The Zuckster and His Tent Erections
June 12, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
Construction is a hassle. Build a traditional data center, and it takes time. Time means people can organize, protest, and become creatures who do not get with the program. Here’s the problem: How can one stand up a data center faster than filing for a zoning permit?
The article “Meta Putting Up Tents Across the US to House AI servers, Like ‘a Scene Out of the Movie Mad Max’ — Structures Take Three Months to Build and Use Jet Engines for Power” provides a solution:
Meta has moved from building traditional structures for its data centers to putting up tents across the U.S. and sticking AI servers inside them.
Mark is a move fast and break things type of guy. Some tents cover 125,000 square feet. Once permits are filed, the tent-like structures are ready for their hardware. The write up points out that Elon Musk allegedly built a data center in 19 days. The original zuck-it dude, wants a fleet of these puppies in three weeks. Cowards like Google[type firms can waste up to four years. The speedsters named Elon and Mark have a better idea.
What body part responds to 30 cycles per second? This family learns about resonant frequencies when the data center turbines fire up. Thanks, MidJourney. Good enough and semi-tasteful.
How does one power 300,000 square feet of data centers? If you said, “Plug it in,” you are behind the times. The approach is “behind the meter”. The article reports:
Another factor that allowed Meta to bring its data centers online at a much faster pace is its use of “behind-the-meter” power, in which the company installed its own turbines to produce power on-site rather than relying on grid power. This is similar to what Musk did with his Memphis Supercluster, which he initially powered with portable power generators. However, Meta’s turbines would be a permanent feature on the Ohio site, as it’s designed to run independently of the power grid.
I like the inclusion that the zucked up approach is compared to a scene in a Mad Max film. That’s a good analogy, but I think it fails to capture the thrill of living near the turbines buzzing and humming 24×7. If you paid attention in physics, you may recall that certain frequencies are not audible to the human ear, but certain body parts can resonate to these frequencies. One of the more interesting body parts affected relates to bodily functions. In some conditions, the consequences can be interesting and often unexpected. The brown note is four to five hertz, the sweet spot for intestinal resonance.
Meta’s PR professionals struggle to understand why the Zuck’s public image is not as sparkling as the Zuck would like. With the termination methods, the yachts in city harbors, and the tents plus turbines, I too wonder why more people don’t recognize the Zuck as the master of the universe, digital emperor, and creator of chaos that he is.
Net net: Zuck him.
Stephen E Arnold, June 12, 2026
Smart Software, DRGs, Treatment Chains, and Money. Yep, Money.
June 12, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
I read “Inside the Accountability Vacuum: Why Clinical AI Errors Have No Owner.” This is an “inside baseball” type of write up. However, if you are into medical fraud innovations, you may find the article food for thought. I want to hit the highlights of the report from my dinobaby perspective, of course. Then I want to outline how the baked in methods of a medical software might allow someone to put billing before patient well being.
The main point of the story is that when an AI module does something as part of the diagnostic process, it may be difficult to pinpoint why something went off the rails. One example: An X-ray interpretation of granny’s lungs output a probability that suggested she had a problem and provided probabilities for treatments. Ooops. Granny died after a series of “actions” were implemented by humans relying on the probabilistic outputs of the smart software in the diagnostic and treatment chain.
Good enough, Midjourney. Hey, do you output DRG treatment chains too?
Yeah, it happens. And, because I am a dinobaby, I think it will happen more often.
The cited article says:
Clinical AI is no longer speculative, no longer the next thing, no longer a topic for a panel discussion at a digital health conference. It is already embedded in care. The question is no longer whether it will arrive but whether the institutions that deploy it can evaluate it honestly once it has.
The long article asserts:
By the early months of 2026, the United States Food and Drug Administration had authorized more than 1,350 AI-enabled medical devices, roughly double the figure from 2022. The technology is propagating into clinical workflows on three continents simultaneously, and the institutions tasked with policing it are still drafting the rulebook in public.
