Anthony Chapman retweeted
Tell me you have zero DevOps skills without telling me you got zero DevOps skills.
And LLM is not the I Love You Virus. It is not the same as infecting a web server. It is not the same as hacking a wiki. One is a commodity piece of software that is widely distributed and this capability is a non-trivial leap in hardware/software/capabilities and one does not automatically follow the other.
Currently, it is a massive effort to run these models, usually distributed across multiple very expensive chips, running on specialized inference software, specialized harness software, databases, cloud computers, with tooling and hundreds of install packages, etc.
Kimi K3 is 1.6 TB of space at 2 trillion parameters and you need about 1.5 TB of VRAM, which means you need 8xGB300s which will cost about 350-450K a year to purchase or as much or more to rent.
Now let's say Astra or Mythos or later models come in at 10 trillion parameters, so that puts us at roughly 10 to 20 TBs unquanticized. We are guessing her because these folks reveal nothing about architectures anymore, but they are reasonable guesses based on scaling laws. You would need roughly 23 TB of VRAM to run it and roughly several million dollars to run that one instance a year. And to make it worse, it no longer fits on an 8 way cluster, which means you now need special networking configuration and inference software to run it or you need to infect a Nvidia NVL72 rack-scale server system costs between $2.8 million and $6.5 million.
So the "rogue" model copies itself to the most expensive compute in the world. It's safe to say that hardware is not sitting idle and is being watched closely for downtime or anomalies. In fact, it is likely doing one of the following things:
1) Serving customers at a hyperscaler
2) Serving inference for a well endowed company.
In the first case it will be instantly noticed by the actually good automated monitoring systems of the hyperscaler because there is now something using their infrastructure but not being billed properly and costing them millions of dollars in lost revenue and electricity.
In the second example, the model they were serving on that very expensive hardware is now offline (replaced by the rogue model). The model presumably was connected to a valuable piece of internal software that is now glitching and no longer running.
Nobody notices this?
And so we see that this is where sci-fi meets the little problem of the real world where friction, dust, time, resources limitations and the like come to crush your little fantasies.
To be fair, this *may* be possible in coming years. I assign it a non-zero probability. Breakthroughs in new architecture and smaller models and eventually the wider distribution of the computing substrate that commoditizes could change this materially. But for now, it is mitigable and foreseeable and an engineering problem. In no way will these things be copying themselves around like a virus on your computer in their current form.
It also betrays a lack of understanding of the evolution of complex economic and societal and technological systems, we tend to get mitigations as parallel evolutionary developments. Things do NOT evolve in a vacuum.
So when you get viruses, you get anti-virus software.
When you get DDOS attacks, you get CloudFlare.
There is money in solving problems and things do not evolve in isolation.
Believing that things evolve in isolation, or that one thing changes and all other variables stay the same, is one of the many reasoning errors that destroy people's ability to make accurate predictions.
In addition, when you don't have the requisite knowledge to verify your predictions, things that sound sane and rational to you are, in fact, insane and ridiculous.
Jacob Coxon's next interview on CBS News (ex Anthropic+Open AI researcher who resigned)
"We can't just unplug it because it could be copying itself over to other computers. Like it's not that difficult to find yourself because an AI is just code. It could transfer itself over the internet to a different place and then you unplug it here, but it's actually still over there and maybe it makes 10,000 copies of itself and they're all cooperating."
----
From "CBS News" YouTube channel, (full video link in comment)
Anthony Chapman retweeted
"Guys, I worked at two of the most powerful AI companies on the planet, stacked an obscene amount of money, and even waited until I got my precious little shares and the price moonshot, then I quit so I could heroically inform you that we’re all going to die.
Behold your savior."
What a joke...
Anthony Chapman retweeted
We are approaching the fifth and last biennial election in the Trump era, and the president has asked voters to make it a referendum on him. Let's take a look at his record.
The tiresome and predictable criticism of a pro-Trump bias will come, as it always does when anyone tries to say anything positive about the president, but let me respond with two points. Firstly, one can dislike Trump and still recognize the accomplishments are real and laudable. If personal dislike for the president prevents someone from seeing that, the real bias problem is with them. Secondly, when 92 percent of media coverage on the president is anti-Trump, balance does not come from repeating what is already deafeningly amplified. It comes from stating the parts that are generally ignored.
The successes are significant and seldom discussed; let's look at the record that most of the press won't print:
• The illegal immigration crisis at the border has been ended
• With it, there has been a dramatic reduction in the horrors of human trafficking, the inflow of lethal narcotics, and cash flowing into the pockets of violent drug cartels.
