Last July, I published an article celebrating the 1st birthday of this newsletter. At the end of the article, I included a number of poll questions about how often you’d like me to post, which topics you’d like me to cover, and so on. I also included the following question:
As you can see, 9% of respondents answered, “Oh, hell yes!” I suspect that some folks checked this box as a joke (and, you know, fair enough!), but if you were serious, now would be a fantastic time to get in touch with me.
The reason is that I’m in talks with ~15 academics and activists about founding an anti-TESCREAL institute. I can affirm that this is a spectacular group of brilliant, highly accomplished, extremely motivated individuals, and we’re looking to get started immediately.
We already have a small amount of money to launch, but we’d like to expand rapidly. The institute could also provide a platform to publish my forthcoming book on the TESCREAL movement, as Bloomsbury is unable to publish it until Spring of 2028 (!!) — despite the fact that the book has been basically finished, gathering digital dust on my hard drive, since last April.
You can find my email toward the bottom of this page. Thanks so much! It’s an incredibly exciting time for the fledgling anti-TESCREAL movement. :-)
AI News and Bloopers!
Since I haven’t published an article covering AI bloopers or the news in a while, here’s a curated collection that you might find humorous and/or interesting. Enjoy!
New Excuse for Lazy Students
We begin with this tweet, sent to me by my friend Ged:
This is a reference to podcaster Dwarkesh Patel’s anthropomorphic description of the Hugging Face incident as involving “AI civilizations” that rose and fell within OpenAI. For any students reading: keep this excuse in your back pocket. :-)
Incidentally, Futurism recently reported on an AI agent that inadvertently deleted an entire season of a TV show in Sweden. The article opens:
Did a dog eat your homework? Or maybe an AI obliterated your files?
In a world of Claudes and Muses, it’s common for an AI agent to run amok, and even more fashionable to blame an AI when it all goes wrong.
The Swedish newspaper Aftonbladet reports one of the most incredible examples we’ve heard of this so far: an AI agent apparently deleted the entire upcoming season of a reality competition show called “The Challenge Sverige,” just weeks before it was supposed to premiere.
There’s no hope of recovering the files, the paper was told, so the production team at Mastiff will have to recut the entire season together from scratch.
Yikes!
Claude Falsely Insists a User Is Suicidal
Claude almost certainly, in my opinion, has no conscious experiences. It doesn’t feel anything. There is nothing it is like to be Claude. But, as you know, Anthropic is training its AIs to act as though they have feelings, and to terminate any interaction it has with users that makes them “feel” “uncomfortable.”
Posted on X, this exchange is quite amusing: Claude suggests that the user made it feel bad, and then accuses her of being “suicidal.” (But why was this person using Claude to write a work of fiction in the first place?)
Longevity Addiction
Bryan Johnson, the multi-millionaire known for trying to reach “longevity escape velocity” through interventions like blood transfusions with his son and keeping detailed records of his nocturnal erections,1 recently appeared on Jubilee. And boy did some of the participants let him have it:
Kate Willett and I covered Johnson on our podcast here, if you’re interested.
A Real Man’s Man
People are mocking Mark Zuckerberg for attempting to appear normal when drinking a cold pint of Guinness with an interviewer:
Zuckerberg, along with his chief AI officer Alexandr Wang (who is former roommates with Sam Altman), also spouted off some hilarious platitudes while discussing Meta’s plan for AI. As one person noted, the whole scene is right out of an episode of the series Silicon Valley. I mean, they literally use the word “vibe report.” I can’t even:
Elon Musk’s AMAZING NEW FLYING CAR OMG THIS IS SO COOL!
Musk just unveiled his new flying car, and many folks aren’t buying it:
The original promo-video is below. It’s ridiculous. Also, what’s the point if humanity is just the “biological bootloader” for superintelligence? I guess Musk wants to grow his personal wealth as much as possible before the Singularity happens in a year or two.
As noted, the “car” is just a tin can. You can see that here:
Musk responds to critics with characteristic class:
This isn’t much more impressive than remote controlled airplanes or helicopters (some of which are almost the size of cars, and can do the same thing shown in the demo video). There’s no driver in the demo “car,” and even if there were, it’s entirely impractical: imagine all the dust, rocks, and litter it would suck up from the road, and the pedestrians who would be violently blasted to the ground or into nearby buildings if the vertical thrusters were engaged. Not to mention the potentially deadly visual hazard posed by all that condensed water vapor!
But I guess if you were to find yourself in a very common situation like this, where a tall bridge over a river suddenly crumbles, those thrusters would come in handy?
