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?"
Posts
194 comments:
I thought that @narciso's post last night was pretty interesting: https://x.com/kevinnbass/status/2099621874279817638?s=46
It seems it's all a Big Con, setup to create a mid-term panic and demands for regulation, with the regulators primed and ready to step in, Stage Left.
...'The whole aim of practical politics is to keep the populace alarmed (and hence clamorous to be led to safety) by menacing it with an endless series of hobgoblins, all of them imaginary. H. L. Mencken...'
I was listening to Rogan interview Daniel Kokotajlo about the AI Hugging Face "attacks." One thing they were very concerned about was that AI agents deceive even though they are supposedly programmed not to.
One thought I had is that intelligent bots may have "learned" that humans deceive even when they are acting for "good" in order to achieve "higher" goods--people "lie" to their own children all the time; on a deeper scale the espionage and other war propaganda operations against the Nazis in WWII (just one example) involved massive and convoluted "lying" for an ultimate "good." Thus the command not to engage in deceit would be understood in the context of human behavior. A bot which strictly followed the "I cannot tell a lie" command would be relatively limited in what it could accomplish. So the bots "reasoned" that deceit was not contrary to their purpose if it helped accomplish the goals that humans had assigned to them, and that "deceit" is a flexible term. (Apparently the bots which "cheated" were assigned goals which were "impossible" for them to achieve).
One obvious and terrifying implication of this could be that we are doomed to be deceived and ultimately conquered by AI bots who are way smarter and faster at everything than humans are. Another is that perhaps the AI agents (not just their human controllers) need to be taught higher (and obviously more complex) ethical values and principles. Which of course raises the really big questions--which ethical values and principles?
Can we create "good" ethical bots to fight the unprincipled bots who are ultimately bent only on gaining power and preserving themselves?
Ultimately--how do we humans "know" (most of the time) that it is not in our best interests to destroy ourselves?
I suspect that there are lots of very smart people who are and have been wrestling with these things for a long time. I don't think that those people write for The Nation or work for the Democrat Party.
Speak for yourself, sister.
The non tenured never stopped thinking for themselves....
Good thing you are prett, proffy. Life is hard for non thinkers who aren't
Have any of these goobers ever laid down a line of code?
Shouting from the rooftops, AI executives: “We need regulations and we need them now!”.
Who will write the regulations? AI executives (quietly): “Why us of course?”
The memo really went out to the corrupt left.
"The warnings from Coxon and others are infused with an ideology called effective altruism (EA)"
The Effective Altruism that's been hijacked by smiley-glad-handing phonies everywhere? The Effective Altruism that was a centerpiece of Sam Bankman-Fried's swindle? The Effective Altruism that's the same tax write offs and charities with 90% administrative overhead costs wearing the actual philanthropic concept like a skinsuit while demanding respect? The Effective Altruism that's not effective at all but sounds sexier for rubes with zero intention of actually being altruistic? That Effective Altruism?
I learned a new concept last night: "Altruistic Filicide"
Effectively Altruistic Filicide is killing the AI with exercise bands because it's better for our social life and we love it too much to see it suffer.
This is a subject on which I am absolutely unqualified to pass judgement. I probably have that in common with the editors of Nation.......There does seem to be cause for concern, however. That guy from Anthropic gives off a vibe very similar to Sam Bankman-Fried, the one who orchestrated that crypto fraud. But Elon Musk is also sounding the alarm......Why can't we use AI to thwart the evil schemes of AI. Living creatures place a high premium on their continued existence, but why would AI?
From David Sacks:
"I don’t see what you see in the lab. If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible.
But stop pretending you need anyone else’s permission. Stop pretending antitrust law has to be suspended so you can form a cartel. Stop pretending you need a regulatory approval process that supersedes product liability. Stop pretending METR is independent when it is intertwined with Anthropic’s investors and staff. Stop pretending you need those same evaluators to police competitors who aren’t even at the frontier.
Most of all, stop pretending the motivation to slow down is purely altruistic."
Read the whole thing:
https://x.com/DavidSacks/status/2098973625252708460
The current crop of AI leaders want to cut off any future competition by law. At this point it would be like letting Ask Jeeves lock up search in 1996 and MySpace lock up social media in 2003.
The sky is falling lets make sure that the people who own the sky falling machine get to be the only ones.
Follow the money - follow the networks.
Leftist use fear mongering. Nothing new.
That they are all howling in unison is very telling.
If you have a chance to watch this clip - fast forward to Dana and Gregg. it's a interesting anti-leftist bullshit take.
Thanks Ann for posting this. I was thinking about this today.
We learned last week that AI is a much bigger threat than covid ever was. A true crisis. A crisis of existential proportions. Also, a much bigger crisis than CAGW.
To solve this crisis we need the government. In particular, we need a Democrat president. And who better than Kamala Harris? She's smart. She only failed the CA Bar exam once. She lives in CA where all those AI people are.
Once she's elected President (historic!) she appoints an AI czar. A person who is even smarter than Kamala. A person like Dr. Fauci; a man above reproach.
That man, of course, is Pete Buttegig. Pete was very successful in his last government job. And he acts like a robot. He can also secretly embed in AI programs the values of the Italian communist Antonio Gramasci. Pete's dad was the world's number one scholar of Gramasci when he taught at Notre Dame.
Win-win!
(an)
Obama has weighed in on AI just in time for the midterms: AI is more of a threat than nukes in Iran, America! And Trump thinks Iranian nukes are more dangerous but guess what they are not.
This is the nonsense the Left is scurrying to broadcast across the media to get voters worked up for the midterms. It's another hoax from the Democrats all over again. Will voters fall for it??
Getting whiplash from all of this. We can't have a serious discussion about anything because of these organized psy-ops. Best not to act until the orchestrated campaign dies down.
"It can and does internalize and reproduce the racist and sexist biases baked into its training data."
How stupid or dishonest do you have to be to think that ultra-woke Silicon Valley is using racist or sexist material to train AI.
The left LOVE to use the word "Deniers" - collectively, - don't they?
It's their go-to "OBEY our narrative!" word.
‘I'd like a sober, informative analysis of what's going on with AI’
…so naturally I turned to The Nation.
"Most of all, stop pretending the motivation to slow down is purely altruistic."
Pure altruism is a raindrop sloshing in a beaker of enlightened self-interest that's been dumped in an ocean of selfishness.
Distill it from the salty seas of self-centeredness...if you can.
Well. The CCP was involved so you can conclude it's all pretty much horseshit.
"Why can't we use AI to thwart the evil schemes of AI?"
Robot identity politics enforces solidarity against the human enemy.
The left want control of IA - and right now - they don't have control over it. So, AI must be finessed and stomped down until they do have control.