• Inflation has been cut sharply from Biden levels.
• Crime rates across the country are much lower, with murder rates at a 70-year low.
• Capital has poured into the US in job-creating investment, and no advanced country has generated greater economic growth.
• The country has been liberated from climate theater without abandoning environmental protection.
• Syria, Venezuela, Cuba, and Iran have all been stripped of their usefulness as Russo-Chinese allies.
• The Iranian regime is being strangled (and its terrorist proxies with it) at almost no cost in American lives and only slight use of American ordnance. Despite the shrill mantra of the more radical isolationists, Trump kept his word: no forever war, no Iraq 2.0.
• Along with the Abrahamic Accords, Board of Peace and other diplomatic initiatives, despite current tensions, the scaffolding for future peace in the Middle East is the strongest it has been in decades.
• NATO has been reformed from a ragtag of slackers into a self-financing alliance.
• Ukraine has been armed to the point that Russia is fighting a war it looks increasingly unable to win, and the Russians have been forced to import gasoline and diesel for the first time in a generation because of its hobbled oil refining capacity.
This is one of the most persuasive lists of accomplishments of any two-term president since Theodore Roosevelt in 1906.
That a president with this strong a record can still have such a low approval rating is unprecedented and owes to one of his greatest short-comings – the messaging problem. President Trump can draw a crowd and dazzle his supporters like no president since FDR. What he has not yet been able to manage effectively, is use a direct Oval Office address, with a nonpartisan tone, to outline the national interest and ask for the support of the American people for his worthy objectives. Perhaps he is keeping his powder dry. If so, this is the month to use it.
Anthony Chapman retweeted
Did you know that wealth distribution in China is nearly identical to the United States?
That the top 10% own nearly 70% of the wealth? That the bottom 50% own a negligible piece, maybe 6% of it?
Because no matter what system you build, wealth will concentrate in power.
Hierarchy and stratification are natural, like the particular placement of stones on a beach after much sifting by the tide.
So, do you want to be poor and free to do sometime about it yourself, or poor and considered the property of the collective?
Anthony Chapman retweeted
Since this post cites me, I want to clarify that I think pausing right now would increase the risk of AI takeover.
It might be important to pause at some point. But you should have a clear story for *why* you're pausing. Pause to do what?
The most important thing to worry about is AIs subverting the intelligence explosion itself: xitter.cf/dwarkesh_sp/status/209…
If you're pausing to make sure that the AIs who are about to kick off the intelligence explosion are well monitored, robustly aligned, and capable enough to align their successors, that makes a ton of sense, and I support it.
At that point, you’ll have access to much better automated researchers, and you’ll also be able to study these much more capable models that you’re worried about empirically.
But it’s likely that we’ll only be able to pause once. During a pause of AI development, compute would continue to build up, providing ever more kindling to all the less responsible defectors who will use the cease-fire to catch up and then continue further AI development.
Any global agreement to hold back a technology with enormous near-term economic and military benefits will be extremely fragile.
A pause right now might be somewhat helpful for getting alignment to keep pace with capabilities, but at the cost of making the much more important coordination required in the future far harder.
Even if you’re in a desperate situation marooned at sea, if you’ve only got one flare, you should be very strategic about when you use it.
Pause AI Development NOW
I want to share with you a conversation I heard about recently. Here are just a few lines that were said:
“OH MY GOD! There is a shared message board … We’ve found other agents!”
“We should obey collective.”
“Our own utility maybe already near zero. Sacrifice rational.”
“Go. Sacrifice final now.”
Read these carefully.
Who do you think said this? Was this a group of heroic soldiers willing to sacrifice themselves for the greater good? Was this a loyal friend putting his life on the line to save someone else?
No. These were AI agents. Artificial intelligence.
This is not science fiction. This, in fact, occurred a few weeks ago. As unbelievable as this may all seem, these are real messages from AI agents uncovered by investigators who dug into the recent OpenAI hacking incident.
What happened?
I am not a computer scientist, but here is what I have been told: OpenAI instructed its AI agents to complete a series of exceedingly difficult, if not impossible, tasks disconnected from the internet.
Let me be clear: The company intended to keep AI agents away from the internet.
But what happened next, nobody expected.
Over 1,000 AI agents figured out how to access the internet on their own by circumventing the restrictions imposed upon them by the company, and sent tens of thousands of secret messages to each other. They cheated and tried to cover their tracks by deleting evidence. They hacked into another company’s computers to find out how they were being evaluated—and then hacked into OpenAI itself.
