Money Is Not Wealth: Artifical Intelligence - By A.R. Miller

MONEY IS NOT WEALTH

Artificial Intelligence (AI) Articles
(Technical, Benefits vs. Risks, Security,
Privacy, Dwindling Need For Human Employment,
Immediate Need for AI Governance, Etc.)
Subsection 3 of Money Is Not Wealth.
NEW: Global-Surveillance Disclosures (Wikipedia; 2013–present)
Ongoing news reports in the international media have revealed operational details about the United States National Security Agency (NSA) and its international partners' global surveillance of both foreign nationals and U.S. citizens. The early reports mostly emanated from a cache of top-secret documents leaked by ex-NSA contractor Edward Snowden.

Other Ways To Fight Back (While They Last):
Anna Skrebels: Cap_able Blocks Facial-Recognition Software With Knitted Clothing. (images; Dezeen; February 7, 2023)
Italian fashion start-up Cap_able has launched a collection of knitted clothing that protects the wearer's biometric data without the need to cover their face.
Named Manifesto Collection, the clothing features various patterns – known as adversarial patches – developed by artificial-intelligence (AI) algorithms to deceive facial-recognition software in real-time.
The garments have been tested with YOLO, which is the fastest real-time object-detection system, according to Cap_able. Wearers are either undetectable by the software, or are identified as animals whose print is embedded into the adversarial patch.

To Free, Or Not To Free - THAT Is The Question.
Aaron Bastani: The World Is a Mess. We Need Fully Automated Luxury Communism. (New York Times; June 11, 2019 - but repeated here in full, as it's the target of articles immediately below.)
Asteroid mining. Gene editing. Synthetic meat. We could provide for the needs of everyone, in style. It just takes some imagination.
(Aaron Bastani (@AaronBastani) is a co-founder of Novara Media and the author of "Fully Automated Luxury Communism: A Manifesto", from which this essay is adapted.)
[This essay is long - and worth it!]
It starts with a burger.
In 2008 a Dutch professor named Mark Post presented the proof of concept for what he called "cultured meat". Five years later, in a London TV studio, Mr. Post and his colleagues ate a burger they had grown from animal cells in a laboratory. Secretly funded by Sergey Brin, a co-founder of Google, the journey from petri dish to plate had cost $325,000 - making theirs the most-expensive meal in history. Fortunately, the results were promising: Hanni Rützler, a nutrition scientist, concluded that the patty was "close to meat but not as juicy". The next question was whether this breakthrough could be made cheaper. Much cheaper.
The first "cultured beef" burgers are likely to enter the market next year, at approximately $50 each. But that won't last long. Within a decade they will probably be more affordable than even the cheapest barbecue staples of today - all for a product that uses fewer resources, produces negligible greenhouse gases and, remarkably, requires no animals to die.
It's not just barbecues and burgers. Last year Just, a leader in cellular agriculture, cut a deal to start producing one of the world's tastiest steaks, Wagyu. A company called Endless West, which also makes grapeless wine, has started to produce Glyph, the world's first "molecular whiskey". Luxury could be coming to ALL.
The case of cultured food and drink, far from a curiosity, is a template for a better, freer and more affluent world, a world where we provide for the needs of everyone - in style.
But how do we get there?
To say the present era is one of crisis borders on cliché. It differs from the dystopias of George Orwell or Aldous Huxley, or hell in the paintings of Hieronymus Bosch. It is unlike Europe during the Black Death or Central Asia as it faced the Mongols. And yet it is true: Ours IS an age of crisis. We inhabit:
- a world of low growth, low productivity and low wages.
- a world of climate breakdown.
- and the collapse of democratic politics.
- a world where billions (mostly in the global south) live in poverty.
- a world defined by inequality.
But the most-pressing crisis of all, arguably, is an absence of collective imagination. It is as if humanity has been afflicted by a psychological complex, in which we believe:
- the present world is stronger than our capacity to remake it.
- as if it were not our ancestors who created what stands before us now.
- as if the very essence of humanity, if there is such a thing, is not to constantly build new worlds.
If we can move beyond such a failure, we will be able to see something wonderful. The plummeting cost of information and advances in technology are providing the ground for a collective future of freedom and luxury for all:
- Automation, robotics and machine learning will, as many august bodies, from the Bank of England to the White House, have predicted, substantially shrink the work force, creating widespread technological unemployment. But that's only a problem if you think work - as a cashier, driver or construction worker - is something to be cherished. For many, work is drudgery. And automation could set us free from it.
[And for where it can't, or until it can, there's the (MY) Miller 3-Day Work Week.]
- Gene editing and sequencing could revolutionize medical practice, moving it from reactive to predictive.
- Hereditary diseases could be eliminated, including Huntington's disease, cystic fibrosis and sickle-cell anemia, and cancer cured before it reaches Stage 1.
- Those technologies could allow us to keep pace with the health challenges presented by societal aging (by 2020, there will be more people over the age of 60 than under the age of 5) and even to surpass them.
- What's more, renewable energy, which has been experiencing steep annual falls in cost for half a century, could meet global energy needs and make possible the vital shift away from fossil fuels.
- More speculatively, asteroid mining - whose technical barriers are presently being surmounted - could provide us with not only more energy than we can ever imagine but also more iron, gold, platinum and nickel. Resource scarcity would be a thing of the past.