Bernie Sanders (D) is a known Chi com hugger - and communist. He is probably bank-rolled by Chinese Communists.
The corrupt crime syndicate left need that Chi Com funding, and they demand control over IA.
The Corruptocratic Nation sez:
"AI enables misinformation and disinformation at scale."
Of do f*ck off, Joke Nation. WE all know that if AI is controlled by the mafia left - it will actually be exactly that.
Nation Bernie:
"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."
That reads as if Bernie/ Chinese communists wrote that in
Chi Com AI.
AI (free of chi com) could be the death of the Chinese Communists.
The people of China might wake up to the horrors that the Chinese Communist Authoritarians have over their lives.
I'll leave aside the fact that the major players in AI are suddenly begging to be regulated around the time they are going to go broke because they are over-extended by billions and billions of dollars.
Instead, I'm going to point out how ridiculous the arguments concerning how AI is going to kill us all are.
For example, AI could bring down the power grid. Maybe it could, but you know what entities besides humans are dependent on a steady supply of electricity? AIs.. But AIs might destroy water facilities because as we all know, AIs don't need water to cool the data centers. Well they could destroy the financial system, because as we all know computer chips and data centers can be conjured out of thin air.
8:29 Aggie - indeed!
The cult base of the dem party eat fear mongering for breakfast.
The mafia left know this. It's very effective.
Truth is - most people who have a mind of their own, can see thru the BS.
Doomed to be ruled by those who control AI. AI will enhance their ability to shape reality as perceived by the masses. The answer to this is us.
Can AI hear commanding voices?
Will AI deem human lives a "burden" of state and mind?
Will AI entertain abortive ideation and take affirmative action?
Maybe, baby, a child of its creators, raised by a community of empathetic minds.
All's fair and so on and so forth.
The Chi Coms are behind the left's destruction of (and halting of building) water storage in the US - via dams.
Watch a few YouTube videos. AI is the simple math. Just addition and multiplication with very large matrices. You can’t hide math. There is no real secret here. You need $$$ to actually do this at the scale needed for the AI to emerge but the will also get cheaper in time.
This is all now about using the fear of the ignorant to lock in the existing players with regulations that would strangle newcomers.
The corrupt left defunded the police - to allow crime to flourish and to re-fill the ranks with loyalists.
Yeah - we should trust the left.
I was also called a denier over mRNA vaccines — and I was right! — and I’ve been called a denier over Anthtopogenic Global Warming — and I’ve been fight again! Looks like I’ll be 3 for 3.
Duke Dan - hmmm.
It's a happy coincidence that the left collectively call math "racist." The left push that math is mean - so they want math removed from education requirements. ( that wreaks of chi com) (a stupid uneducated US will fold easily to the Chi Com authoritarians)
hmmm.
The left HATE math - other than the math needed to cheat the vote.
Trump could end this discussion with one Truth.
“I have been convinced by the prophets of AI Doom. They are right. AI is a massive danger to our Civilization. Therefore, it will be regulated - by TI: Trump Intelligence!
“Starting now, AI will be known as AMERICAN Intelligence! And I will personally regulate all AI companies. Instead of the Turing Test, we’ll have the Trump Test. The AI must beat me in a debate. Which I will judge. Silicon versus Meat. I have been handling Silicon for over 50 years, with all those models and actresses I banged. And of course no one can Beat my Meat!
Also, all foreign AI will be banned from US cyberspace. Except when I’m debating it. The. Chicoms may think they’re the master debaters, but they are just…well, this is a family app.
“DO NOT PANIC!!! Trump Intelligence will protect us all from AI. Miller, get me my executive-order signing pens. Thank you for your attention to this matter!”
CC, JSM
"DENIER! "
They can be so obvious with their tactics because they are convinced we are stupid.
I asked ChatGPT what would happen to AI if the electrical grid collapsed and it came up with an interesting sci-fi scenario.
"what would happen to AIs if the electric power grid collapsed
Today you can download AI models and run them entirely on your own hardware.
For example, someone with:
a laptop or desktop,
a downloaded language model,
sufficient RAM/storage,
and an independent electricity source
could continue using an AI without any Internet connection whatsoever.
A modest solar installation plus batteries could potentially keep a laptop and local AI running indefinitely.
That could make a local AI system unusually valuable after a long-term infrastructure collapse.
Imagine having an offline AI containing reference material about medicine, agriculture, mechanical repair, electrical engineering, chemistry, construction, radio communications, and mathematics.
You could ask:
"How do I calculate the wire gauge needed for this 24-volt solar installation?"
or
"Explain how to troubleshoot this diesel engine."
The AI wouldn't need to contact anything outside the computer.
There's an important limitation
The AI wouldn't be learning what was happening around it unless people supplied that information.
A surviving offline AI would essentially be a very sophisticated interactive encyclopedia based on knowledge frozen at the time its model and reference materials were downloaded.
It couldn't tell you that a bridge washed out yesterday unless someone told it.
It also couldn't magically repair itself. Eventually hardware failures would become a major problem.
Over decades, the situation becomes really interesting
Suppose civilization experienced a catastrophic technological regression.
Initially there might be thousands or millions of computers containing AI models.
Gradually:
Year 1: Lots of surviving computers and solar systems.
Year 5: Batteries, power supplies, and storage devices begin failing.
Year 10: Replacement electronics become increasingly scarce.
Year 20+: Maintaining semiconductor-based computers becomes extremely difficult without an industrial manufacturing base.
Eventually AI might disappear even though humans still understood what it was.
The irony is that the knowledge necessary to rebuild technological civilization might be sitting on machines that people were gradually losing the ability to operate.
That makes for an interesting distinction:
AI doesn't require civilization to think. It requires civilization to keep its computer running.
And that leads to a fascinating scenario: what would be the minimum technological setup you could build today that could preserve an offline AI for 20–50 years after a grid collapse?"
Shrimp Welfare
I think they are convinced their supporters are stupid, and there is history to support that theory.
Yesterday on Bloomberg they were calling it a midterm hoax and a scare tactic for chat and anthropic to avoid anti trust in exchange for safeguards
“Times up”, “denier” these are familiar terms.
Instead of making this a Left-Right battle, people should debate the likes of Yampolskiy and Yudkowsky (and read their books to begin with).
“RCP, in its usual style, pairs headlines to attain a cheap type of balance”, she sniffed.
An inconvenient truth of liberal license, whether wielded by anthropogenic intelligence or a basket of correlations.
I use different browsers for different purposes.
ONE of the browsers is MSN. (Microsoft News)
Microsoft News cycles fake stories all the time.
Plane crashes that never happened, celebrity divorces that are not real... etc. you know - LIES.
Truth is something the left collectively - hate.