After a romp through some baseball type information, the article reports:
The Stanford-Harvard report’s central anxiety is not that clinical AI is bad. It is that nobody yet knows how to tell when it is….A model that performs flawlessly at one teaching hospital can quietly degrade at a community hospital ten miles away because the patient population is different, the equipment is older, or the implementation team configured the alert thresholds in a slightly different way.
I have a healthy skepticism for information from both of these estimable certifying institutions. However, I do want to mention that both outfits have been linked with made up information. Yep, research fraud is everywhere folks.
Two examples of smart software fancy dancing appear in the cited article. The first is the old chestnut about Watson’s cancer foibles. The other is from Epic Systems. I think the anecdote is indeed epic. I quote:
The Epic Systems sepsis prediction model is the more instructive one. Documented in a series of investigations published from 2021 onwards, the Epic Sepsis Model had been deployed across hundreds of American hospitals when an independent external validation by researchers at the University of Michigan, including the work of Karandeep Singh, found that the model missed sixty-seven percent of sepsis cases and that eighty-eight percent of its alerts were false positives. Epic had claimed accuracy of between seventy-six and eighty-three percent. The independent figure was closer to sixty-three. What made the Epic story matter was less the performance gap than the institutional dynamics it revealed. Hospitals had bought a tool, in some cases under financial incentives that included payments of up to a million dollars to use the algorithm, without seeing an external validation study. Clinicians had spent months responding to alerts that turned out to be wrong most of the time, building up the very automation fatigue that ECRI now warns about. By October 2022, Epic had overhauled the model and was recommending that hospitals retrain it on their own patient data before clinical use, which is itself an admission that the original product was not fit for the purpose for which it had been sold.
The cited article tiptoes into the question of killing granny as malpractice. I quote:
Talk to a medical malpractice plaintiff’s lawyer about AI cases, and the conversation eventually arrives at a particular kind of frustration: the audit trail that does not exist. … When AI sits in the chain of decisions, that reconstructibility starts to break down. The first is technical: many of the models in clinical use, particularly those based on deep neural networks, do not produce outputs whose reasoning can be inspected after the fact in any meaningful sense. There is no chart of inferences. The model produced a probability, and the probability turned into a flag, and the flag turned into a recommendation, and the recommendation either was or was not heeded.
I am certainly no medical professional. I am definitely not a legal eagle. I am 82, and I view health care with some skepticism. My mother told me when I was in college, “Stephen, you know you could have lost your leg when you had osteomyelitis.” Okay, that was a surprise. I still have two legs and both seem to work okay.
I do have some observations:
- Smart software embedded in “diagnostic workflows” can be incorrect. Busy people may miss the bad outputs. People can lose their leg. (See above.)
- The outputs from embedded software may recommend high payoff treatments; that is, the incentive to fiddle with data can translate into more money.
- Overworked or under trained staff can just accept what the smart software outputs. Fatigue, problems at home, doom scrolling, whatever can distract the health care human. Granny? Hasta la vista.
- Regulations in the US and elsewhere are not in step with smart software applied to health care. When a fraud or treatment issue arises, the rules output will, by definition, be reactive and inapplicable to those who create smart systems and are quite adaptable to guard rails.
Net net: When did IBM Watson make its capabilities known to cancer docs in Houston? I think it was 2012 or 2013. That was more than a decade ago. That tells you something about regulations, guard rails, oversight, and medical fraud opportunities.
Stephen E Arnold, June 12, 2026
Surprise! Bad Actors Use AI
June 12, 2026
Bad actors are usually the first to learn how to manipulate AI for nefarious means. The trouble is learning how to catch them at their own game before they commit horrible crimes. OpenAI wrote a report about, “Disrupting Malicious Uses Of AI” and here’s what they summarize about the latest findings:
“In the two years since we began publishing these threat reports, we have gained important insights into the ways threat actors attempt to abuse AI models. In particular, the case studies in this report, as in our earlier reports, illustrate how threat actors typically use AI in combination with other, more traditional tools such as websites and social media accounts. Threat activity is seldom limited to one platform; as our report on a Chinese influence operator shows, it is not always limited to one AI model. Rather, threat actors may use different AI models at various points in their operational workflow. We share these insights in our threat reports so that our industry, and wider society, can be better placed to identify and avoid such threats.”