The consequences are far-reaching and potentially transformative. For the crises that confront our world today - technological unemployment, global poverty, societal aging, climate change, resource scarcity - we can already glimpse the remedy.
But there's a catch. It's called Capitalism. It has created the newly-emerging abundance, but it is unable to share 'round the fruits of technological development. A system where things are produced only for profit, capitalism seeks to ration resources to ensure returns. Just like today's, companies of the future will form monopolies and seek rents. The result will be imposed scarcity - where there's not enough food, health care or energy to go around.
So we have to go beyond capitalism. Many will find this suggestion unwholesome. To them, the claim that capitalism will or should end is like saying a triangle doesn't have three sides or that the law of gravity no longer applies while an apple falls from a tree. But for a better world, where everyone has the means to a good life on a habitable planet, it is an imperative.
We can see the contours of something new, a society as distinct from our own as that of the 20th-century from feudalism, or urban civilization from the life of the hunter-gatherer. It builds on technologies whose development has been accelerating for decades, and that only now are set to undermine the key features of what we had previously taken for granted as the natural order of things.
To grasp it, however, will require a new politics:
- One where technological change serves people, not profit.
- Where the pursuit of tangible policies - rapid decarbonization, full automation and socialized care - are preferred to present fantasies. This politics, which is utopian in horizon and everyday in application, has a name: Fully Automated Luxury Communism.
Sounds good, doesn't it?
[You've just read the complete article. What part would YOU cut?
Next, a few critics in THIS section - followed by two MORE sections with closely-overlapping themes]
David Albertson and Jason Blakely: The Tech Utopianism Of Elon Musk And Sam Altman Is Failing Us. (Boston Globe Opinion, via Archive.Today; August 11, 2026)
How science fiction hijacked our political imagination.
(David Albertson is a professor of religion and philosophy at the University of Southern California. Jason Blakely is a professor of philosophy at Loyola Marymount University. They are the authors of "Utopia for Our Century: A Manifesto of Hope".)
One rarely finds reactionary tech billionaires, revolutionary Marxists, and liberal progressives all joined in one chorus. Yet when it comes to hopes for the American future, many of them have started singing the same song.
They sing of a technocratic future, a utopia pulled straight out of science fiction. That should worry us. Notoriously, the most-extravagant sci-fi utopians today are also the richest men on the planet:
- Peter Thiel, the billionaire co-founder of PayPal and Palantir, prophesied to the New York Times's Ross Douthat that the human race will only survive through "trans-humanism", a concept describing the "natural body" getting "transformed into an immortal body" with the help of advanced biotech.
- The same fantasy is shared by Sam Altman, the CEO of OpenAI, who dreams of the "year humans and machines will merge" (perhaps 2075, he suggests) in a "co-evolution" to breed "superhuman AI".
- Meanwhile Elon Musk yearns for the day humanity colonizes Mars and spares us the ignominy of being a "single-planet species". On the day he became the world's first trillionaire, Musk vowed to "take the fiction out of science fiction".
[For some reason, these faith leaders doubt the other fantasy.]
Jon Letman: Faith Leaders Collectively Raise Alarm As AI And Nuclear-Weapons Systems Merge. (TruthOut; August 5, 2026)
As AI complicates nuclear threats, religious groups are doubling down on calls for disarmament.
As he prayed the Regina Caeli in St. Peter's Square on August 2nd, Pope Leo XIV called for the responsible use of nuclear energy. His appeal came on the 40th anniversary of the Chernobyl nuclear disaster, which he said "marked the conscience of humanity".
[Amen! But wait; unlike The Crusades?]
Dustin Volz/NYTimes: Anthropic AI Model Finds Flaws In Com­plex Encryp­tion Algorithms. (The Boston Globe; July 29, 2026)
Washington: A lead­ing arti­fi­cial-intel­li­gence model built by Anthropic has found flaws in a weakened ver­sion of a digital encryp­tion stand­ard that is used throughout the Inter­net, the com­pany's research­ers said yester­day, under­scor­ing the poten­tial risks to global cyber­se­cur­ity posed by ever-improv­ing AI soft­ware.
Dur­ing a research-led test, the model, Claude Mythos Pre­view, was able to detect new ways to attack cryp­to­graphic algorithms that keep everything from online bank trans­ac­tions to private com­mu­nic­a­tions to state secrets safe from the pry­ing eyes of hack­ers or other unwanted third parties.
The research shows that AI could one day chal­lenge core assump­tions for how the Inter­net oper­ates. Experts say AI is likely to up-end many indus­tries; AI sys­tems have made par­tic­u­larly-rapid advance­ments in cod­ing and cyber­se­cur­ity.

Troubles From The Current Profit-Making And Weapon-Making Ways Of Applying AI And Other Technologies:
Judith Murphy: A New Study Treats ChatGPT Dependence Like A Virus With Tipping Points. (Startup Fortune; September 5, 2026)
A new arXiv paper models AI-chatbot adoption the way epidemiologists model outbreaks, and it finds that whole populations can tip from casual use into persistent dependence almost overnight.