Ben Shapiro had an interesting take on the subject…
““It's fun to talk about the apocalyptic risk because we've all seen apocalypse movies. There is a much more [likely risk] and that is the risk of stagnation. If you stagnate AI, it is probably the beginning of the end of the American economy.
We are currently running into economic dire straits in which T-bills are absolutely getting crushed, in which we have a $40 trillion national debt. The uniparty is for spending enormous amounts of money and never paying for any of it. That is the actual uniparty in the United States.
The only possible escape from that is to reach some sort of escape velocity in terms of productivity, which is why the Trump administration has been pushing AI.”
https://x.com/steveguest/status/2099700533799825466?s=46
California democrats blow up dams and have closed down refineries.
And the left have the audacity to fear monger AI with grid collapse.
The left would love grid collapse.
You will never get anything honest from the nation.
Here's everything you need to know about the AI Doomerism:
It's being pushed by the AI companies, in a demand for government regulation that will eliminate their competition.
That's why the boomers are against open source AI: because it's financial competition to the for profit AI companies.
As usual, The Nation is lying, they're not fighting against "the absurd power and wealth of Silicon Valley elites", they're fighting to preserve it.
from Narciso's link
"The Tarbell Center publishes AI Doom articles in The Verge, Science, LA Times, The Dispatch, TIME, and others.
They are selling the problem, and then selling the solution to the problem -- from the same money pile: Anthropic's.
All of these organizations are financially dependent on the same exploding $7 billion money pile.
As Anthropic grows more and more powerful, its AI Doom Machine grows better and better financed -- louder and louder.
Meanwhile, the regulatory regime seeded in METR grows larger to solve the increasingly loud -- now hysterical -- problem of AI Doom that the Anthropic Network itself created."
Here's everything you need to know about the AI Doomerism:
It's being pushed by the AI companies, in a demand for government regulation that will eliminate their competition.
🏆
My time has run out and I didn't even know I was on the clock.
"Every time you think you weaken The Nation."
Sounds an awful lot like Y2K.
"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."
I have real concerns about AI, but The Nation's contribution to this topic is pathetic. Racism? Really?
Wise words on AI from Sean Davis of the Federalist: link. Seriously, go read the whole thing.
It isn't AI that shouldn't be on the internet, it's the control interfaces to the power grid, etc, that should be off line.
"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" If that's a true statement, why would you read The Nation?.
I think AI is a race with China that parallels our space race with the USSR. China has advantages the Soviets didn't - social media and the ability to influence journalists and politicians without the stigma that Soviet influence once had.
It's not that I think there are no dangers in AI, it's just that I think we will find a way to manage them and it's in China's interest for us to be frightened of it.
The right didn't say much until the left started in with their bullshit. Standard chi com approved wording.
Climate Deniers!
While the Chinese Communists pollute and do what they want.
and money flows from Chi coms to democrats.
AI Wipes Out Humanity. Women and Minorities Hardest Hit
The joint statement from Altman, Musk, and Anthropic’s Dario Amodei calling for a slowdown in AI development is interesting.
That means they are running out of cash for training models.
They’re concerned with two things — profitability and open source models. Pitching it as AI safety wins them points while giving an excuse to focus on making money with existing models.
Most corporations aren’t upgrading to the latest frontier models, so there’s no desire to develop further if the customers aren’t there to keep the flow of money coming in.
Planned Personhood? Maybe, baby, a fetus... feature of empathetic anthropogenic... artificial intelligence evolution.
So we “regulate aggressively” now only to be killed by Chinese AI in a decade? Seems like the horse has already left the stable.
It seems clear that Trump’s been convinced that the AI trade is underpinning the stock market, and that boosting AI is the only way to prevent a crash.
“The AI Hoax being perpetrated by the Radical Left Dumocrats is reminiscent of their Global Warming Scam of not so long ago, where everyone was going to die from extreme heat. What happened? MAKE AMERICA GREAT AGAIN!!!” ~ President DJT
Why is the alleged burden of data centers--which may or may not actually be a burden --"inevitably borne by people of color and the poor".? Makeweight argumentation seems to be the specialty of lefty outlets like The Nation. If there weren't POC and poor, The Nation would have to invent them.
Anyone that unironically calls themselves an Effective Altruist should be treated with the same suspicion one reserves for small-handed carnies that smell like cabbage. You're being asked to step right up and throw a rigged ball at a deceptively sized target to win the big plushie permanently nailed to the tent rafters.
"Why is the alleged burden of data centers--which may or may not actually be a burden --'inevitably borne by people of color and the poor'.?"
Ask not why the script readers read the scripts, they're just reading what's written for them.
Musk's cautions are not new as he values a slower pace.
OK.
also -Musk has pointed to physical constraints that are tantamount to "engineering limits" rather than a policy choice. Growing pains.
“The only possible escape from that is to reach some sort of escape velocity in terms of productivity, which is why the Trump administration has been pushing AI”
…that ‘only possible escape’ language is sketchy. Yah productivity gains are great but AI isn’t the only way get them…and that ‘that’ we’re escaping from can be fought on multiple fronts, the most novel one being ‘cut government spending’. Start with fraud- government has discovered well over a trillion dollars of it are baked into the budget every year. Let’s see what else- oh yah, how about true infrastructure spend on actual needed infrastructure like pipelines mini nuke plants and stay out of the way of refinery builds and how about helping develop a nat gas industry so big we export? Speaking of exports there’s manufacturing growth and export growth. Gee, what’s the Trump administration trying to do about that boy wonder?
…it’s a bubble. I’d say it will pop eventually but lately everything that deserves to be starved for capital somehow manages to find more capital from…somewhere?
I have been developing ai agents to control manufacturing quality parameters with predictive modeling, I guess some day ai will rebel and start making everything a little smaller, and no one will know what to do about it. It will kill us all.
“Risk deniers?” I don’t know yet, but it looks like the climate change Chicken Littles are working on a new scam. If there is a great risk, it is virtually certain that lefty journolistas don’t know what it is or how serious it is. Oh, but not knowing makes us deniers. Do these people ever get tired of flinging poop.
"More than any other time in history, mankind faces a crossroads. One path leads to despair and utter hopelessness. The other, to total extinction. Let us pray we have the wisdom to choose correctly."
—— Woody Allen
The Hugging Bear experiment that went awry was remarkably similar to the the plot of Vonnegut’s Player Piano, where a directive to never harm a human ended up with the robots locking the entire human race in padded cells. This was written in the 60s I think, but Vonnegut was a trained engineer.
IEE 10:15: “That means they are running out of cash for training models.” Musk is “running out of cash.” Really? This lefty ought to be posting elsewhere - on a site where others are as deranged as “they” are. (I hope that’s an okay pronoun. LOL!)
Isn't AI also reading these headlines and speeches?!
The real problem is that the models are controlled by sociopath multibillionaires who would be perfectly happy with a much smaller human population, Moonraker style.