Here’s the scenario we have to remember: all of these situations are a game of cat and mouse. The bad actors are the mice raining havoc down less-tech savvy individuals, while the white hat actors are the cats trying to catch them with similar and sometimes better resources, i.e. government funding. The problems are that mice multiply faster than cats and also learn how to adapt quicker. They don’t rely on the same tricks and tools, plus they operate in countries with lax laws or they’re smart enough to cover their tracks.
A reminder: Most bad actors act without too much bureaucratic friction. Defense can be tiresome tactic. Offense is, for some bad actors, more fun and with AI-enabled smarts, lucrative. Employment phishing and malware delivery uses different systems to allow agents to launch attacks at anyone whose names can be scraped or snagged from a Dark Web store front.
Whitney Grace, June 12, 2026
AI Does Mathy Stuff and Mathematicians Are Annoyed
June 11, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
Believe it or not: I had a friend. His name was James Terwilliger, Ph.D. He studied mathy stuff. He worked at Halliburton Nuclear with me until a tragic car accident claimed his life. Nevertheless, I remember fondly his standing up in the company cafeteria and saying, “This is Stephen E Arnold. He has a degree in medieval poetry.” The nuclear engineers in the cafeteria roared and hooted. Then, Terdwilliger (my name for him) announced: “Stephen will now recite a poem.” I stood thanked Dr. Terdwilliger and recited William Carlos Williams’ “The Red Wheelbarrow.” It was one sentence and drew a number of nuclear engineering and math wizard responses. Terdwilliger’s took the prize for remarks. He said, “I never read a poem. That’s why.” Halliburton Nuclear employed at that time about 600 nuclear engineers, nuclear physicists, mathematicians, and one person who had a degree in poetry.
Thanks, Midjourney. Good enough.
When I read “Over 150 Mathematicians Warn Governments Not to “Believe the Hype” About AI,” I thought about Terdwilliger. His mindset, as I recall, was different from mine. We had similar interests and our view of humor was congruent. The article could have been assembled by talking to Terdwilliger’s AI simulacrum or other real mathematicians.
I noted this passage:
"There is currently a strong commercial incentive on the part of the technology industry to overstate the capabilities of their products." Terdwilliger would have said, marketers cause more problems than they solve.
The write up adds:
In perhaps the strongest public rebuke yet, a new declaration signed by over 150 mathematics experts from around the world warned governments not to “believe the hype” when it comes to AI’s capabilities to solve complex mathematical problems, throwing cold water on claims of a revolution in the field. In a statement accompanying the 11-page “Leiden Declaration on AI and Mathematics,” International Mathematical Union vice president Ulrike Tillmann argued that AI “raises questions that cannot be left unexamined.” “The future of mathematical research must be guided by human judgment, fair and transparent practices, and the shared values of the global mathematical community,” Tillmann said.
Terdwilliger and other professional mathematicians I have known like Dr. Zbigniew Michaelewicz among others love problems. The whole point is figuring out how to solve, prove, disprove, or mix and match how their brains work to come up with a solution, a proof, or a negative example. Letting a machine solve a problem is probably not what would inspire confidence in the outputs. The mathy pros like to scribble on whiteboards. Dr. Julian Steyn would often mumble something and Terdwilliger would say, “It’s obvious that…” This is the human part of mathy types. A probabilistic collection of algorithms is only of interest when an equation is a problem. The outputs are definitely not going to get hugs and kisses from the whiteboard types.
Here’s another statement presented in the Futurism article warranting a look:
“Current automated techniques can produce plausible but unreliable (or even incorrect) arguments which are difficult to distinguish from correct mathematical proofs,” said signee and University of Oxford head of computer science Leslie Ann Goldberg in a statement. “This is a serious problem: research in mathematics (and in mathematical disciplines like theoretical Computer Science) almost always builds on previous research, so it is essential for researchers to know that the results in the literature are correct.”