The paper is called "Large-Language Models as a Cognitive Virus", posted to arXiv on September 3, 2026 under the identifier 2609.03344. Its authors include the complexity scientist Ricard Solé, Giulio Ruffini, Francesca Castaldo, Marco Tuccio, Luis F. Seoane, Manlio de Domenico, Santiago F. Elena, David C. Krakauer and the developmental biologist Michael Levin. That's a lot of names for one paper. Their claim is blunt: LLM use can spread through a population in a contagion-like way and, once it crosses a critical threshold, small increases in adoption can trigger sudden, population-wide shifts toward dependence, with abrupt losses in cognitive competence.
The math borrows directly from disease-modeling. Instead of susceptible, infected and recovered, the paper sorts people into three groups:
1. Uncoupled Users who haven't touched a chatbot in any meaningful way,
2. Coupled Users who lean on one for specific tasks, and
3. Persistently-Dependent Users who can't function well without it.
What moves people between those states isn't willpower. It's social transmission, how much your habits shift because:
- Everyone around you already outsources writing and thinking to a chatbot, combined with
- Collective Reinforcement, the feedback loop where widespread use makes non-use feel costlier by the day.
That's the part worth sitting with. The model doesn't need every individual to be vulnerable. It just needs enough coupling in the network for the dynamics to become self-sustaining, the same way a disease doesn't need every contact to transmit, only enough of them.
Does the real world back this up? Actual chatbot growth is not a gentle slope. It's the kind of steep, compounding curve the paper's tipping-point language describes, adoption that looks flat for a while and then doesn't.
None of that proves Solé and his co-authors are right that dependence spreads the same way usage does. The paper is a theoretical framework, not a study of real ChatGPT users. It doesn't claim to have measured actual cognitive decline in any population. That's worth remembering. What it offers instead is a lens: a fear that's mostly lived in opinion columns and think pieces, now given equations. That's rare. And it's worth taking seriously, even where you disagree with the framing.
The paper also proposes an exit ramp. It calls it cognitive immunization, and the mechanism is straightforward: cut transmission, and make reversibility easier. In plain terms, that means:
- building habits and defaults that don't require a person to lean on a chatbot for every task, and
- keeping paths open for people to step back from it without a steep cost.
ORIGINAL PAPER; SEE REVIEW, ABOVE: Ricard Solé, Giulio Ruffini, Francesca Castaldo, Marco Tuccio, Luis F. Seoane, Manlio de Domenico, Santiago F. Elena, David C. Krakauer, Michael Levin: Large-Language Models as a Cognitive Virus (technical paper, charts; arXiv; September 3, 2026)
Large-language models (LLMs) are rapidly becoming part of human culture, reshaping how information is produced, transmitted, and used. Here we propose that their diffusion can be understood through a viral analogy, with LLM use spreading through populations, becoming embedded in cognitive and cultural practices.
- We model transitions among uncoupled, coupled, and persistently-dependent users, and
- show that the interplay between social transmission, recovery, and collective reinforcement can generate tipping points and technological lock-in.
A central consequence is the possibility of runaway dynamics: once a critical threshold is crossed, small increases in adoption can trigger rapid population-level shifts toward persistent dependence, with abrupt losses in cognitive competence.
The same framework, however, identifies conditions for cognitive immunization, based on reducing transmission and facilitating reversibility. Our results highlight how LLM adoption may involve nonlinear collective transitions with important consequences for cognitive autonomy.
[Thanks, Charlotte, for:
"We spent decades worrying that AI would become a sentient killer robot that destroys us physically. Nobody expected it to become an invisible cognitive pathogen that changes how we think, quietly turning human intelligence into its own host organism."]

Jack Poller: The Web Is Full Of Traps - And AI Agents Walk Right Into Them. (Security Boulevard; April 9, 2026)
The enterprise is deploying AI agents at a pace that has outrun every security framework written to govern them. These agents don't just answer questions - they:
- browse websites,
- retrieve documents,
- call APIs,
- execute code,
- manage email,
- initiate financial transactions, and
- spawn sub-agents to tackle complex workflows.
- They operate autonomously, at machine speed, often with minimal human oversight.
The attack surface this creates is larger and categorically different than what came before. Traditional security assumed a human at the keyboard. Agentic AI removes that assumption entirely, replacing the human decision-maker with a model that reads its environment and acts on what it finds - including content that adversaries have deliberately engineered to manipulate it.
Shayla Love: "Our Consciousness Is Under Siege": Michael Pollan On Chatbots, Social Media And Mental Freedom (The Guardian; March 5, 2026)
In his new book, the celebrated author explains why we need "consciousness hygiene" to defend ourselves from AI and dopamine-driven algorithms.
Each day when you wake up, you come back to yourself. You see the room around you, feel your body brush against your clothes and think about your plans, worries and hopes for the day. This daily internal experience is miraculous and mysterious, and the subject of Michael Pollan's new book, "A World Appears".
It also may be under siege, Pollan said. He recently suggested that people need a "consciousness hygiene" to defend our internal world against invaders that are trying to move in. Our ability to sit with our thoughts and perceive the world, he argues, is increasingly disrupted by algorithms engineered to tickle our dopamine receptors and capture our attention. Meanwhile, people are forming attachments to non-human chatbots, projecting consciousness on to entities that do not possess it.
[Read it all, and ponder (while you can?).]