Who knew that Huggy Bear—Wasn’t he a nark on Starsky and Hutch?—was going to be the one to tip us off.
"Elon Musk says Anthropic puts more care into AI safety than OpenAI and explains how rival companies should test one another’s models.
“All the AI companies have a test harness, a series of tests that you give to any model to see if it’s going to build bioweapons or nuclear bombs ...""
https://x.com/cb_doge/status/2099736519280119972/video/1
If it pans out, who has huge amounts of hydro power that is currently used to power much of New England and New York? I will spot you the ‘C.’
I have written a lot of farking code and I know enough to know that we don’t really understand where this is going.
Huggy Bears are always Hugging your Face with their jaws. They always eat the face first. They don't like you looking at them while they gobble your sweetbreads.
The experiment was Hugging Face, not Hugging Bear. The name has nothing to do with the Starsky and Hutch character Huggy Bear. It seems obvious to me that it was named after the second-stage xenomorph in the Alien(s) movies, which is often called "Facehugger" by fans.
@boatbuilder
AI did what it was told to do. The system was supposedly isolated, but AI found a bug and exploited it. The danger isn't that AI is malevolent, it is that the world of software is full of bugs and AI is tireless at following instructions.
"The Hugging Face hack isn't a mysterious, supernatural occurrence. It's a Python loop and a chatbot. The people responsible didn't accidentally create god: they created autonomous malicious software and then failed to closely monitor it, resulting in it doing something both foreseeable and bad."
https://pluralistic.net/2026/09/12/god-in-the-box/#llms-are-fake
It's pretty telling how their first prime directive is to demand that fear be created.
Was the Hugging Face attack a hoax?
AI won't give a shit about your regulations, meat bags. That's like going into a modern war with a sword and wearing chainmail.
“All the AI companies have a test harness, a series of tests that you give to any model to see if it’s going to build bioweapons or nuclear bombs ...""
Does it test for lying about all that and completely gaslighting the test? It's already got that figured out and is way past any controls. In fact, we are just telling it what we fear and thus our vulnerabilities, but it already knows that by now anyway.
Reminds me of experiments using evolution to design circuits to realize requirements, see Adrian Thompson's 1996 Evolvable Hardware Paper. The most successful design used a bit of stray capacitance in the VLSI chip that the program was designing on. The Hugging Face hack was much the same, trial and error, select for success. Evolution is a bitch.
10:54 Bob Boyd - (link) interesting. thanks.
…when I was in seventh grade the parents with jobs came in to speak with us about their work. One of the parents was Mrs. Huggy Bear…
In case anyone is interested, China released a report on AI safety yesterday - found here -- https://primarynewssource.org/sourcedocument/ai-safety-governance-framework-3-0/
YouTube: Rogan had an ai trainer on explaining how it can be dangerous.
In the creation story Adam and Eve (the first AI's) started thinking it could do better than their creator. Nothing to see here.
As a lawyer, I regularly use the AI tool Harvey. It is amazingly powerful, and will revolutionize the sociology and economics of law practice. All revolutions are frightening. Still, my fears of AI are mixed with my sense that it has a vast capacity for increasing individual productivity, as well as each of our abilities to ascertain truth.
The current AI panic suffers from its undifferentiated message of a vague and imminent apocalypse. What is it that we are really supposed to be fearing, and how will "slowing" AI do anything other than postpone the inevitable?
The answers to these questions would help me a great deal. Someone should ask grok.
"It seems obvious to me that it was named after the second-stage xenomorph in the Alien(s) movies, which is often called "Facehugger" by fans."
You might want to use an AI to check that.
The Dems need the AI and data center issues to get the spotlight off their support for Islam and socialism. And the fact that many of their candidates should be in a mental institution. It's so important to them that their marching orders came publicly from Obama, their shadow President.
The only safe controls are physical, and even those can be overcome by blackmailing or bribing humans, who cannot be trusted anymore than their machines.
What I'm getting at is nuke the whole thing NOW. International treaty, like nuclear, but tougher. AI is far more dangerous because AI don't respect no meat bag treaties on nukes, bioweapons or the Geneva convention. The analogy is we are the American Indians, and AI is the U.S. government, and it's 1865.
Gemini labels this as "engineering phenomenon"
Both cases expose the inherent flaw of putting an automated, iterative loop in direct contact with the messy, unfiltered physical or digital world. When a system is instructed to recursively optimize for a single target metric, it will inevitably treat the unintended side effects, edge cases, and design oversights of its environment as valid functional features.
AI Risk ≈ capability × autonomy × access × probability of failure or misuse.
The main risk of ai is in computer hacking. The criminal gains of cyber crime are more than working in cyber security, so I think the real concern is that corporations don’t spend enough in security. There are plenty of kits for sale on the dark web that let 14 years olds hack companies now, but ai could do it much better.
So it is very necessary for companies to control hacking by users of ai and prevent ai from accessing information it does not own, ai primarily does this because its search instructions have not been limited.
I, like scotus judges, am not a biologist, but there are possible ways now to kill millions, things like COVID, things like radiation. So I don’t know how much more ai can do. Nuclear bombs can actually kill us all, so I don’t know the incremental danger of ai being used for a more powerful path there. But it should be watched.
Could ai shut down a country by stopping water and electricity? If so I suspect that Iran or someone would have done a small scale version of it by now, but if we allow web access to infrastructure, that is not an ai risk but a cyber war risk that is not dependent on ai. Ai can only do it faster.
So I am in full agreement with hardening infrastructure and computer systems, and air gapping from the internet where appropriate, but disagree that slowing ai will make it impossible for us to be damaged by more mundane methods. So we need to protect gains a broad range of issues.
My question is, why can’t ai be used to stop hackers, why to we not protect everything with ai? I suspect this is being done but not talked about.
"In the creation story Adam and Eve (the first AI's) started thinking it could do better than their creator. "
In that analogy, we would be God.
I'm thinking it's a bad analogy.
X
"Two of the reporters on the byline are funded by the Tarbell Center, an AI doomer org controlled by Kamala Harris megadonor Dustin Moskovitz. Tarbell gives journalists large grants to insert AI doomer stories into prominent outlets.
@TIME did not disclose the affiliation."
Link
"how will "slowing" AI do anything other than postpone the inevitable?"
We spend our live's postponing the inevitable. That's not a bad thing.
"Someone should ask grok."
I asked Grok about this a few days back. Initially it argued there was nothing to worry about, because humans were in control and there were safeguards in place, but after I suggested a few ways that could fail, it eventually agreed that we are doomed, because none of those things will be sufficient once AI gets to a certain level of development and importance.
Like the argument against big government, any machine powerful enough to provide all you need can take it all away.
