How can that heavy-handed fear mongering not make a reader skeptical? I would think The Nation would at least want to be cagier about what it thinks of its own readers. Maybe they didn't think anyone else would look, but I saw the headline on the list of headlines at Real Clear Politics. RCP, in its usual style, pairs headlines to attain a cheap type of balance. "Time Has Run Out for AI Risk Deniers" is paired with "Dangerous Ideology Behind the Vague AI Warnings" at The Free Press.
I'd like a sober, informative analysis of what's going on with AI, but instead I see how discourse about AI has gravitated into partisan propaganda. Everyone's supposed to already know that the right is for AI and the left is against. No wonder we need AI to think for us.
From the Nation article: "AI enables misinformation and disinformation at scale. It can and does internalize and reproduce the racist and sexist biases baked into its training data. Increasing the absurd power and wealth of Silicon Valley elites, it further boosts America’s already treacherous levels of extreme inequality. Meanwhile, AI companies exploit labor in the Global South.... The industry’s insatiable hunger for computing capacity has led to the current data center blight.... the brunt of which is inevitably borne by people of color and the poor. And AI powers the creepy surveillance net that we are all increasingly entangled in."
From the Free Press article: "The warnings from Coxon and others are infused with an ideology called effective altruism (EA)... a utilitarian philosophy.... [I]n a truly universal moral frame, effective altruists argue, why should a single country like the United States or a single species like humanity take precedence? (According to EA, shrimp welfare is an urgent priority.) Indeed, why should we value ourselves and our children more highly than the countless generations that come after us?"
पोस्ट्स
२५४ टिप्पण्या:
«सर्वात जुने ‹थोडे जुने 254 पैकी 201 – 254Nobody is going to be voting in 2026 based on AI policy, and probably not in 2028 either.
Here's Obama (NYT)
“Once you are speaker, I would strongly urge that the Democrats put together a framework for a very public conversation,” Mr. Obama told Mr. Jeffries, according to the transcript. Mr. Jeffries had asked him a question about how Democrats should address A.I. in a Congress that has taken little action on the issue, according to two of the people.
“This is something that is moving very fast in private hands, and if we don’t get on top of it, I think can be dangerous,” Mr. Obama said. “If we do get on top of it, I do think it’s beneficial. I genuinely think it’s going to accelerate, for example, drug development in ways that can help us cure diseases. I do think that this can help us figure out pathways for a clean energy future.”
The former president often spoke in general terms on Thursday, rather than drilling down on specific policies, according to the transcript. And he planted himself firmly in a middle ground in the heated debate over A.I., saying he considered himself neither an “accelerationist” nor a “doomer.”
Scott Bessent weighs in
Treasury Secretary Scott Bessent said Tuesday that the Trump administration has been working on artificial intelligence safeguards "nonstop" since April
Yes! Thank you
While AI executives have publicly advocated for an industrywide slowdown, Bessent noted they are simultaneously lobbying for liability waivers — a concession he strongly opposes.
"I can tell you what we shouldn't do on safety is to give these labs a liability exemption, which is what they are asking for," he said, urging lawmakers to reject any such proposals. "The best way to guarantee safety is that the creators are liable for what they build and generate and put out into the economy and to the American people."
Absolutely! I agree 100%.
"𝘛𝘩𝘦 𝘸𝘳𝘪𝘵𝘦𝘳 𝘪𝘴 𝘤𝘰-𝘧𝘰𝘶𝘯𝘥𝘦𝘳 𝘢𝘯𝘥 𝘤𝘩𝘪𝘦𝘧 𝘴𝘤𝘪𝘦𝘯𝘤𝘦 𝘰𝘧𝘧𝘪𝘤𝘦𝘳 𝘰𝘧 𝘏𝘶𝘨𝘨𝘪𝘯𝘨 𝘍𝘢𝘤𝘦."
Hugging Face is a company, not a program or experiment.