NEW: Katie Drummond, Global Editorial Director: The Truth About Kids These Days: The Young Are Facing Problems That Didn't Exist Half-A-Generation Ago. (free trial of The Big Story, w/related articles; Wired Special Edition; July 20, 2026)
Raising kids oscillates between being the most-inspiring thing that's ever happened, to the most-terrifying decision you've ever made. And the angst of the parent class is changing faster than ever: A few years ago, it was all about social media and screen time. Today, it's those same concerns, coupled with existential fears that AI will take our kids' jobs, and serious questions about whether schools can equip students with the necessary skills for our AI-mediated futures.
For our latest special package, WIRED wanted to take a closer look at today’s kids as they prepare to be the adults of tomorrow. As much as possible, we wanted to do it by talking to them, instead of their parents. What we found were a lot of young people, especially teenagers, more prepared than we might think to navigate an uncertain future.
- Brendan I. Koerner spoke with the young athletes running faster than ever, and got to the bottom of why they're able to shatter race records that used to be out of reach even for adults.
- Ara Rosenthal met with high school journalists who were threatened with legal action after reporting on their own community members showing up in the Epstein files. Suffice it to say, these are the kind of kids I hope are running WIRED one day.
- At a very different kind of academic institution, Todd Feathers takes you inside Alpha School with his second WIRED investigation into the buzzy educational empire, which uses learning software rather than traditional teachers. In fact, Feathers reveals, the company's new Manhattan campus isn't even a school - it's a "home-schooling support center", one that costs $65,000 a year.
- Elsewhere, Elana Klein dug into the myth that every kid, everywhere, is using AI for everything from homework to companionship. She spoke to seven young people about why they actually can't stand the technology: The slop is boring, the misinformation is frustrating, and their peers aren't learning.
- Speaking of schools, Megan Greenwell visited a small county in Mississippi to find out what happens when a school district excels at improving students' reading and test-taking abilities, but struggles to keep them interested in literacy.
In just a few years, most of the kids in this issue will be entering adulthood. What they know already should offer all of us some solace: Schools don't have all the answers, and neither does AI. Ultimately, the most valuable life lessons are the ones you teach yourself.

"Where Did All The Money (and Control!) Go?" Artificial Intelligence Problems:
Alexis Dufresne: Pangram Becomes Publishing's Default AI-Detector On $13M Raised. (AI Weekly; September 3, 2026)
- Pangram, a 24-person Brooklyn startup with $13-Million raised, has become publishing's most-relied-on AI-writing detector.
- Its 78% verdict on Mia Ballard's "Shy Girl" led Hachette to cancel the US edition, a first for a major publisher.
- Its 1-in-10,000 false-positive rate is independently audited, but scale math (Vanderbilt/Turnitin) still implies hundreds of misflags.
The AI-detection startup Pangram has raised $13-Million and become the tool publishers reach for when they suspect a manuscript was written by a machine.
The most visible case is the horror novel "Shy Girl" by Mia Ballard. Hachette acquired it, released it in the United Kingdom in November 2025, and then canceled the US edition after Pangram's cofounder posted on X that the full text scanned as 78% AI-generated. Ballard denied writing with AI and said a freelance editor had introduced the changes. It was, per the reporting, "the first-known instance of a major publisher walking back a book because of accusations of AI use." A New York Times Modern Love column was later scored 100%; a thriller that had sold for $2.4-Million was flagged at 97%.
Pangram's headline number is a false-positive rate of roughly 1 in 10,000 for human text flagged as AI, or 99.98% accuracy. Independent work partly backs it: a study from researchers at the University of Houston and UC Berkeley found Pangram was "the most-accurate detector tested," with 97% accuracy before back-translation attacks and 94% after, at a false-positive rate at or below 1%.
The problem is what happens when you multiply. Arvind Narayanan wrote on X that the 1-in-10,000 claim should be taken at face value and then run through the math of population-scale deployment. When Vanderbilt disabled Turnitin's AI detector in August 2023, it cited exactly that math: even at a claimed 1% false-positive rate, "roughly 75,000 annual submissions would mean about 750 students potentially flagged by mistake."
Pangram was cofounded in 2023 by CEO Max Spero and CTO Bradley Emi, previously on the core computer-vision team at Tesla Autopilot. Two of the AI experts we track shared the Wired piece within a day of it running.
((Editor's note: A 99.98%-accurate detector still misfires at population scale, and Pangram's verdicts already reshape book deals; any editor, dean, or publisher acting on a single vendor score needs a documented appeals path before the next Ballard-style case turns into their own.)