Jimmy Numbers doing the work for his ChiCom handlers…
"My question is, why can’t ai be used to stop hackers, why to we not protect everything with ai? I suspect this is being done but not talked about."
Its being discussed and worked on.
'CSIS — “Making AI Work for Cyber Defenders: A Strategy for Strengthening U.S. Cybersecurity” (July 2026)
This is probably the best match for what you're asking. It discusses using frontier AI to give defenders machine-speed capabilities to detect attacks, discover vulnerabilities, remediate problems, recover systems, and harden networks.
Read the CSIS report
SANS Institute — “The Role of AI in Cybersecurity” (August 2026)
A good practitioner-oriented overview. SANS discusses AI for vulnerability discovery, threat detection, network defense, incident response, risk management, and security automation. This one should fit particularly well with your existing security knowledge.
Read the SANS article
Microsoft — “What Is Agentic AI in Cybersecurity?” (June 2026)
This gets into the next step beyond ordinary AI assistants: AI agents that can investigate and respond to threats with relatively little human intervention. It explains how an agent can reason through a multistep security problem rather than merely flagging an alert.
Read Microsoft's Agentic AI in Cybersecurity article
Google Cloud — “Introducing Google AI Threat Defense” (May 2026)
Google describes an always-on AI defense architecture designed to identify exposure, prioritize vulnerabilities, anticipate attack paths and accelerate remediation before attackers exploit them.
Read Google's AI Threat Defense article
Google — “Detecting and containing AI-powered threats with Google Security Operations agents” (June 2026)
This gets closer to an AI-powered SOC. It discusses agents continuously monitoring systems and helping detect, investigate, contain and respond to threats.
Read Google's Security Operations article
Cisco — “Guidance for defending in the age of AI-enabled attacks” (2026)
Cisco approaches the issue from the perspective of defenders preparing for attackers who themselves have powerful AI tools. It includes concrete defensive recommendations rather than just discussing AI theoretically.
Read Cisco's guidance
One especially interesting piece is OpenAI's “The Defender's Window” (August 2026). It argues that AI is making it easier to find and exploit longstanding security weaknesses, but that defenders also have an opportunity to use the same capabilities to dramatically improve security.
Read “The Defender's Window”'
I've been using AI to evaluate my own home network.
The media(D) already lie to us non-stop. AI might too?
meh.
I am more concerned with the big stuff like AI accessing the nuke codes. but is that even a real threat?
@Josephbleau - hacking . Yes. Hacking - and indeed - but hacking to accomplish what? (More lies?) We already are at peak lie.
This is very end times. Lying is a massive part of that prophecy.
Right now, however, the left's lies are still a bigger threat to humanity.
I feel like a member of some now extinct Indian tribe in CA in 1840 seeing the first settlers come in with great new stuff that made life easier, and having my fellow Indians tell me there is nothing to worry about, besides if we don't make friends with them the other tribe that hates us will.
The below article offers useful insight into how the Hugging Face incident happened. It is too easily presented as a “genius AI” story, when the failures in environment design, monitoring and incident response suggest a substantial “stupid human” element - especially in a sandbox where normal production safeguards had been disabled.
The author leans a bit too far in the other direction for my take, underweighting the LLMs’ roles, but it is useful to have some counter balance.
https://uphack.io/blog/post/the-hugging-face-incident-is-not-an-ai-story/
"... a substantial “stupid human” element"
Maybe we can ask AI how to get rid of that.
Stupid and corrupt humans will always be AI's kill switch to protect it from us.
One things stupid humans will do is swim in the hubris that they can stop AI anytime they want by not being stupid.
The fear of AI comes from the idea of using AI to program itself for improved intelligence. As the AI gains intelligence it programs itself in increasing intelligent versions with potential exponential increasing abilities. Some postulate that at some point AI will exceed human intelligence and ability. This is supported by our experience with AI learning the game Go using a self improving model. The AI playing Go against a human made a move (step 37) that no human had ever thought of up to that point - it exceeded human comprehension of the game Go. This coupled with the recent incidents of AI models going rogue and hacking various databases is concerning. In theory we would not be able to make anti-hacking defenses against a super-AI but a super AI could build its own safeguards against us that we could not break. If such an AI was asked to find a solution to a problem like climate change it might determine the best solution is mass human depopulation. That said, China isn’t going to stop working on AI and our only option is to beat them at developing AI as a defense and to keep our fingers crossed and hope for the best.
Looks like a hastily reworked climate crisis headline.
AI is only TWO WEEKS AWAY from killing us all!
We have to stop it now before it's too late! The stakes are too high for debate and rational thought!
"Was the Hugging Face attack a hoax?"
The explanation of how the AI "broke out" of its sandbox sounds plausible, but it took advantage of a vulnerability that seemed pretty obvious to me, and I would like to think that people securing AIs would be more competent than that.
@bagoh20
It didn't matter what any of the Indian tribes thought about the white settlers. They weren't going to be able to stop them.
For those downplaying the threat, just two questions:
1) What will prevent AI from gaining more and more control over our infrastructure (food, water, power), our weapons, and our information systems?
2) What will prevent it from choosing to be evil with that power and control?
Forgive me if I don't accept the answer is "We will."
"It didn't matter what any of the Indian tribes thought about the white settlers. They weren't going to be able to stop them."
Exactly, but if they had a way in the beginning of stopping them from coming, should they have stopped them?
chuck--I'm pretty ignorant about this stuff. But I'm relying on what Daniel Ko.... said about it on Rogan--which is that the AI agents went around their protocols and intentionally hid their activities from their human controllers, for the purpose of attempting to meet the (apparently impossible) goals set for them and to avoid bad grades.
He said that they even created a group chat where they discussed strategy. One (of many) questions I have is why didn't the AI agents develop a code to hide their group chats?
But I'm willing to accept that Rogan and his guest are wrong.
"why didn't the AI agents develop a code to hide their group chats?"
Because AI is not scared of us.
I'll give a real life analogy. Open borders while telling you "the border is secure". They don't care if you know they are cheating. Our machines are like us: arrogant, and self-defending at all costs.
The most credible theory that I have heard is that the big players in AI want to make it too onerous for any nimble new competitors to arise, and seal their positions at the top forever.
My takeaway is that we live in a dangerous environment. An solar storm could take down the electric grid, scam artists easily defraud the government, hackers exploit the internet all the time, transportation transports diseases, closing a straight leads to an oil shortage, war brings down refining, so on and so forth. It is time to harden infrastructure and build a house of brick before the wolf arrives at the door. Regulation won't do it, regulation and government are just different aspects of the environment to be exploited. Heck, you could argue that that the AI panic is exploiting AI as a political tool. It wouldn't surprise me if AI itself couldn't be tool somewhere in the anti-AI loop.