The “statement” is a document signed by a set of mathy types. It is worth reading.
The conclusion to the Futurism article is, and I quote:
In short, it’s a ringing denunciation of the persistent hype surrounding AI, and a call for reining in its use that reverberates far beyond the world of mathematics. The broader scientific community has been reeling from a flood of papers that make heavy use of AI, risking contaminating the peer-review process with hallucinations. It’s also a pertinent reminder that AI models are being trained on cutting-edge research, often without sign-off from the original authors. “Mathematicians who never intended to contribute to AI development are having their work used for this purpose without their consent,” Leiden University anthropologist of AI Rodrigo Ochigame, who helped draft the declaration, told Scientific American. “I think that’s a deeply concerning situation.”
Terdwilliger had a great sense of humor. He liked whiteboards. However, he preferred old fashioned chalk and some weird, thick, crumbly chalk. He said, “I can’t do anything without chalk or a dry erase marker.”
Net net: I think the mathy types have output a truth the BAIT (big AI tech) tribe have yet to recognize.
Stephen E Arnold, June 11, 2026
AI Kung Fu: Better, Faster, Cheaper
June 11, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
In the last few weeks, I’ve noticed some product announcements about speeding up smart software.
On item was “This PCIe AI Accelerator Card Can Run 700B LLMs Locally With 384 GB Memory at Just 240W, Less Than Half The Power of RTX PRO 6000 Blackwell.” Is the “card” real? I have no idea, but the write up makes it clear that Taiwan’s Skymizer is thinking about add-ons that might change the math for some AI use cases… it it works.
I also spotted “New Device Could Make Processors Run 1,000 Times Faster without Additional Waste Heat — Scientists Say It Could Reduce Data Center Energy Demands.” Less heat means lower costs. Good news if one is arm wrestling chip centric AI systems.
Thanks, Midjourney. Four tries to get the word “cheaper” spelled correctly. Not even good enough.
Two examples, one from Japan and one from Taiwan. I assume other silicon whiz kids are beavering away in other shops around the world. I am not sure when, but I think there will be some useful hardware and software that can benefit AI applications and reduce costs… sometime.
VentureBeat, one of my favorite sources for content marketing juiced novelty, has offered another example of this push to make AI better, faster, and in some ways cheaper. “MiniMax-M3 Debuts, Eclipsing GPT-5.5 and Gemini 3.1 Pro on Key Benchmark Performance for Just 5-10% of the Cost” says:
The company’s leadership also announced plans to deliver the model under an open source license including “open weights,” allowing for full enterprise downloading and customizability free-of-charge, coming sometime in the next 10 days. For now, it is available via the MiniMax API at a special discounted price of $0.3 per 1 million input tokens and $1.20 per million output tokens (on fresh cache) for the next week — beating proprietary U.S. giants like Google, OpenAI and Anthropic handily on cost, while also eclipsing the performance of the latest models from the former two on selected benchmarks.
As US vendors push up fees, this Chinese outfits uses the “does more with less” angle and then slams the front part of this digital iron maiden. Oooof. The write up includes some razzle dazzle about how the costs are kept low. No problem, particularly if a company is Chinese and just maybe linked to China’s funding of mission-critical projects.
The write up also trots out some benchmarks that “show” how fast the software is. But the whipped cream on the marketing cup cake is that the MiniMax M3 includes “agentic capabilities.” Here’s an explanation from the VentureBeat write up, and I will leave it to you to deconstruct this chunk of prose:
The system relies on a “Producer + Verifier” adversarial harness loop. As one agent instance generates code, a secondary verifier instance aggressively tests and reflects upon execution outputs, allowing the network to self-correct and operate autonomously for days without human oversight. Because of its native visual grounding, MiniMax Code supports direct computer use. A developer can issue a cross-application voice prompt via their phone to have the model open a localized enterprise ERP client and batch-populate data tables directly from an open Excel spreadsheet. For custom setups, developers can pipeline M3 directly into existing workflows using an API key (
sk-cp) compatible with common alternative IDE environments like Claude Code, Cursor, Roo Code, and Cline. The API introduces a toggleable “thinking mode”.