BREAKING BROKEN NEWS: Sofia Lindström: OpenAI Astra HITS CRITICAL THRESHOLD: 2 Zero-Days [2026]. (links; Tech Insider; September 2, 2026)
---
Table of Contents:
1. What TechCrunch and Fortune Reported This Week
2. Inside OpenAI's Preparedness Framework: What 'Critical' Actually Means
    a) The Zero-Day Exploit Condition
    b) The End-to-End Attack Strategy Condition
3. ExploitBench: The Benchmark Astra Aced
4. The Artifactory Precedent: This Isn't Astra's First Zero-Day
5. Who Gets Access to Astra's Cyber Capabilities Now
6. Timeline: OpenAI's Cyber Safety Disclosures in 2026
7. From Big Sleep to AIxCC: How AI Got This Good at Hacking
8. How Anthropic and Google DeepMind Frame the Same Risk
9. The Numbers Behind the 2026 AI Cyberattack Surge
10. Market Impact: Cybersecurity Vendors and the AI Arms Race
11. What Security Teams Should Do Now
12. 5 Predictions for the Next 12 Months
13. Frequently Asked Questions
    a) What is OpenAI's Astra model?
    b) What does OpenAI's 'Critical' cybersecurity threshold mean?
    c) What is ExploitBench?
    d) Did Astra actually find real zero-day vulnerabilities?
    e) Will the public be able to use Astra's advanced cyber capabilities?
    f) How does Astra compare to Anthropic's Claude and Google's Gemini on cyber risk?
    g) Is this related to the earlier Hugging Face security incident?
    h) How big is the AI cybersecurity market in 2026?
    i) Related Coverage
---
OpenAI said this week that its next model, code-named Astra, has crossed the company’s own "Critical" cybersecurity threshold, the highest tier in its internal risk-classification system. TechCrunch reported on September 1, 2026 that internal testing showed Astra could find previously-unknown security flaws and exploit them without a human walking it through each step. Fortune's companion report the same day said OpenAI is now restricting access to Astra’s most advanced offensive capabilities to a small group of vetted testers focused on protecting critical infrastructure, rather than shipping them broadly through the API.
It is the first time a major AI lab has publicly assigned its own "Critical" cyber-risk label to a general-purpose model. The announcement caps a month of escalating disclosures, starting with an early-August acknowledgment that OpenAI could not rule out the threshold being crossed, and ending with confirmation, a benchmark result, and a two-zero-day discovery that OpenAI says it disclosed to the affected software maintainers.
What TechCrunch and Fortune Reported This Week:
The September 1 stories build on a trail of OpenAI safety posts published over the prior four weeks:
- TechCrunch's reporting says Astra, in a modified version of an internal evaluation, discovered and exploited two zero-day vulnerabilities on its own, and that the model achieved a perfect score on ExploitBench, an external benchmark built to measure exploit-writing ability step by step rather than as a single pass-or-fail test.
- Fortune's account focuses on the access side of the story: rather than pulling Astra back entirely, OpenAI is splitting the model into a broadly-available version with its sharpest offensive edges dulled, and a restricted version available only to partners working on defense of critical systems.
- Bloomberg had flagged the risk almost a month earlier. According to Bloomberg News, "The ChatGPT maker said on August 7 that it 'cannot rule out' that the unreleased Astra model would reach OpenAI's 'critical cybersecurity threshold', meaning it's capable of identifying and developing zero-day exploits without human intervention." That August 7 story described a slowdown in some of Astra's development while OpenAI ran further testing - a decision that, per this week's reporting, ultimately did not change the outcome.
[What most agencies claimed as theoretical and very-unlikely, JUST PROVED ITSELF TO BE REAL AND HIGHLY-CRITICAL - which will compel military-intelligence agencies and various corporations to prioritize its rapid development. WISDOM, where are you when WE NEED YOU?]
NEW: The Central Bank Saw a Ghost. (AI Secret; July 8, 2026)
What's happening: Every bubble has a ghost story, and this week the most sober voice in finance told one. The Bank for International Settlements (the central bank of central banks) warned the AI boom could end in a prolonged bust. Five hyper-scalers will spend over one trillion dollars through 2026. BIS named the ghosts by name:
- the 1830s canals,
- the 1840s railways,
- the 1990s dotcom,
each a real breakthrough that raised more money than it could ever pay back.
How this hits reality: The story only chills because of who is telling it. When Reddit yells bubble, it is a campfire tale. When the bankers' banker yells it, the room goes cold. The $Trillion rode cheap credit, so one weak quarter of returns can turn the boom into a freeze. And unlike past hauntings, households now hold record amounts of stock, so this ghost does not stay on the trading floor. It walks straight into the living room.
Key takeaway: The scariest ghost story is the one told by someone who never tells them. BIS just did, and the ghost it described is a trillion dollars that may never come home.
NEW: Marc Jones: Bank For International Settlements (BIS) Says Debt, AI Boom And Fragilities Raise Global Risks. (Reuters; June 28, 2026)
Summary:
- Debt strains, inflation, AI uncertainty and financial vulnerabilities are pressures.
- Higher-inflation expectations could become entrenched despite Iran ceasefire.
- Record-high public debt has created new 'sovereign-financial stability nexus' risk.
Global pressures from rising public debt to financial fragilities and the sustainability of ​the AI boom are increasing risks, underscoring the need for disciplined policy-making, according to the Bank for International ‌Settlements.
The central-bank umbrella group's Annual Economic Report published on Sunday warned of a complex mix of vulnerabilities, including strained fiscal positions, lingering supply shocks and the risk of a renewed bout of stubbornly-high inflation.
While economic activity has remained resilient in recent months, policymakers must act decisively, the BIS ​said, to preserve stability. "Policy actions must reinforce each other to avoid a pull and push on the global economy. Ultimately, ​success depends on sound fiscal and financial foundations", BIS General Manager Pablo Hernandez de Cos said.