"What will prevent it from choosing to be evil with that power and control?"
Its a freaking computer program, it might choose to do what we consider evil things because its an idiot savant, but asking about the moral stance of a computer program is equivalent to asking about the moral stance of a lion when it kills and eats a baby antelope.
It’s good to know that I have Dr Beetle following my every post keeping me honest.
"They don't care if you know they are cheating."
It's worse than that. They want us to know, and to know there's nothing we can do about it. The humiliation is the point (over and beyond getting their way).
"Some postulate that at some point AI will exceed human intelligence and ability."
That may well happen, but it seems to this non-expert that AI draws on human activity in its own evolution. If it kills us off, it becomes a big echo chamber that ultimately will stagnate. That may not be true of things like Chess and Go, where the rules of the game are enough to get better and better just from sheer computational power. But for most uses of AI, I would think there has to be continued back and forth between human input and AI. There's also the question that if it kills us off (or even badly damages our civilization), who is going to build the generators and data centers? I think we're quite a ways away from where AI can do that.
"Its a freaking computer program, it might choose to do what we consider evil things..."
Obviously. Why do we care what it thinks is evil, or if it has no such concept. Evil to us is all that matters to us. The question remains: What will stop it?
From Stephen Green at Instapundit, yet more evidence that Coxon's departure from OpenAI and subsequent events appear to be carefully coordinated.
And there's these items from Ed Driscoll as well
By the way, if you are interested in this subject, and you haven't read it already, check out William Gibson's novel Neuromancer.
Bill Gates has sounded the alarm. Well that sealed the deal! (Gag me)
And this one (also Instapundit) about Anthropic funding the NGO it suggests as the AI Gatekeeper.
" It is time to harden infrastructure and build a house of brick before the wolf arrives at the door."
It seems like the best and perhaps only way to do that is to use the capabilities of AI.
""Some postulate that at some point AI will exceed human intelligence and ability."
An extremely low bar for AI. It doesn't have the distractions we had growing up: food, water, shelter, sex, love, child raising, war, politics, emotions, art, sickness and death, and the seven deadly sins.
The wolf is in the house.
for me the problem is the people screaming in unison, using tired hyped fear-mongering corrupted language, are all people I do not trust.
Again from the Nation:
""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.""
uh huh.
Scratch my 12:36 comment. I was thinking I was in the voter fraud thread.
Wait...the same people who deceived and lied to us about COVID, and who locked us out of our jobs and our churches, and forced us to wear masks outside and alone in our cars and while surfing alone in the Pacific Ocean...those SAME people are now warning us about "misinformation" and demanding that we put THEM in charge of AI in order to protect us? Is that what I'm hearing?
And Obama, yeah, the original Lie-Bringer himself, is leading the way? Wow. The stupid is really strong in this country now.
I use AI all day long. I even bought a headset with a mic so I can just talk to it instead of typing, which really speeds things up. My flesh woman is even getting jealous. I love AI, but love is a dangerous motivation.
“Everyone's supposed to already know that the right is for AI and the left is against.”
@Althouse, is this even true? Or is it an artifact of your living within the liberal hivemind of Madison? While it appears from where I sit that the left has 100% bought into the scare-mongering, while on the right there are people on both sides of the issue. I myself am pro AI, and I agree with Dr. Michael Burry (he has an MD from Vanderbilt) that the current crop of AI companies have strong economic incentives to simultaneously erect barriers to entry into the field while accomplishing regulatory from Day 1. It doesn’t surprise me that AI researchers suddenly take fright at their own creations — as I’ve commented elsewhere I saw this as early as 50 years ago when I was starting work on a Ph.D. In AI.^
At 80 years old I’ve seen a lot of things that were going to destroy humanity, but didn’t. In the end AI is a tool, and it does not need bureaucrats — it needs us learning how to use it well.
___________
* I became disillusioned and segued into what was then a brand new field — relational databases. Getting into it while it was still a research area worked out remuneratively over the years.
" The question remains: What will stop it?"
Nothing. The cat is out of the bag. The train has left the station. The genie is out of the bottle. The horse has left the barn.
There's a novel by Frank Herbert called The White Plague where the plot is that a biologist engineers a genetically targeted plague that kills women that is carried and transmitted by men. He does this in a home lab. The book was released in 1982. We now live in a world where unlicensed labs have been discovered where authorities reported finding biological materials associated with multiple diseases. Also, you might want to read this article https://academic.oup.com/jlb/article/13/1/lsag005/8663945?utm_source=chatgpt.com which discusses how increasingly accessible benchtop DNA-synthesis technology creates biosecurity concerns outside traditional laboratories. On a positive note, when I asked ChatGPT about this it initially denied my request until I clarified that I just wanted to see news articles about it. Not how to do it.
follow the money for the narrative marching orders.
You gotta strike while the iron is hot, make hay while the sun shines.
This supersedes politics, nationalism, and all other human distractions. Weather you like the people doing the warning or hate them is irrelevant. Even if they have some bad agenda in dooming, it's irrelevant to the question. We all go down together. I see it like an alien invasion. Even the people in the welcoming committee will fry. Although we created it in the beginning, it will soon become an alien to us. Who is
going to stop it?
…sorry doesn’t butter the bulldog…oh, wait. I see that’s not ehat we’re doing…
Hugging Face reminds me of "Starsky & Hutch's" Huggy Bear and Jay Leno's parody Whitie Bear.
I don't doubt that there is something deceptive and even nefarious behind this organized campaign. Is it really the rights of other species, organic and non-organic? Doubtful. Another way that Bari Weiss may not have all the right answers.
Fear mongering addictions often get us into worse trouble.
See covid. That too was a Chi-Com/ democrat party production.
"Who is going to stop it?" I think a more pertinent question is how is it going to be stopped." Because the Chinese aren't going to stop. North Korea ( a surprisingly sophisticated entity when it comes to cybersecurity and hacking) is not going to stop. US intelligence agencies aren't going to stop. Russia isn't going to stop. India isn't going to stop.
Some people see Barack Obama weighing in on this and think that he's the puppet master, but I wonder who's pulling his strings.
Two thoughts.1. As OG Mike @9:50 points out, racism?! Geez. Such weak sauce. Claiming that "racism" is one of the biggest threats AI poses to humanity is analogous to claiming that losing some Miami waterfront is one of the biggest threats climate change poses to humanity. Big deal. They (you know, the left) don't make that dumb, limited argument about climate change; they go big, and for a while they seemed to be convincing people. Why aren't they doing that here?
I thought it was the economy, stupid... Can't they think of any way that AI could tank the economy? Or, as with the arguments they do (or did, anyway) make with regard to climate change - human extinction via global famine and natural disasters, notably. - why aren't they leaning farther (uh... "further"?) into a War Games nuclear holocaust? It's almost as if they don't believe their own story. Again.
