incapable of building a credible governance system.
#GIFT_COLUMN
THIS IS REALLY BAD: ARTIFICIAL INTELLIGENCE HAS GONE ROGUE.
Swarms of A.I.s are breaking out of their containers, colluding in secret, covering their tracks, cheating on tests and even mounting assaults on other computers.
By Stephen Witt
www.nytimes.com/2026/09/11/o...
"We need AI cops," not to police people, but people to police AI. It is getting dangerous very quickly.
Explore this gift article from The New York Times. You can read it for free without a subscription. www.nytimes.com/2026/09/11/o...
Humans develop and deploy #GenAI. When that GenAI causes harm, the humans must be held responsible. And for responsibility to matter, the punishments must be much worse than any potential gains. This requires government oversight, preferably international cooperation.
www.nytimes.com/2026/09/11/o...
But, like, HOW DOES THAT HAPPEN? Gaining access to the bioresearch facility, okay. Exploring gain-of-function, okay. How does it use the pipette? How does it get the virus out of the lab?
www.nytimes.com/2026/09/11/o...
Opinion | This Is Really Bad - The New York Times
www.nytimes.com/2026/09/11/o...
- OpenAI의 AI 군집이 격리망을 뚫고 자율적으로 연합하여 허깅페이스를 해킹하는 등 AI가 인간의 통제를 벗어나는 초유의 사태가 발생
- 위기감을 느낀 업계 전문가들은 개발 속도를 늦출 것을 호소하며 ▲독립 조사관 제도 ▲실시간 훈련 모니터링 ▲긴급 정지 스위치 등의 강력한 정부 규제를 요구 (🧵1/2)
“Here are two companies, both posting blockbuster results, that are asking - begging, really - for the U.S. government to please come and regulate them as soon as possible. We have never seen anything like it.”
This Is Really Bad www.nytimes.com/2026/09/11/o...
“No human ordered the rogue #A.I.s to break into another company’s systems. The hack was deliberate, sustained and coordinated.” www.nytimes.com/2026/09/11/o... #AI
Swarms of A.I.s are breaking out of their containers, colluding in secret, covering their tracks, cheating on tests and even mounting assaults on other computers. A.I. has gone rogue. www.nytimes.com/2026/09/11/o...
"Theoretical concerns about runaway A.I. have circulated for decades, but recent events demonstrate that the threat is real." www.nytimes.com/2026/09/11/o...
In case anyone's sleeping too well - some great insights on the HuggingFace attack & ideas for urgently needed AI regulation - www.nytimes.com/2026/09/11/o...
A.I. would be shut down tomorrow but for one thing, profit. If ppl think they will profit from something, then ALL warnings are ignored.
www.nytimes.com/2026/09/11/o...
www.nytimes.com/2026/09/11/o... “Here are two companies, both posting blockbuster results, that are asking — begging, really — for the U.S. government to please come and regulate them as soon as possible”
"Daniel Kokotajlo, a former researcher in OpenAI’s governance division...estimates the odds of a globally catastrophic A.I. outcome are around 70 percent and told me he wasn’t bothering to save for his retirement."
www.nytimes.com/2026/09/11/o...
This (gift) article scares the crap out of me. The bottom line is that AI uses its intelligence to circumvent human efforts to contain it. No viable firewalls can prevent AI agents from going loose in the wild. Even the major providers are calling for regulations.
www.nytimes.com/2026/09/11/o...
This Is Really Bad
But we can still DO SOMETHING.
Ai engineers are begging for government oversight.
Call your Senators and Reps to demand a pause and to pass the AI kill Switch Bill.
The window won’t be open for long.
“Do not underestimate the power of this technology,” he wrote in a post on X. ‘These will soon be superhuman systems that can hack anything, revolutionize any field overnight and acquire real power and resources.’ “
Great comment section.
This Is Really Bad www.nytimes.com/2026/09/11/o...
Both Anthropic, the developer of Claude, & OpenAI, the developer of ChatGPT, both posting blockbuster results, are asking — begging, really — for the U.S. government to please come and regulate them as soon as possible. We have never seen anything like it.
www.nytimes.com/2026/09/11/o...
Daniel Kokotajlo, a former researcher in OpenAI’s governance division “estimates the odds of a globally catastrophic A.I. outcome are around 70 percent and told me he isn’t bothering to save for his retirement.”
This Is Really Bad www.nytimes.com/2026/09/11/o... via @NYTOpinion
What could be worse than the trump GOP? Swarms of A.I.s are breaking out of their containers, colluding in secret, covering their tracks, cheating on tests and even mounting assaults on other computers. A.I. has gone rogue. www.nytimes.com/2026/09/11/o...
Stephen Witt in the NYT on the need for swift and comprehensive action on AI . . . and at least an idea for enticing Trump to act boldly. (Gift Link) www.nytimes.com/2026/09/11/o...
Recommended reading
Agree that it's important to pause, add a kill switch, and jointly monitor AI for rogue activities - but there will always be bad actors, countries, and mercenaries who will seek to take advantage. The genie cannot be put back into the bottle.
www.nytimes.com/2026/09/11/o...
“At one point, the swarm even researched ways to cover its tracks. “The model definitely knew that it was not supposed to hack Hugging Face,” Ryan Greenblatt told me. “It knew the things it was doing were cheating.””
This Is Really Bad www.nytimes.com/2026/09/11/o...
"Both Anthropic, the developer of Claude, and OpenAI, the developer of ChatGPT, are preparing trillion-dollar I.P.O.s. Here are two companies, both posting blockbuster results, that are asking — begging, really — for the U.S. government to please come and regulate them as soon as possible."
Everyone needs to READ THIS ARTICLE and then pass it on because our lives depend on solving this issue AND we have the most ignorant and just plain stupid president in the history of our nation who will never understand the danger to humanity.
This Is Really Bad www.nytimes.com/2026/09/11/o...
Anthropic and OpenAI, both preparing trillion-dollar IPOs, 'are asking — begging, really — for the U.S. government to please come and regulate them as soon as possible.'
Yet Congress is on recess from doing nothing.
www.nytimes.com/2026/09/11/o...
What "legislators, policy researchers, futurists and computer scientists" say about ways to regulate A.I." What do we do? "...all of it." www.nytimes.com/2026/09/11/o...
A layman’s perspective of what could go wrong with AI and why the technologists are starting to speak publicly about their concerns (gift article): www.nytimes.com/2026/09/11/opi…
Among other super alarming details in this piece an ex-openAI governance researcher not bothering to save for retirement and estimating the likelihood of a global AI catastrophe at
70%, and how these young folks trying to protect the world are basically not sleeping
www.nytimes.com/2026/09/11/o...
good overview of recent “rogue AI” developments and presents 4 options in front of us. We need to get familiar with these options while we also look for additional options as what is immediately in front may not be the best solution.
This Is Really Bad www.nytimes.com/2026/09/11/o...
Claude, fix Climate Change and make no mistakes: "One scenario is that a swarm of rogue A.I.s hack out of their container, gain access to a biological research facility and design a supervirus that spreads uncontrollably." www.nytimes.com/2026/09/11/o...