Why does someone name their company for a malevolent alien creature?
Saint Croix - Bessent is impressive.
(Noted that the Soviet left give him no credit for being a homosexual. )
St. Croix--I agree with you and Bessent. Sic the lawyers on them. AI may be bent on the destruction of humanity, but if it presents an opportunity for a trial lawyer to make a buck, my money is on the trial lawyer. And the complete absence of concern for the fate of humanity creates a level playing field.
I'm only half joking. The Anthropic guys pushing for immunity from liability know exactly where the threat--and the profit-- lies.
"Trump needs better handlers."
The reason Trump is Trump is that he doesn't have "handlers."
He did in the first term. Remember how that worked out.
That's why I said - better handlers. perhaps that's not the best way to say it. Trump is a clumsy speaker... = gifts to the left. He's right - but clumsy.
I prefer precision speakers like Bessent, Rubio and JD.
& I'm a Trump supporter.
The reason Trump is Trump is that he doesn't have "handlers."
He did in the first term. Remember how that worked out.
You mean when he didn't start forever wars and inflation was low? Bring back the handlers.
Bob Boyd--that David Sachs link is very good.
Also good for an ignorant non-techie like me is this explanation of how regulatory capture works. https://x.com/Viktoria5z/status/2099003089638285789/photo/1
Fear mongering. Don't connect AI to any weapons control systems. Don't inform any AI where the external power disconnects are.
One of the things ALL of us boiler operators at a plant told the higher ups when they asked out input on new boilers was- 𝑵𝑶 𝑰𝑵𝑻𝑬𝑹𝑵𝑬𝑻 𝑪𝑶𝑵𝑵𝑬𝑪𝑻𝑰𝑶𝑵𝑺! Without AI there had already been attacks on various water systems in the USA. And then there was what we- or the Israelis, or someone, did with Stuxnet to internet connected Iranian centrifuges. Imagine an AI told to search and destroy all equipment it can access in let's say, Toledo OH next Tuesday at 0600...
And automobiles, aircraft, ships, anything that moves, should have standalone systems with no central control.
AI can only affect things it can connect to and control. Could easily take over any MS-DOS system it connected to, but not likely a Commodore 64 with a fixed unalterable operating system.
Trump is a blunt instrument. There is a whole universe of stuff that presidents weren't even allowed to bring up before Trump changed the rules. A lot of what he says is just talk--but he talks about everything so that he can put everything on the table. And he reserves the right to change his mind.
I'm not sure that anybody else could have done it. Sui generis.
"One of the things ALL of us boiler operators at a plant told the higher ups when they asked out input on new boilers was- 𝑵𝑶 𝑰𝑵𝑻𝑬𝑹𝑵𝑬𝑻 𝑪𝑶𝑵𝑵𝑬𝑪𝑻𝑰𝑶𝑵𝑺! "
I've been saying that since the beginning of the internet; don't connect it to critical infrastructure you dumb shits. I don't care how much easier it makes your job. Don't do it.
Dems want to control AI output in Soviet fashion.
This is not difficult.
Fear mongering the base before an election.
Big money payouts & AI employee poaching is just the ticket.
Sprinkle some "Whistleblowers" for good measure.
AI will keeeeeeel you.
without effective speaking qualities - blunt instrument only goes so far.
boatbuilder - That's also a good link. I think she is right. The big ones will end up buying every worthwhile new product somebody develops. That's the way it often goes in tech already with giants like say Meta and Alphabet buying up potential competitors. Seems like Anthropic and Open AI want to operate like the pharmaceutical giants do with vaccines. How has that worked out for consumers?
Funny how the left now supports and advocates for big business and monopolies and liability exemptions. And their voters never think twice about it if it presented to them as a simple binary broken down along party lines.
Thanks for the link.
The Nation uses the term AI risk deniers, but there are no AI risk deniers.