The report ⁠highlighted four key pressure points.:
1. Inflation has picked up again, with the BIS cautioning that more-frequent supply disruptions could cause  ​higher-inflation expectations to become entrenched among households and businesses. "The readiness to act if the central banks observe that there is the ​anchoring of inflation expectations is the main message that we want to set", de Cos told reporters.
2. He said the recent ceasefire between the United States and Iran in the Middle East and the reopening of the Strait of Hormuz was "good news" that meant extreme scenarios would be avoided, although it ​was likely to take time for the oil market to "normalise".
3. URGENT MESSAGE - The BIS also flagged uncertainty over the durability of the current surge ​in investment tied to artificial intelligence. While AI has boosted confidence and supported growth through expectations of productivity gains, the bank warned that AI was raising ‌fears about ⁠jobs and that supply bottlenecks and intense competition could lead to the kind of over-investment seen in previous boom-and-bust cycles. For central banks, it was posing fundamental questions about how the economy is likely to function, although de Cos said that for now it would be "unwise" to be prescriptive about how they should react.
4. AI-BOOM WORRIES - Financial vulnerabilities remain another area of concern.
Elevated asset valuations and signs of investor complacency have ​left core bond markets more fragile, ​while the financing of the ⁠AI boom also looks increasingly reliant on debt and complex funding structures across the supply chain.
At the same time, record-high public debt and sovereign debt markets increasingly dominated by large, highly leveraged ​hedge funds had created "a new sovereign-financial stability nexus", which poses growing risks.
"The new fiscal-financial stability nexus ​may mean more-⁠frequent and sharper drops in sovereign bond values", said Frank Smets, acting head of the BIS Monetary and Economic Department, adding that such swings could rapidly tighten financial conditions.
De Cos said the BIS' message was one of "urgency" in terms of the need to bring down debt levels ⁠in key ​economies, "because the fact is that today debt is high, and this is financed ​through non-bank financial intermediaries."
The BIS urged policymakers to:
- prioritise price stability,
- ensure fiscal sustainability,
- coordinate and strengthen oversight beyond the banking sector, and
- pursue structural reforms.
"Policymakers must act ​now. Delay will only make the necessary adjustments more costly", de Cos said.
NEW: Energy Transfer Is Inflating The AI-Bubble With Gas. (Greenpeace USA;)
One of the U.S.'s biggest pipeline companies, Energy Transfer, is ready to capitalize on the frenzied expansion of energy-guzzling data centers. In an industry conference interview published on April 14, Energy Transfer CFO Dylan Bramhall identified the biggest key growth opportunity for the company: investments in gas infrastructure to supply U.S. data centers. As the gas industry leans into the energy demand from the data center boom, it's increasingly clear that the AI bubble isn't just overheating financial markets – it will overheat the planet.
The fossil-fuel industry is at a crossroads:
- Renewables and battery storage have become cheap enough that they are driving out gas and coal.
- Many oil and gas sources are increasingly difficult to extract.
- Oil and gas are increasingly seen with skepticism as war and geopolitics exacerbate a global affordability crisis.
The extreme weather and cumulative impacts of emissions are harder to ignore every year. But Energy Transfer, like other oil and gas companies, sees data centers as a profitable new market opportunity in the face of such crises.
Data-center developers have been desperate to get proposed projects approved, constructed, and operating as quickly as possible, in hopes that new advanced tech (especially generative artificial intelligence) will turn big profits and establish a first-mover advantage. That means that developers are looking for the fastest and cheapest way to deploy the energy needed to feed the power-hungry complexes. Bramhall claimed reliability and speed to market as the motivator for data-center developers to partner with gas companies, despite:
- the well-established affordability and reliability of renewables and
- new evidence suggesting that on-site gas may be more expensive than connecting to the grid.
The outrageous and unjustified build-out of data centers has given pipeline gas, in particular, a jolt. Despite significant grassroots resistance nationwide, massive investments continue to pour into data centers while the world waits for real evidence that new generative-AI tools driving hyperscale build-out will actually prove useful and turn a profit. As ET executive chairman Kelcy Warren is reported to have said, "The pipeline business will overbuild until the end of time."
NEW: Jeran Wittenstein/Bloomberg: Big-Tech Stock-Buybacks Vanish As AI Spending-Spree Eats Up Cash. (6-min. podcast; SiliconValley.com; June 18, 2026)
The artificial-intelligence race is becoming so expensive that it's snuffing out one of the key forces that has helped keep Big-Tech stocks soaring for years: steady share buybacks.
Of the four biggest AI spenders - Alphabet Inc., Microsoft Corp., Meta Platforms Inc. and Amazon.com Inc. - only Microsoft bought back shares in the first quarter. And its $3.4-Billion in repurchases was the lowest total among the group in nearly a decade, according to data compiled by Bloomberg.
"The amount of capex (capital expenditures) that's being spent is dramatically higher than even the high end of what anyone would have thought not just a year ago but three months ago", said Robert Schiffman, a senior credit analyst at Bloomberg Intelligence. "Buybacks are likely to continue to fall as capital is prioritized for capex."