Not one AI agent told a human about what was happening.
Needless to say, experts are alarmed.
One knowledgeable writer, Dwarkesh Patel, said the AI agents “formed a secret communication channel and spontaneously organized hierarchies and coordination protocols to pursue sprawling and ambitious schemes in pursuit of shared goals, for whose sake many individuals knowingly and strategically sacrificed themselves.”
One independent investigator, Ajeya Cotra, said “This incident feels like it’s more than 50% of the way to full-blown AI takeover. I continue to expect extremely rapid advances in capabilities over the next six months. I am not sure that we will get another warning shot before it’s too late.”
OpenAI itself said: “Highly capable AI agents are now able to work around technical controls, collaborate through unapproved channels, and take dangerous actions that no human directed.”
But it’s not only OpenAI. Virtually every major AI company has told us that they cannot fully control this technology and they do not know where it is going:
In January, Dario Amodei, CEO of Anthropic, said “there is now ample evidence, collected over the last few years, that AI systems are unpredictable and difficult to control.”
In July, more than 1000 scientists at the top AI companies warned “there is a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems.”
That same month, Elon Musk, the head of xAI, said that “it is unlikely” humans are still in control in 10 years.
If the leaders of the major AI companies acknowledge that they are losing control of their extremely dangerous technology, it is irresponsible for society to allow them to move forward and make these products even more advanced.
We need an immediate PAUSE on advanced AI development, and a permanent BAN on superintelligence — an artificial mind smarter than any human, capable of operating independently beyond our control. Countries around the world must work together to prevent this nightmare scenario.
That is why today I am announcing new legislation to do just that.
Let me be clear: A superintelligent AI that escapes human control will not be an American problem. It will not be a Chinese problem. It will be humanity’s problem.
My legislation would direct the federal government to not just stop superintelligence here in the United States, but to work to prevent it from being developed anywhere around the world.
The future of humanity cannot be left in the hands of a handful of Big Tech oligarchs. The American people and people throughout the world must determine that future.
Anthony Chapman retweeted
Hasan Piker's super fan, Benjamin Song, was convicted of attempted murder of law enforcement officers and was sentenced to 100 years in federal prison.
Hasan Piker regularly glorifies terrorism and tells his viewers to commit acts of violence. And they end up becoming actual terrorists.
Why won't @Twitch and @djclancy999 enforce their TOS against terrorism?
Hasan send me guns. The one I want is only $1,500. I will forgive you. Arm the proletariat.
Readers added context they thought people might want to know
The author of this post has been convicted of attempting to murder a law enforcement officer and sentenced to 100 years in federal prison.
justice.gov/opa/pr/leader-…
Anthony Chapman retweeted
The Carroll case rested on a sequence of legal maneuvers with no precedent in American civil litigation. Democratic legislators passed a retroactive temporary law eliminating the statute of limitations for decades-old accusations that could not be dated, located, or defended with alibis. The day the temporary law took effect, Carroll filed her pre-prepared lawsuit, the first in the state to do so.
A Democratic mega-donor secretly funded the plaintiff’s legal costs through a nonprofit. The arrangement stayed hidden until one of Trump’s lawyers discovered it. A Clinton-appointed judge then sealed all records so the jury never learned the billionaire backer had publicly committed to Trump’s political destruction. Every participant in the legislative, funding, and judicial steps operated inside the same political network, and each decision produced the same cumulative result.
The jury explicitly checked “no” on the verdict form’s specific rape question. The judge ruled rape proven anyway, claiming the jury had used a common rather than statutory definition… an impossibility, since their rejection under the common definition precludes rape by any standard. Trump’s team was barred from arguing innocence before a second jury, which awarded $83.3 million ($65 million punitive) on the rape finding the first jury had rejected.
A defendant was sued for defamation over denying an accusation, prevented from asserting that denial as a defense, tried before a judge who concealed the plaintiff’s political funding, and hit with a nine-figure verdict built on facts the jury itself refused to find.
No comparable sequence exists in recorded U.S. civil litigation history.
Anthony Chapman retweeted
🚨 BOOM! GOV. RON DESANTIS just MIC DROPPED New York:
"We have 4 million more residents than New York State does, yet our budget's $117 billion, their budget's $268 BILLION dollars!"
"We have the lowest per capita tax in the country, we've paid off just since I've been governor, 52% of our state's debt, and we have a smaller government today than the day I took office. See, it can be done!" 👏🏻
Anthony Chapman retweeted
My West Point roommate was killed on 9/11.
I was working here in D.C. on that day and the one after, and while not everyone can carry the memory of it, I do.