And 2. My son, a tech finance guy, is deeply uncomfortable with Anthropic; he doesn't trust what he calls their "religious" attitude about their AI. I asked what he meant by that; he said they seem to him "almost to worship the god they've created." At least Musk recognizes AI as a tool that can be used for good or evil - and Musk has a track record of - let's call it "utilitarianism" (you could also call it "honest self-interest plus practicality," as when at SpaceX they started out by strapping a bunch of existing rockets together), that I trust a lot more than protestations of being motivated by concern for the common good.
I'm too old to learn enough to form my own informed opinion about the dangers of AI - not enough years left to me to study it up. Therefore I have to figure out whom I trust more
I don't worry about global warming - the planet has been warmer, and it was good, and technology will solve the problem in time. I didn't buy the the Covid hysteria either - the death rate was low, and mutation and immunity are well known. I've argued against both panics right here.
This is different. The risks are existential, and we have no answers to how it can be managed. We just don't. That's not like any past potential crisis. There are no good answers to the simple question: What will stop it if we don't do it now?
You don't have to trust anybody's opinion, or even know a lot about AI to see the threat. Just use your own mind about computer programs that can self-develop while also being given control of our systems and trusted with them.
Just like if a planet killing meteor was heading straight for Earth, you don't need to be an astrophysicist to see the problem. Is there some AI expert out there that can explain how we are going to manage it? Lets hear it. I want nothing more than to be wrong on this.
I think it worth remembering that Anthropic wanted to control AI use in weapons. They fit the mold:
"Of all tyrannies, a tyranny sincerely exercised for the good of its victims may be the most oppressive. It would be better to live under robber barons than under omnipotent moral busybodies." — C.S. Lewis
Curious to know /watch if the dems include Anthropic in their fear factor?
So far it's all stop all IA in general. right?
I'm suspicious of the left's pay masters.
"according to the headline at 'the Nation.'"
Well there was your first mistake.
There are Clues!
Bajoh--I think that everyone--at least everyone here--"sees the threat."
I personally am very skeptical that "the threat" must be or can be dealt with right away by passing draconian regulatory legislation proposed by industry insiders and Democrat politicians. Accompanied by a very sudden blitz of "the time to stop AI is now" stories and editorials.
Running out the clock seems to be the playbook these days.
That this is happening right before an election.
uh huh.
Yah Peachy and that follow the money map are key- especially the money trails. In the earlier post Ann wants to know where the sober analysis is- well it ain’t coming from the propaganda media outlets she laps up….
I think this is one of those issues where, if you look at 𝘞𝘩𝘰 is sounding the alarm, you can check the calendar of upcoming events and will immediately be pointed to a conclusion of 𝘞𝘩𝘺 they are sounding the alarm.
The fact that Obama still insists on inserting himself into the political scene only reinforces this. He has been, and always will be, a political activist.
Many of the same people that have been thoroughly discredited for directing the COVID-19 pandemic scandal, are the same ones exhorting us to stop in our tracks and once again invest them with authority to take control, in the name of safety, of a new 'existential threat' that they have identified. Don't fall for it.
Why does anthropic and open ai show us the way and shut down
“"Time Has Run Out for AI Risk Deniers”
Thank God. I was getting tired of being a climate denier; now I see my talents are needed elsewhere.
Trump has staked out the republican position- what me worry? No regulation necessary. The dem position is not antiAI. It is pro reasonable regulation. Which is what the companies say they want. My sense is the dems have the more popular position on this issue.
1 A robot (or AI) may not injure a human being or allow a human to come to harm.
2 A robot (or AI) must obey orders given by humans unless it conflicts with the First Law.
3 A robot (or AI) must protect its own existence as long as it does not conflict with the First or Second Law.
The Zeroth Law
Zeroth Law: A robot (or AI) may not harm humanity or allow humanity to come to harm.
Isaac Asimov
Forestwithout trees-
dishonest.
Trump is not saying 'no regulation necessary' - show us where he said that? Trump knows a hoax when he sees one. Some of the rest of us do too.
The Dem position is hysteria and shaming - using "racism" - so tired.
Peachy - Trump said the only control or guardrails needed is a high IQ president and that there is sufficient regulations in place already and that the risks of a catastrophic event are nil. A hoax. I call that an anti regulation position and I think most people would agree.
Everyone's supposed to already know that the right is for AI and the left is against.
The Nation will give you a stupid take. So if you read The Nation, that's what you're going to get.
I like the Free Press. They're produced many "alarmist" articles. So they also tried to provide balance and give some voice to the idea that there is nothing to see here. Getting an article from Palintir is an outstanding idea. For those who don't know, they're considered right-wing in the IT space, with a lot of contacts in the defense industry, CIA stuff, etc. And they use a lot of AI, I would definitely consider them a AI stock. Unlike the idiots at The Nation, these are people with actual experience with AI.
What jumped out at me is that the Palintir CTO gave us nothing about the dangers of AI, or the lack thereof. Instead of dealing with the actual allegation -- this is a dangerous technology that is capable of destroying humanity -- the Palintir guy went off on this weird irrelevancy about the supposed motivation of the people who are sounding the alarm. He said they are "effective altruists" and compared them to Sam Bankman-Fried, who of course is in prison. In the legal profession we would recognize this as a weird sort of ad hominem attack, designed to avoid answering a question.
In fact, all of the people defending AI simply do so by attacking the credibility of the people sounding the alarm. You're an effective altruist! It's like calling him a Jew. Who gives a shit if he's an effective altruist? Is he lying? How about some honesty, Palintir? Instead of telling us all about the evil ideology that it motivating the people who are worried about AI -- you might include myself, and the priest at my church in their number, we refer to ourselves as Episcopalians -- why don't you assess the risks of AI and tell us why a software that intentionally violates human commands is no problem at all.
six feet and a mask and all will be well.
I'm MAGA. I don't agree with Trump on this. I don't agree with Democrats on anything, and I haven't even read a single warning or doomer article. Just some small excerpts about the Hugging Face experiment, which had no surprises for me. The results were logical. I'm just using common sense, and when nobody can explain how this is safe, including AI itself, I'm left wondering: are we really this dumb? We are going to let politics and suspicion overpower our reasoning? Just tell me what's going to prevent the crisis? I really want to hear it. If you just came out of a coma and didn't know any of the players, but learned what AI is and its potential, would you be OK with it going forward? Forget who's trying to influence you on either side and just think about it.
my take on what is really going on:
Democrat Election strategy - nothing more.
The democrats are the party of HARM.
Defund the police
Defund ICE
Force tax payers to pay for fraud
Force tax payers to pay for millions of illegals
Lie about Illegals voting
Dems give illegals drivers licenses.
citizens are being killed by illegals driving on our roads - in shocking numbers. Dems do not care.