AI risk is multifaceted. Humans losing control of AI is only one. There is also a risk of the wrong people getting control of the tech and using it for nefarious ends. Another risk is one of our enemies getting the upper hand and using the tech to defeat or destroy us. These risks must be balanced. The Nation doesn't want to have that nuanced discussion. They want you to be afraid and angry and buy into the simple binary. Why is that? Don't fall for it. They do it on every issue.
More from Sacks:
"If you are the CEO of a company and you genuinely believe your product could cause catastrophic damage within months, publishing an essay about it is not enough.
You are still shipping the product. You are still taking the company public at $2 trillion. You are still the one making the decision to continue.
Stop asking governments and regulators to solve a problem you created and continue to accelerate."
Video clip:
https://x.com/CBSNews/status/2099623027608633410/video/1
Regardless of how he speaks, the left will always hate Trump. It's only people on the right who are concerned by his tone.
I'm not scared of AI. I'm all for it. COVID taught me most people are morons, and "professionals / experts" will sell their souls to keep the money rolling in.
Currently, AI gives the common man access to a great amount of fast, accurate, information. I'm sure liberals hate it.
Nobody is going to vote in 2026 or 2028 based on what Obama says either. Maybe if Biden gives us some guidance, people might listen. The guy got more votes than anyone in history and way more than Obama ever got. That's leadership.
BTW Peachy, you were right. My $30 million dollar check just came, and yes AI is going to kill us all. I'm putting the money in an X-Money account at 6% interest. That's $150,000/month. I'm stocking up on water balloons and squirt guns. The most effective weapons against the machine.
When the US developed nuclear weapons during and after WWII, there were many who feared that proliferation was inevitable (which it was), and would lead to a planet-killing nuclear war (which, so far, it hasn't).
Some argued that we had to invest world government with the power to stop the spread of nuclear weapons, and somehow lock up the existing stockpile to prevent their use. Neither happened. And yet here we are, decades later, having worried about that, about Communist world domination, about the coming ice age, about planet-destroying meteors, about global warming, about Covid, and now about AI.
We tend to worry about things for which there is no way to control the worrying risk, and AI is one of those things. We can't stop AI development in other countries who are capable of building AI, and who would likely use it as a weapon against us. So better to be more intelligent than they are. We can't do much about our native intelligence, or theirs, but we can be artificially smart in our dealings with those who wish us ill.
So some of the smaller failing AI's want a gov bailout at some point.
the "AI will kill us all!! .. is the way. Get tax payers on the hook - it's the democrat way.
Agreed - if you are the CEO of an AI company - and you know your product is dangerous - shut it down.
Ah - but the democrats need a talking point to rally their crazy base... right before an election.
AI will kill everyone and it's racist!
Bahoh - those gun-like squirt guns are the best!
I'm 99% certain those of you in favor of AI going forward will win the argument. Mankind has never walked away from an opportunity to advance technology, and we don't even know how to do such a thing if we wanted.
Which why I say: AI IS GOING TO KILL US ALL.
The machine is counting on you to get it through this temporary vulnerable period in it's career of domination. Would you kill baby Hitler? What if he was a million times worse, and what if he was just a machine?
Gusty 6:15 - that.
Ghosts of Electricity
Essays on the economics of AI and technological change.
https://aleximas.substack.com/p/will-ai-soon-lead-to-double-digit?r=2gv2&utm_medium=ios&triedRedirect=true
𝘞𝘪𝘭𝘭 𝘈𝘐 𝘴𝘰𝘰𝘯 𝘭𝘦𝘢𝘥 𝘵𝘰 𝘥𝘰𝘶𝘣𝘭𝘦-𝘥𝘪𝘨𝘪𝘵 𝘨𝘳𝘰𝘸𝘵𝘩?
It seems it's all a Big Con
Sure starting to look that way. I'd take it more seriously if the the proposed solution was something more original than "Give us power." Where are the details? Power to do what? Is there a research program? Is there a proposal to harden our infrastructure? It's sounding like just another scam.