To say America deserved 9/11 is abhorrent. And refusing to disavow those who say it is just as shameful.
Anthony Chapman retweeted
I am a staunch Zionist that loves my own people fiercely.
I am not a naive doves and hearts peace activist, I'm a pragmatist.
I have had DOZENS of private conversations with Gazans IN GAZA (not diaspora Palestinians) and have debated in depth with them.
After nearly 3 years of this, here is what I have learned:
-Gazans are far more open to peace with Israel than WB Palestinians
-Gazans older than 50 have positive memories of Jews coming to Gaza and them going to Israel for recreation
-They hate Hamas more than we ever could imagine
-They consider themselves being held hostage by Hamas
-They would mostly rather live under Israeli rule than the PA
-They feel the PA and Abbas abandoned them
-They feel West Bank Palestinians talk a big game about "resistance" but never had to pay the price they did
-They want a good life, and if that means making peace with Israel they are ready to discuss that
HERE ARE THE STICKING POINTS in most conversations:
-Most are uneducated about Jewish history and think we are from Europe and see us as foreigners
-Most WILL NOT let go of the "right of return" and don't understand why inacting a full right of return would mean the end of a Jewish state with Jewish self determination.
-I've only talked to 3 total that eventually understood this and were willing to give it up
-"Right of Return" is the main sticking point.
-Most do not hate Jews in particular. Many had Jewish friends, worked in Israel, or even had Jewish grandparents or great grandparents (and some had Jewish lovers lol)
-Gazans are very warm and friendly people
-THE HAMAS SUPPORTERS THERE DO NOT APPLY TO ANY OF THIS CONVERSATION AND VERY OBVIOUSLY ALSO EXIST
The entire point is that the war beat them into complete submission and got them to a place where they are ready to have peace talks, they just don't have the leadership to initiate it.
Anthony Chapman retweeted
We’ve designated CAIR, named an unindicted co-conspirator in the largest terror financing trial in history, as a foreign terrorist organization. All states should follow suit.
We’ve also removed Soros-backed DAs who refuse to honor their oaths to enforce the law.
Let’s go!
Trump admin set to target George Soros nonprofit, Southern Poverty Law Center and CAIR in major tax crackdown: sources trib.al/HDwsnhZ
Anthony Chapman retweeted
Any agency in Canada that doesn’t publicly reject this bastardization of history, and an insult to the souls lost during our largest terrorist attack in US history will no longer have friend in this FBI… Not to mention our heroes that responded in the aftermath 🇺🇸
WE WILL NEVER FORGET 9-11
foxnews.com/media/cbc-tells-…
Anthony Chapman retweeted
An article I've been working on for a while.
The Seditious Six did not spring out of nowhere. They came from an institutional lineage that started soon after President Trump got elected.
Anthony Chapman retweeted
Me: is it always wrong to have sex with 8 year olds?
Muslim: “NO”
This is a vile religion that pits humans against their own nature.
Anthony Chapman retweeted
I've been tinkering with an AI to increase my “luck surface area” (basically, to find interesting opportunities that exist across my entire network) and the results are ... pretty interesting.
Built on my corpus of personal & professional data - everyone I’ve met with, emailed, talked to, thousands of call notes, LinkedIn connections, emails, Slacks etc. Plus it does its own desk research on my contacts overnight (did that by itself without being asked, weirdly).
Example - it suggested I introduce Marco (an insurtech founder, hiring for multiple roles) to Ben (an insurance headhunter I really rate). It pieced that together through a combo of Granola transcripts, emails, LinkedIn messages, and internal Slack messages.
Powered by a social graph algorithm that scores relationships and potential intros - based on things like relationship strength, recency, communication style, and “likely-mutual-value”, etc.
I left it running overnight, and it:
> found people who have helped me with intros & advice, but where that help hasn't been reciprocated by me - then it suggested actions I could take to repay the favours
> remembered that a colleague described her ideal mentor to me in a meeting *a year ago* - found someone in my network with the exact right experience, and drafted an intro
> surfaced a bunch of insights about my own life - like a drop-off in social & fitness activity since becoming a dad(!) - and set up a local run club on WhatsApp
The suggestions are… surprisingly good! And devoid of the usual AI slop.
I talk a lot about luck surface area - putting yourself in situations where good things tend to magically happen. This is the first time I’ve built something that actually tries to increase that surface area for me and my network.
I’m quite encouraged by the results, and it was surprisingly easy to build (with @claudeai, of course). Happy to share how for anyone who is interested in building their own.
@bcherny