My view of this announcement is that the Chinese firm wants to make clear that Chinese firms innovating in China are doing what China does: Better engineering.
Several observations are warranted:
- The AI battles are changing. Instead of the slow, ponderous tank-like systems, China is creating more agile solutions. This is similar to Russia relying on tanks and Ukraine recycling hobbyist drones to blow up pipelines. It’s a new era of warfare in AI too.
- The combination of the baby steps with speed ups and modern AI systems from China means that big, quickly out dated data centers might be the anchor than sinks the Silicon Valley super athletes in the long distance swim. Outmoded or not, those big expensive data centers have to paid for by someone.
- The “cheaper” angle is like one of those Bruce Lee one hand strikes. Sometimes they work.
Net net: With incremental improvements coming, the US approach particularly with massive training systems and even more elephantine data centers, one must beware of the competitor with a T shirt that says, “Better, faster, cheaper.”
Stephen E Arnold, June 11, 2026
Learning Is Hard. AI Is Easy. 6-7, Folks
June 11, 2026
AI is prone to hallucination which is another term for taking digital drugs. AI is a prediction system, happily regurgitating what has been converted to vectors. If the algorithms are not smart enough, whiz kids will write software wrappers to fix up problems that arise. Sort of.
In today’s zip-zip world what is a college student supposed to do? Read books? Use an old-fashioned card catalog? Let AI do the work and hit the local pub?
Aaron Tay wrote “AI Academic Search And The Missing Middle Of Literature Discovery.” Librarians are unimpressed with AI search and rightly so, because it removed the sources from the search. Here’s what Tay found in his own query:
“In my own testing particularly for difficult queries, the difference is often visible in the first screen of results: the better AI search tools place relevant work high in the ranking, while conventional databases often require more query reformulation and screening before comparable papers appear. For example, I recently ran a study on reproducibility, running the same query five times across multiple AI search engines. It was not even meant to compare AI search engines relevancy but even with a broad query that had 200+ possible relevant results, only Undermind maintained high precision down to rank 50.
He added:
Other respected AI search tools could barely manage this for the top 10 to 20, showing the variance in performance. For “hard queries” the difference was even more stark between AI search tools and conventional databases.”
Tay also summarizes more of his findings:
“Freshman undergraduate work: mid-to-high precision, recall does not need to be high. Almost any database with the right coverage will do.
Evidence synthesis: high recall is non-negotiable, and you accept whatever precision you can get. Multiple sources, documented strategies, traditional databases plus maybe AI tools as supplements.
Narrative reviews: moderate recall for the time spent, high early precision, and strong top-ranked relevance. This is where the better AI search tools sit, and it is the use case where librarian voices are thinnest.”
In other words, perhaps there is something missing AI-infused search? I wonder if the old-fashioned methods will return to favor for students who want to learn? AI strikes me as a complementary tool to standard information literacy.
His findings verify that a mixture of the old and the new is the royal road to information literacy. That’s easy to say. Even some “smart” students struggle to read. Or, as GenZ says, “Learning is high key stressing.” Okay.
Whitney Grace, June 11, 2026
Telegram Note: MTONGA and Buy a Gram to Support Telegram
June 10, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
Pavel Durov has been taking leadership-type actions. He pulled crypto responsibility from the TON Foundation. The TONcoin has been rebranded at GRAMcoin. This is the original name of the firm’s push to crypto in 2019. The self-styled GOAT of Russian entrepreneurship has plans to make TON great again. Yep, MTONGA. The name reminded me of a song my mother loved. The name also inspired me and my team to write “MTONGA: Buy a Gram and Support Telegram.” I assume there will a baseball cap with MTONGA proudly displayed, T-shirts (form fitting, of course), and probably a yoga mat. Mr. Durov is an aficionado of yoga and other physical activities.