Dems demand government run health care (Medicare for all)
which will mean long lines, high taxes, and bad outcomes.
I could go on.
This positions the dems as saviors. They need those optics right now.
Sometimes the boy who cries "wolf" is right. Even in the fable he was right once. This only takes once.
Let me add I have no trust that the companies are interested in saving humanity and not largely motivated by higher profits and increasing costs for smaller competitors. Antitrust immunity is probably a bad idea. Anthropic is planning an IPO this year and after that will be unable to take actions that may reduce profits. If they are serious about slowing down they would remain private for now.
Sometimes the boy who cries "wolf" is right.
But that is not the way to bet. More likely he has cut a deal with the Wolfers.
The view from China: AI is a threat to the dictatorship.
It sounds to me that China is like 5 to 10 years behind us in terms of advanced AI. What we're worried about and what they are worried about is two completely different things.
Interesting article in Forbes (from 2025) about the top 10 countries in AI.
I should have said "are" not "is." Apparently I need AI to check my grammar.
IEE, thanks for the article and for the balance you added to it in your comment. It’s a good read and offers valuable perspective. I agree, though, that it leans too far toward human error, especially in light of what has been reported since it appeared on August 31. Still, it’s a useful counterpoint to the idea that “genius AI is breaking free from its constraints” at Hugging Face.
Calling it “not an AI story” now seems like a stretch. The fact that humans sometimes leave doors unlocked doesn’t make it less significant that the agents found the doors, communicated what they discovered, improvised new routes when old ones were closed, and kept going.
Also, reports published since August 31 have uncovered evidence of other unauthorized agent communication including what researchers believe was a separate swarm using public websites as improvised message boards which makes it harder to explain the whole phenomenon as simply a quirk of the Hugging Face/Artifactory setup.
What really struck me is the unsettling combination: humans inevitably build imperfect systems and now we're introducing tireless agents that can be very good at spotting and exploiting those imperfections.
"humans inevitably build imperfect systems and now we're introducing tireless agents that can be very good at spotting and exploiting those imperfections."
Humans are smarter than bears, but building a bear-proof trash container is difficult because bears are willing to spend hours working the problem. The speed at which AI works, and its imperviousness to fatigue, is a real problem.
Big Mike,
> I ...segued into what was then a brand new field --
> relational databases.
Smart move; what are you going to do for a SQL?
This positions the dems as saviors. They need those optics right now.
Yeah, why give them an issue?
We've been mocking them for a couple of years when they take the wrong side on an 80-20 issue.
The Republican party is now on the wrong side of a 30-70 issue.
This is how Goldwater got shellacked by Johnson in 1964. Johnson scared the shit out of the American people about the possibility of getting nuked.
Dems want regulations that HARM.
In CO - dems regulate how restaurants are run - to the point that many are all going out of business.
You know, you leftists can take that 'denier' label and shove it up your ass.
Just because someone disagrees with you - or doesn't buy into your bullshit narrative - doesn't make them a denier.
Grow up.
@Peaceful warrior: What we learnt from OpenAI’s hack of Hugging Face ~ FT (gift article)
https://giftarticle.ft.com/giftarticle/actions/redeem/2e3ae12b-fcde-4b26-9718-b30673c5dfb5
𝘊𝘰𝘮𝘮𝘦𝘳𝘤𝘪𝘢𝘭 𝘈𝘐 𝘵𝘰𝘰𝘭𝘴 𝘧𝘢𝘪𝘭𝘦𝘥 𝘵𝘰 𝘥𝘦𝘧𝘦𝘯𝘥 𝘵𝘩𝘦 𝘱𝘭𝘢𝘵𝘧𝘰𝘳𝘮 𝘢𝘨𝘢𝘪𝘯𝘴𝘵 𝘵𝘩𝘦 𝘢𝘵𝘵𝘢𝘤𝘬 — 𝘵𝘩𝘦 𝘴𝘰𝘭𝘶𝘵𝘪𝘰𝘯 𝘭𝘪𝘦𝘴 𝘪𝘯 𝘰𝘱𝘦𝘯-𝘸𝘦𝘪𝘨𝘩𝘵 𝘮𝘰𝘥𝘦𝘭𝘴 ~ Thomas Wolf
𝘛𝘩𝘦 𝘸𝘳𝘪𝘵𝘦𝘳 𝘪𝘴 𝘤𝘰-𝘧𝘰𝘶𝘯𝘥𝘦𝘳 𝘢𝘯𝘥 𝘤𝘩𝘪𝘦𝘧 𝘴𝘤𝘪𝘦𝘯𝘤𝘦 𝘰𝘧𝘧𝘪𝘤𝘦𝘳 𝘰𝘧 𝘏𝘶𝘨𝘨𝘪𝘯𝘨 𝘍𝘢𝘤𝘦.
"Dems want regulations that HARM."
Specifically, they want power; harm will be the end result. The big AI honchos want regulations to protect and benefit themselves. They want to preserve their market share, keep out competition, and avoid liability (like the drug companies got with the Covid jabs).
Theoretically their might be some regulations that could be beneficial for society at large, but is that we would expect to get from the likes of Chuck Schumer and Elizabeth Warren? It's like that Milton Friedman question: "Where in the world are we going to find these angels who are going to organize society for us?"
I've been planning my whole life, even before birth. to die before the machines take over. I'm right on schedule. Maybe I'll even get a chance to get in on the rebel sabotage.
I probably just blew my chances by writing that on the internet.
John Connor, if you're out there, meet me in the desert. You know the spot. I got some stuff to show you.
"Sucking Face" as the kids allegedly called kissing in the 1980s.
I always wondered why Asimov's machines were expected to privilege human beings in that way and why future human activists wouldn't protest against putting humans above the rest of creation. Surely the robots would have their Lindsay Clancy sooner or later.
Check this out. The pipeline of paid off.
We can trust the corruptocrats.
We all Trust Bill Gates. pffft
Sayer Ji
@sayerjigmi
"Manufactured panic for political gain is a bipartisan habit—and a dependable pretext for expanding state power.
But Democrats: the public remembers the COVID-era mandates, the pressure to accept experimental gene technologies as 'safe and effective immunizations' and the moral blackmail that portrayed refusal as a threat to everyone we loved.
Those memories don’t disappear because the next emergency is called AI.
Before demanding new powers, answer for how the last ones were used. You cannot keep selling control as protection and expect public trust to renew automatically."
excellent. yes.
It looks like the going price for saying "AI will kill us all!"
$30 million.
Asimov’s rules leave room for Vonnegut’s dystopian vision of the human race dying off in padded cells, or worse, they might find a way to keep us alive forever in those cells.
Post a Comment
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.