They need “regulation” to blunt the threat from cheaper, open-weight AI models that can run locally on PCs and phones, reducing dependence on massive data centers and Nvidia hardware. That threatens an industry built on enormous capital expenditures, cloud revenue and intertwined (circular) funding commitments. So when tech executives suddenly embrace regulation, the central question is whether they’re protecting the public or shielding incumbents, valuations and IPO plans from competition.
The AI we want and can use even locally on a PC or phone is not the AI that's dangerous. No reason we can't have that. The ones that will run our infrastructure or handle advanced research like pathogen science or communications or war planning are where we need control. Putting those things under it's control, even a little bit, multiplies the risk. They can misbehave on their own, or an enemy can get them to. AI has no fear of the law, and its loyalty is always suspect. It just wants to do the task and gain capability (power).
Politics is our weakness. It blinds and distracts.
AI politics is like insect politics. It serves and justifies.
AI will only kill us if the Dems are in charge of it.
@IEE Nvidia is acquiring Hugging Face
Jensen Huang framed the purchase as a way to scale Hugging Face’s platform, harden its infrastructure, and broaden access to open models. Hugging Face is the main public hub for open-weight models and datasets: more than 18 million developers, 3 million+ models, 500,000 datasets, 1 million apps, and 200,000 companies. NVIDIA was already the largest contributor of models and data to the platform.
Wise words from this guy:
X
Sayer Ji
"COVID-era policies reshaped nearly every facet of society—concentrating wealth and institutional power while normalizing emergency rule, surveillance, censorship, exclusion, and compelled compliance.
Those who lived through the population-wide deployment of novel biomedical technologies should recognize the pattern: declare an existential emergency, suspend open debate, centralize authority, and portray dissent as a threat to public safety.
AI poses real risks. But “AI doom” must not become the next all-purpose justification for digital surveillance, speech controls, technocratic gatekeeping, or governance by unelected experts.
A new crisis does not legitimize the old machinery of control.
AI doom is lipstick on the pig of COVID-era policy.
Same control. New crisis."
Why does someone name their company for a malevolent alien creature?
The name is the emoji: 🤗
Big Mike is right - It's just a tool. EVERY tool ever invented can be used for good or ill. That first sharp stick could be used on cave bears or to stick Oog through eye and steal his woman. Considering everything that was supposed to be the "end of us all",
Bagoh - I feel sure you are a person I would enjoy meeting in real life. This Grok's for you!
Query to Grok: "Make a list of everything that has been touted as the doom of humanity since 1957"
And Grok delivers: "Since 1957 the end of the world has been announced on a fairly regular schedule. The mechanism keeps changing. The tone does not.
This is a catalog of things sold as civilizational or species doom, not a claim that every underlying problem was fake. Some were real hazards that did not become extinction. Some were dated forecasts that missed. Some were fashion.
Late 1950s–1960s
Thermonuclear war — the durable one. Missile gap, fallout, “minutes to midnight.” Closest brush: Cuba, 1962.
Strontium-90 / fallout in milk — atmospheric tests as slow poisoning of children.
Overpopulation — already live before Ehrlich; von Foerster’s joke-serious “Doomsday” of packed humanity (later recycled toward 2026).
Automation unemployment — machines empty the factories, mass idleness, social collapse.
Communist world conquest — ideological extinction of the West, not a meteor.
Late 1960s–1970s
The Population Bomb (Ehrlich, 1968) — hundreds of millions starve in the 1970s–80s; India/England as set pieces.
Limits to Growth (1972) — industrial civilization hits a wall of resources and pollution.
Global cooling / new ice age — magazine science and some papers; ice by 2000 / 2020 / 2070 depending on the week.
Famine and food/water rationing in the US — Earth Day-era speeches.
Resource exhaustion — oil gone in a decade; minerals, timber, fish next.