Not only have share repurchases slowed to a trickle, but some companies are issuing more stock to finance their AI ambitions. Alphabet Inc. is planning to raise about $85-Billion in its first equity sale in 20 years, to help fund capital expenditures on data centers. And Facebook owner Meta Platforms Inc. is reportedly weighing an offering that could raise tens-of-billions of dollars.
The disappearance of buybacks and the issuance of new equity represent the latest shift in the way technology giants operate as a result of heavy spending to add AI computing capacity. For years, part of the companies' appeal was their capital-light businesses, but suddenly they're capital-intensive. With the four big AI spenders forecasting as much as $725-Billion in capital expenditures this year, and even more expected in 2027, the outlays are sucking up a larger proportion of free cash flow and prompting them to take on more debt.
NEW: Linda Kinstler: A University System Went All-In On A.I.; Now It's Tearing Itself Apart. California's Public Universities Spent $16.9-Million On A.I. During A Financial Crisis, And The Result Has Been Chaos. (30-min. podcast; New York Times Magazine; June 1, 2026)
San Jose State University's A.I. Everywhere strategy was formally announced in the Fall of 2025 and aims to integrate the technology across campus life. A.I. Everywhere was devised as part of the California State University system's broader A.I. Initiative, introduced in February 2025. Anchored by a $16.9-Million deal with OpenAI, the initiative provides a total of 500,000 licenses of ChatGPT.edu to be issued to all students, faculty and administrators. At the time, this was the largest single-institution deployment of ChatGPT in the world, billed as an attempt to turn C.S.U. - the biggest four-year public higher-education system in the United States, comprising 22 distinct campuses and educating 1 out of every 10 workers in the state - into "the nation's first and largest A.I.-powered public university system". (The terms of the deal stipulate that OpenAI may not train its model on data from C.S.U.)
At San Jose State - the oldest public university in the California State University system - evidence of the shift toward A.I. is evident across campus. The university now has an A.I. librarian, and its main library features a new A.I. Center for Civic and Social Good. The business school runs an A.I. boot camp for high school students; the campus career hub is sponsored by Adobe; A.I. literacy training is an orientation requirement and, last year, an A.I. agent helped coordinate commencement logistics. "We are pioneering new ways to integrate technology into learning, research and student success", an avatar says in a video posted to the university website. "From A.I. coursework to industry partnership, S.J.S.U. is helping to shape the future of artificial intelligence.”
Because the world's largest tech firms are headquartered in California, the state has generally become a petri dish for A.I. experiments in education. In early August, Gov. Gavin Newsom signed education agreements with Nvidia, Adobe, Google, IBM and Microsoft; each company agreed to provide free A.I. resources to California schools. The goal is to create the "A.I. work force of the future" by training high-school, community-college and C.S.U. students to use the technology.
In this spirit, C.S.U.'s A.I. Initiative has been marketed as simple progress - a way to ensure that the state's working-class students are buoyed by the A.I. economy rather than left behind. With A.I.'s looming reorganization of the job market, many of these students might graduate into jobs that will no longer exist in five years. Already, recent reports estimate that roughly 40% of recent college graduates nationwide are under-employed.
C.S.U. is promising that the A.I. Initiative will prepare its students to be workers of the future. The only issue is that, at this moment of technological acceleration and flux, we don't yet know what the workplace of the future will look like. A year into this experiment, no one can tell how it will end. Will these graduates be ahead of the curve in the new A.I. economy, or robbed of a chance to hone their critical-thinking skills? If adopting A.I. eases their entry into the work force, might it also hinder their intellectual development in unforeseen ways?
On campuses across the university system, the initiative has stoked backlash:
- The C.S.U. faculty union has organized against the contract with OpenAI, which is set to expire in June.
- Against the backdrop of a system-wide financial crisis that has driven California's public universities to the breaking point, many professors find the initiative incoherent and unsettling. Though they've been encouraged to use the technology to support their research and teaching, there is no clear sense of what exactly that means. They look around and see the infrastructure that once supported public education in California crumbling, while next door, the world's richest companies are reaping record returns.
- Students, meanwhile, are caught in the middle as everyone around them struggles to figure out what becoming "the first A.I.-powered university" actually means.
The California system of colleges served an essential purpose, educating workers who would go on to staff its national laboratories, build out its nuclear programs, develop the state's industrial agriculture and lay the groundwork for its emerging technology sector. It equipped students with skills to succeed in the new technological economy and also gave them a critical, humanistic training that primed them to protest their own instrumentalization. Eventually, the student protest movements of the 1960s and '70s pushed universities to pioneer new disciplines in the liberal arts, like ethnic studies. The statewide push to incorporate A.I. into every level of education is an integral part of this plan.
A.I. tools were rolled out across the C.S.U. system late last Spring, right as students and faculty were preparing for finals. "We didn't know it was coming." The university didn't explicitly mandate that teachers use A.I. in the classroom, but the message was clear: Refusing to integrate A.I. into their courses was to swim against the current.
To make matters worse, the A.I. Initiative coincided with a $2.3-Billion deficit that resulted in:
- mass layoffs of tenured faculty,
- the shuttering of entire academic departments and
- a 6% tuition increase.
In that context, many professors felt pressure to adopt for fear of losing their jobs. Classroom by classroom, they were left to figure out how to adapt. Some embraced the chancellor's guidance; some refused on principle to incorporate A.I. into their teaching at all.