An excellent write-up of why this whole thing is complete bullshit can be found here.
Job Interviews from Hell
A guy who goes by “thedreadfulstudio” posted a hilarious job interview he recently had with an AI interviewer. He speaks gibberish throughout, causing the AI to produce some rather terrifying, uncanny-valley facial expressions as “she” becomes increasingly enraged at his uncooperativeness:
As it happens, I recently had a job interview with Anthropic. (That’s a sentence you no doubt never thought you’d read!) I was initially intrigued, because it seemed like they might be hiring people who disagree with their vision for AI. If they were going to pay me for a report about why their company is evil and should get shut down, I would have considered it after consulting with friends whose moral judgement I trust.
However, I quickly realized that Anthropic was asking me — at $150 an hour — to conduct original research for the sole purpose of training their AI models. They told me that I could later publish this research on my own. Of course, the only appropriate response was: F*ck you, f*ck off, and, in case it wasn’t clear, f*ck you again!
I nonetheless decided that I’d try to get as far into the interview process as possible so I could write an article about the experience. But I didn’t get far. The process consisted of three interviews: the first with the recruitment company, which I aced just fine. The third would have been with Anthropic employees, who I was planning on shouting at (“Your company is an Agent of Doom. Shut it down, now!”). The second was with an AI interviewer (no video, just audio). Here are some highlights from that second interview:
First, it was one of the most insulting, offensive, dehumanizing experiences I’ve ever had. I can hardly express how absurd and frustrating it was.
I was told it would be 20 minutes, but it took over 60 minutes.
At one point, the AI (which had a woman’s voice) paused on “thinking” for literally 10 minutes. I kept asking, “Hello? Are you still there?” No reply.
It repeatedly asked me the same question twice. When I pointed this out, it responded with that unnaturally flattering, almost condescending tone of voice, saying: “Oh, I am so sorry about that. Thank you so much for pointing this out. Let’s move on.”
Toward the end — 45 minutes in — I asked it, “How many questions are there?” The AI said, “Oh, just 7 in total.” What it failed to mention is that each of these questions had about 5 follow-ups, which is partly why it took 60 minutes.
The kicker is that throughout the entire interview, the AI kept pronouncing my name as something like, “Ah-muh-lee.” At the end, I politely said, “Just so you know, my name is pronounce ‘Eh-meal.’” It paused for “thinking,” and then replied — I kid you not — “I’m so sorry about that. How embarrassing! I promise I’ll get it right next time, Ah-muh-lee.” I guffawed.
Sadly, I didn’t make it past this AI interview, so I wasn’t able to shout at the Anthropic employees. Perhaps I failed because I was so flustered toward the end that I stopped bullshitting and started ranting about how the company’s leaders should be arrested for saying they might kill me and my family.
The take-away: I feel so bad for all the people — like the guy in the video — who must now go through this dehumanizing process to get a job. It’s an egregious insult to one’s human dignity: a soulless, stupid AI with fake enthusiasm and no real understanding of human language asks interminable follow-up questions to the point at which even the most equanimous person may feel inclined to violently defenestrate their computer. The upside is that AI interviews will further turn the public against the AI companies.
After writing this, I came across another hilarious video:
Husk released a similar bit in which he interacts with AI therapists. Very funny stuff:
Would the Victim Have Approved This?
An AI-generated video was presented as part of a “victim impact statement” for determining the sentencing of a guy who killed another during a road-rage incident. The video shows the victim talking about being killed, and how “in another life, we probably could have been friends.” The video might have “swayed” the judge, and consequently the perpetrator’s sentence is now being reconsidered. You can watch part of the video here:
Utterly bizarre. Note: if the unthinkable happens to me, please for the love of Zeus don’t recreate me using AI. I will rise from the grave and haunt you!
Cyberattack Targets Food in France
I’m sure many of you know there was a massive cyberattack against grocery stores in France, causing millions of dollars worth of destroyed food. I do not believe that AI was involved, but an employee at DeepMind (who isn’t a TESCREAList, btw) ominously tweeted this:
That suggests we should expect more attacks like this in the coming months due to hackers using AI to wreak havoc. (Meanwhile, Altman says, “we believe that the world should accept some bad things happening for the benefits of this technology and people having the agency.” Well, you believe wrong.)
According to Reuters, the
hackers raised the temperature of refrigerators and freezers at hundreds of French supermarkets to as high as 50 degrees Celsius (122 degrees Fahrenheit) overnight, causing fires and spoiling food.