A now-removed social media post on a Russian service suggested that the name “gram” was a nod to the company Telegram Mr. Durov set up in 2013. A wag noted that it was a common term among some in Moscow for a controlled substance. The problem with use of Russian language information sources on Russian services is that content can disappear. True or false, I found the name GRAM catchier than TONcoin.
If you want a rundown of what changes Mr. Durov has planned, this article in our Telegram Notes blog will hit some highpoints. We will address the AI features after we poke around making agentic functions work. In September 2026, I have been asked to illustrate how the new AI agent performs tasks previously done by mostly human hands. We also will address in a separate post the new push by Telegram into video. Video is on the platform, but now it is a priority and not just for those with an interest in sensitive subjects in a Telegram Private Channel.
The new post also contains a line from the tune “Bongo, Bongo, Bongo” my mother found enjoyable. One word in that snippet is “bungle.”
We also include four questions that are ones we are currently researching. The article is dated June 3, 2026, but that is a function of the content management system’s penchant for requiring manual entry of dates multiple times. The story went live on Tuesday, June 9, 2026. When we have a date for the French trial, we will provide that information once the wheels of French justice have rolled around to a date. Going slow has definitely destabilized Telegram. Now Apple, another AI laggard, has pulled ahead of the tardy Telegram engineers.
Bongo, bongo, bongo, bungle for us too.
Stephen E Arnold, June 10, 2026
Google and Small EU Publishers: Who Has the Cash for a Long Litigation?
June 10, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
I learned from “European Publishers Seek £552m+ from Google Claiming Ad Market Abuse.” I had to read the headline twice to make sure I understood that publishers believe Google abused the ad market. Now the publishers want money.
Thanks, MidJourney. Good enough.
The write up reported:
More than 20 European news publishers are taking legal action against Google seeking damages of £550m for adtech monopoly abuses.
That’s it. Google is a monopoly. The monopoly makes life tough for non monopolies. Google has been running its advertising railroad for a couple of decades. Now European publishers are charging into action.
The case is being funded by Prague-based litigation funder LitFin, which will cover the costs even if it fails. The publishers involved have agreed to share part of any awarded damages with it if they win. LitFin chief operating officer Matej Pardo said: “Google’s abuse of its position across the ad tech stack has been found unlawful at the highest levels – now it’s time for the publishers who bore the cost of that conduct to be made whole.
Speaking practically, if Google litigates this case until the cows come home and then appeals until the cows come home, who has enough money to pay the lawyers to herd the cows? [a] The publishers and LitFin or [b] Google. If you picked [a], you might want to think about how quickly European outfits cash those Google checks. I would convert this to another exam question, but I think you know the answer.
Several observations are warranted:
- Google is the big dog of online advertising. Meta is doing its best to dethrone the money printing machine. Good luck with that, Mr. Zuckerberg. Maybe your AI can pull this off, but not overnight.
- Google has lots of invisible outfits doing SEO and then telling their clients, “Just buy ads to get traffic.” AI will make this more of an imperative. Want eyeballs? Pay Google. DuckDuckGo may kill Google. Yandex may kill Google. Swisscows may kill Google. But for now, Googzilla looks healthy because traffic…
- Google views governments as a bit of proof that some people get it and others don’t. The EU and the small publishers don’t get it. I hope they win, but even a dying Google knows they won’t get it.
Net net: Publishers are taking on an entity that is larger than most of the countries in the EU in terms of money, value, and power. My hunch is that Google leadership will have Gemini output some PR pablum and then let the Google lawyers burn through whatever cash the publishers and the Prague outfit can muster.
Stephen E Arnold, June 10. 2026
Modern Management Methods: The 30 Minute Meta Exemption
June 10, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
If I were younger, I would catalog some of the Silicon Valley and Stanford MBA precepts into a book called “Modern Management Follies.” I am old. I really don’t have the energy.