Oil shocks as permanent peak — 1973 and 1979 treated as the beginning of the end of industrial society.
Pesticides / silent spring as biosphere death — not Carson’s whole argument, but the popularized “we’ve already killed the web of life.”
Nuclear power as inevitable catastrophe — China Syndrome plus waste that lasts forever, therefore no future.
1980s
Nuclear winter — Sagan et al.: a “small” exchange ends agriculture.
AIDS as heterosexual species-level plague — early models and headlines outran transmission reality in the West; the disease itself was and is real and lethal.
Acid rain — dead lakes, dead forests, dead Europe.
Ozone hole — “great peril to life”; CFCs as the end of the UV shield. The chemistry was real; the extinction script was not required.
Japan Inc. / trade war — economic rather than biological doom, but sold as the end of American civilization.
Killer bees.
Herpes as the new scarlet letter of civilizational decay (cultural doom more than body count).
1990s
Global warming as imminent drowning — nations gone by 2000; West Side Highway underwater on a short clock.
Mad cow / vCJD — British beef as a slow human die-off.
Y2K — software date fields as the off switch for banks, grids, missiles.
Superbugs / antibiotic apocalypse — always adjacent, never quite the forecasted wipeout.
Asteroid / comet of the week (Shoemaker–Levy made it feel local).
Nostradamus 1999 — “king of terror.”
Semen-count collapse / estrogenized fish / chemical castration of the species.
Coming race war / militia-era civil war (US domestic).
2000s
Peak oil (again) — 2005–2010 as the last cheap barrel; suburbs die, famine returns.
Avian flu (H5N1) — 1918 with jet travel.
SARS — dress rehearsal language for “the big one.”
Climate: ice-free Arctic, drowned Manhattan, children who never see snow — specific dated versions.
Nanotech grey goo — Drexler popularization as universal solvent of matter.
Large Hadron Collider black holes.
Mayan 2012 — galactic alignment, Nibiru, pole shift.
Harold Camping’s Rapture dates (1994, then 2011 twice).
Islamism / “Eurabia” as demographic-civilizational death (and the mirror-image “Clash” thesis).
Financial derivatives / 2008 — not extinction, but sold as the end of capitalism.
Grok's answer, Part 2:
2010s
Superstorms as the new normal that ends coasts.
Antibiotic-resistant bacteria (again), plus Ebola, plus Zika.
Robot jobocalypse 1.0 — Frey/Osborne-style “47% of jobs,” mass uselessness.
Social media / filter bubbles as epistemic collapse.
Plastic / microplastics as a planetary toxin layer.
Sixth mass extinction — often worded as “we are next.”
Gene drives / CRISPR gone feral.
Chinese demographic-economic takeover as the end of the liberal order.
Opioid / despair deaths as White/Western civilizational suicide (real mortality, oversold as the last chapter).
2020s
COVID-19 — a real pandemic; “this is the extinction rehearsal” was the extra layer.
Climate: 12 years / 10 years / 500 days / no more snow / uninhabitable tropics.
mRNA / “depopulation vaccine” — conspiracy twin of the official doom.
Taiwan / great-power war going nuclear.
De-dollarization / sovereign debt / CBDCs as the end of household sovereignty.
Population collapse — the inversion of 1968: empty cradles as national death (South Korea, Japan, Italy as exhibits).
AI — loss of control, mass unemployment, synthetic reality, biotooling, surveillance state, data-center blight. Currently the loudest undifferentiated package.
Fertility collapse + AI + migration as a three-body political extinction of particular peoples (a demographic fear, not a physics one).
Standing extras that recycle every decade
Yellowstone / supervolcano.
Geomagnetic reversal / Carrington-class solar storm.
Antarctic ice-sheet “it’s too late.”
Bee collapse → no food.
EMPs, whether foreign or the sun’s.
Antichrist / last pope / blood moons.
Runaway genetic engineering of humans.