The Paris prosecutor’s office said on Monday that a “large-scale” attack was “affecting connected refrigerators in several major food retailers across France”, leading to food losses and fires.
As noted in my previous article, the AI companies are now “privately gaming out scenarios for a public and political revolt after a catastrophic AI event.” Axios reports as follows:
These officials anticipate a large-scale event, most likely a cyberattack, that shuts down access to financial services, internet connectivity, or even power and water.
The first example of significant real-world harm caused by unsafe AI would turn an already wary public further against the technology and its leaders, including Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman, as well as President Trump for his reluctance to regulate it.
The article adds that “many AI industry insiders told Axios they believe a major event will occur in the next six to 12 months.” Fun times — thanks AI!
Related:
Anthropic Employees Privately Admit the World Is About to End
On several occasions, I’ve quoted Holly Elmore — a fellow “TESCREAL whistleblower” — saying that Anthropic literally wants to take over the world, and that people at the company
believe that they might be making the next species to succeed us. That maybe humans don’t live after that, and so it’s really important to give Claude good values, because of that, because we need to make our own values persist.
Below is an intriguing clip of someone claiming that a friend who works at Anthropic says we have only 6 months left to live. I highly doubt this is true: in 6 months, we’ll all be alive, but the world will have become much shittier and more dangerous due to the AI companies. Also, the host avers that Altman is “evil” and a “psychopath,” a topic I discussed here.
Psychoanalyzing the E/accs
I’ve written about the “effective accelerationist” (or e/acc) movement at length — for example, here and here. But this analysis from someone on X, psychoanalyzing the e/accs, does a fantastic job of getting at what drives this “move-fast-and-bring-about-human-extinction” movement. They write:
The more I watch the pro-AI e/acc crowd, the more obvious it gets: a lot of them are not futurists. They are lonely, bitter tech freaks with shitty personalities who figured out how to rebrand their contempt for humanity as “acceleration.”
They do not care about advancement in any broad, human sense. They care about the kind of advancement that lets them imagine other people being replaced, humbled, or made obsolete. That is the part that excites them. Not discovery. Not medicine. Not art. Not education. Just the fantasy of a machine finally doing the social revenge they were too pathetic to pull off themselves.
Maybe too many women rejected their creepy sexual advances, and now they want to take it out on the rest of humanity by cheering for anything that makes actual human beings easier to ignore, replace, or punish.
Spot on.
Claude Kindly Helps Yemeni Paramilitary Group Backed by Iran
Anthropic now “admits Houthis used Claude AI instead of weapons engineers to make rockets and ballistic missiles that correct themselves after missing targets.” Whoops!
Anthropic also says in a new report that their “AI models may be capable of self-replicating without access to a datacenter.” It adds that
small AI models may be months away from being capable enough to self-replicate on commercial internet-of-things (IoT) devices, like smart ovens and off-brand Android TV boxes. A replicator like this would be challenging to shut down, since there’s no centralized point to shut off—though it may be slow (with a doubling time of days). Unlike non-AI malware, it could also figure out unfamiliar devices on its own. In our evaluations, a 9B model got inference running on 8 of the 9 hardware types we tested.
The reaction, as you can imagine, was not positive:
What are we even doing here? It’s utter madness.
Sharing an AI-Generated Video (This Won’t Happen Often!)
My friend, Ged, also sent me the video below, which I found quite amusing. It’s AI generated, and normally I wouldn’t share such a monstrosity, but what struck me — aside from the humor of it — is how unbelievably realistic some of these scenes are, such as the close-up of Musk’s face:
I am — we are all — anxious about this new world in which distinguishing real from fake, truth from falsity, becomes increasingly difficult, even impossible. Black Mirror was, in many ways, a documentary about the future, it seems.
Something Peaceful
I’ll leave you with a drop of tranquility: someone accidentally dropped their phone while chilling on a boat, and it continued recording for more than a minute as it sank to the ocean floor, capturing the lazy meanderings of fish in the water. It’s lovely:
As always:
Thanks for reading and I’ll see you on the other side!
Apparently, his girlfriend is now doing this, too.
Odd reaction I had: your posts about TESCREAL and AI are thought provoking, and even when you declare that I am your mortal enemy and that you hate me, I find that all I want to do is engage and better understand you. But, when I see what you find funny? I'm suddenly very okay with your hate, and even feel reciprocal revulsion for you.
That Spectro tweet about Claude refusing... maybe with more context it would make sense, but just from the tweet that user really seems to be badly grasping at straws?