However, I can muster up the strength to react to the information in the BBC write up “Meta Workers Can Opt Out of Being Tracked at Work – But Only for Half an Hour at a Time.” I like the half hour angle. I recall a time when I was a salaried employee with some responsibility. I also suffered a kidney stone attack. I left my desk, went out in the corridor, and sat down. I then writhed and walked to the cafeteria. I sat down. I writhed some more. I walked out the front door which, believe it or not, provided a grassy area with trees. I plopped under a tree in the shade and writhed. This continued for about 45 minutes when the excitement of a kidney stone went away. I went back to my office and noted that I had been indisposed. No one cared.
Romans had to row the boat. When the praetorian said, “30 minutes,” he meant 30 minutes. Thanks, Midjourney. Four tries and Photoshop. Good enough, sport.
If I were at Meta, I would have been in violation of the Meta exemption: 45 minutes without working.
The write up says:
new controls will allow employees to pause the data collection for “up to 30 minutes at a time” as well as request exemptions from the initiative altogether.
Would I have been terminated? I was an officer of the company, but I just violated a Meta exemption.
The article reports:
…one Meta employee, who asked not to be identified, telling the BBC that having their actions train AI models felt “very dystopian” – as workers expected a slew of additional job cuts.
Why does Meta need the “meta exemption” aka Model Capability Initiative or MCI? Here’s the answer, according the the BBC:
Meta told the BBC: “If we’re building agents to help people complete everyday tasks using computers, our models need real examples of how people actually use them.”
Several observations are warranted because I am an old dinobaby:
- The 30 minute window seems arbitrary
- Workers at home or in the office can experience what I call “stuff happens” moments. A quick break became a 45 minute unexplained absence. No human cared what I did. But if I were plugged into the Meta matrix, something would notice. I have considered operating a taco van but were I a Meta professional, that career would have been forced upon me.
- The Silicon Valley approach to work may not be a positive for some employees’ mental health, their physical well being, their families, and their work itself.
Those 30 Meta minutes, like the consequences of moving fast and breaking things like employee spirit, are because of the values of the “leadership.” Since there is one leader at Meta, what’s that say about a modern captain of industry?
Stephen E Arnold, June 10, 2026
Pay to Do Anything Online Is Coming
June 10, 2026
Social media as we know it may change forever to a pay to use model. It’s being touted as “pay-to-engage” by Techradar: “Meta’s Subscription Plans Are The Tip Of A Terrible Pay-To-Engage Iceberg And May Be The Beginning Of The End For Social Media As We Know It”. What’s happening is that Meta is rolling out social media platforms with the word “plus” added at the end. When that happens it means Facebook Plus, Instagram Plus, WhatsApp Plus, and more will lock features behind a paywall unless users pay a premium price. Naomi Gleit is the Head of Product at Meta and she had this to say about the new “plus” platforms:
“The Plus-sized services are, Gleit says on Facebook, rolling out today and should offer ‘premium features that unlock more from our apps and our AI glasses.’”
Boo.
What that really means is that
“The only tangible change, though, may be Meta AI falling in step with many of its generative AI competitors, and adding more capacity, the ability to handle more complex requests, and “more room to create.” Sure, this is fuzzy, at best, lacking details like how many daily/monthly processing tokens or even how many prompts.”
Social media will follow the same path as streaming services. Amazon wants you to pay so you can pay to view better digital goodies. YouTube wants to be the new cable TV and Hollywood. If you thought popcorn at the dying movie theaters is expensive, wait for the Google’s next price increase. These services include special messages too. Paywalls and subscriptions are the in thing.
There isn’t an upside for adults, but there is an advantage when it comes to the kiddos. If social media is locked behind a paywall like Roblox and Fortnite then kids won’t be able to access services without mommy’s and daddy’s credit cards. Paywalls work better than the Australia-type of fake-aroo age limits.
The good, old Internet is gone.
Whitney Grace, June 10, 2026
-
- Subscribe to Beyond Search
Feature archive
News archive
-