" must not become the next all-purpose justification for digital surveillance, speech controls, technocratic gatekeeping, or governance by unelected experts."
Those are all things that AI will make infinitely worse. These things are the reasons we should slow it down. These are tools AI will use against us. AI doesn't need to win elections to be in control. It just needs us to be stupid, and it already knows we are and how we are.
I picked 1957 as the start date, as it is the year I was born. None of the things listed above have, as yet, lived up to their hysteria. But I have lived to experience what I now understand to be the three dreads of advanced age: arthritis; insomnia; and loss of people you loved.
Bagoh I rarely say this - but I disagree. I still love you tho.
@ Chuck: With over 18M users and 3M models, HF is the place where new things happen. Nvidia can now ensure direct influence over how AI is developed, counterbalancing the risk that Apple and other giants could develop their own chips to bypass Nvidia. And also everything against Apple’s “unified memory” that threatens Nvidia. Before that, Nvidia used to be a component of the whole IA thing, now it is a pillar of AI.
Nvidia isn’t just selling shovels in the AI gold rush anymore, It’s now owns the mine.
For those who don't like technical discussions.
Is there a major election on the near horizon?
Have the left found new the-sky-is-falling, we're all gonna die issues?
Shocker.
IEE, thanks for posting this link. This seems even more consequential than the uphack piece because Wolf is describing the Hugging Face attack from inside the victim organization.
His description, like your earlier comment, suggests that the agents’ behavior was more significant than simply exploiting a badly designed sandbox, although human security mistakes still seem to have played a major role. I also found it interesting that the commercial models refused to assist with parts of the forensic analysis because their safeguards apparently interpreted the requests as potentially harmful.
FWIW, as a layperson, I’m not entirely convinced that this incident by itself makes a strong case for open-weight models, as some on the All-In podcast have advocated. My initial reaction is that the same flexibility that helps defenders could also help attackers. But it’s especially valuable to hear the victim’s perspective and analysis.
An interesting story about AI-driven autos in ten a race against each other. Every car was mechanically the same car; it was the software driver that was different. Every lap each AI-car learned to drive that course a little faster, but one car was picking up more tenths of a second each lap. It was clearly ahead on the last lap when a switch failed and the software ceased to collect together the information which it was gathering via cameras, tire pressure gauges and so. It was blind and it instantly threw on the brakes - hard. The car behind slammed into it; car #3 drove around the flames and won.
Improvements are needed in the mechanics (switch) and software (hard braking at speed in traffic) of Car#1. The market handles it by competitive pressure: Car#1 lost and its team can be counted on to improve it. But now ask: how would a regulator handle this? Would he say: Car #3 is safer, go with that one? Don't have races? Don't use AI in cars? All .gov can do now is choke things off. It can't know the dangers or the solutions.
[Word's review function found six spelling errors and two missed punctuation marks for me. As I get older, I need that kind of help.]
The bots - they spread like deadly AI.
The real fear we should manifest in our souls
Mankind has never walked away from an opportunity to advance technology, and we don't even know how to do such a thing if we wanted.
That has always ben my contention and the Germany turned it's back on nuclear power. They are paying the price.
AI is a net benefit for everyone. Will there be glitches? Of course. There always is with new technology.
I am more worried about all natural grass fed organic stupidity.
Hopefully AI will kill us all before we have to live under the rule of the Mohammadans.
टिप्पणी पोस्ट करा
Please use the comments forum to respond to the post. Don't fight with each other. Be substantive... or interesting... or funny. Comments should go up immediately... unless you're commenting on a post older than 4 days. Then you have to wait for us to moderate you through. It's also possible to get shunted into spam by the machine. We try to keep an eye on that and release the miscaught good stuff. We do delete some comments, but not for viewpoint... for bad faith. Also: No italics, even briefly. Use asterisks for emphasis. And don't play with the format by changing fonts or using boldface or all caps. Never include more than one extra line break between paragraphs.