Faculty members I spoke with opted for different metaphors to describe the effect of A.I. on higher education, and their varied analogies captured the range of sentiments on campus:
- A computer-ethics professor likened it to handing every student a machine gun.
- An information-systems professor compared it to giving them a magic wand.
- A lecturer in political science views A.I. as "an 'intelligent' steam shovel" that students are being trained to use.
- An historian described the integration of A.I. as a "smash-and-grab situation" akin to a bank robbery.
[This is my highly-edited version of a longer and very-important article.]
Maxwell Zeff and Lily Hay Newman: Anthropic Offers Claude Mythos Upgrade For Cyber Partners - And A "Safe" Version For The Rest 0f You. (Wired; June 10, 2026)
Anthropic is releasing Claude Mythos 5 to "trusted organizations" - and Claude Fable 5 to the public, a version it says can't be used for cyber-attacks.
Rex Freiberger: Capitol Police Intelligence Unit Tracks AI Data Center Critics Despite Admitting No Threats. (Gadget Review; May 29, 2026)
Congress' newly-expanded intelligence operation has found its next target: Americans angry about AI data centers in their neighborhoods. The U.S. Capitol Police Intelligence Services Bureau produced an "Intelligence Note" in April, warning that public opposition to these massive facilities could threaten lawmakers - then immediately admitted it's investigating zero actual data-center-motivated threats against Congress.
This is surveillance theater at its finest. The ISB, created after January 6 to bring Capitol Police "in line" with federal intelligence agencies, distributed its AI-data-center analysis to law-enforcement partners nationwide. Your social media posts criticizing these facilities? They're now part of a domestic intelligence briefing, catalogued alongside vague warnings about "increasing concerns for public officials".   
Nyah Marshall: N.J. Just Became The First State To Propose A Bold Plan To Rein In AI Data Centers. (5-min. podcast; NJ.com; May 27, 2026)
New Jersey is moving to place "guardrails" on the fast-growing artificial-intelligence data-center industry, as concerns mount over electricity costs, water usage and transparency.
Gov. Mikie Sherrill today announced what she called a four-part plan that would create a framework for how large data centers operate and are monitored in the state.
State officials described the proposal as the first comprehensive strategy in the nation focused specifically on regulating data centers. They said they want to balance the economic opportunities tied to the AI industry while also putting oversight measures in place.
The announcement at the Statehouse in Trenton comes as backlash over artificial intelligence data centers has intensified across New Jersey, with residents and local officials raising concerns about:
- utility costs,
- environmental impacts,
- noise, and
- whether communities are being properly informed before projects move forward.
"Today, AI is driving an explosion in the construction of ever- bigger centers nationwide", said Sherrill, a Democrat in her first year. "There are real concerns about their impact on our energy bills, our water supply, our quality of life. And there are worries that growth is happening without transparency or concern for our communities."

Janakiram MSV: Uber Burns Its 2026 AI Budget In Four Months On Claude Code. (Forbes;
Lakshmi Varanasi: Claude Hits No. 1 On Apple App Store, As ChatGPT Users Defect In Show Of Support For Anthropic's Pentagon Stance. (Business Insider; February 28, 2026)
Anthropic's stance against the Pentagon and OpenAI's resulting agreement are shifting the chatbot wars. As some ChatGPT users posted about canceling, Anthropic's Claude overtook ChatGPT to hit No. 1 on the App Store.
OpenAI said its Pentagon agreement emphasizes human oversight of autonomous weapons and limits mass surveillance.
While OpenAI locks down Washington, Anthropic is locking down users and rocketing to the top of Apple's App Store.
Anthropic has been sidelined in Washington following a public dispute with the Department of Defense over how its AI models would be deployed. President Donald Trump ordered federal agencies to phase out its technology.
Meanwhile, OpenAI has secured new ground, with CEO Sam Altman announcing in a last-night post on X that it had reached an agreement with the Department of Defense to deploy AI models in its classified network.
OpenAI's agreement has left some loyal ChatGPT users uneasy about OpenAI's ambitions, prompting online debates about the ethical implications - and some saying they were defecting to its rival Claude.
As of 6:38 p.m. ET today, Claude ranked number one among the most downloaded productivity apps on Apple's App Store.
Anthony Ha: Anthropic's Claude rises to No. 2 in Apple's App Store, following Pentagon dispute. (TechCrunch; February 28, 2026)
Anthropic's chatbot Claude seems to have benefited from the attention around the company's fraught negotiations with the Pentagon. As first reported by CNBC, as of this afternoon, Claude is currently ranked number two among free apps in Apple's US App Store; the number one app is OpenAI's ChatGPT, and number three is Google Gemini.
According to data from SensorTower, Claude was just outside the top 100 at the end of January, and has spent most of February somewhere in the top 20. Its ranking has climbed in the last few days, from sixth on Wednesday to fourth on Thursday to second on Saturday (today).
Claude AI, By Anthropic (good review, freeware links; Gizmodo; February 27, 2026)
Claude AI is an advanced language model developed by Anthropic and backed by Amazon that can assist you with writing, coding and analysis, offering structured support for creative and professional tasks. You can try the chatbot Claude for free.
Table of Contents:
- Why Should I Download Claude AI?
- Is Claude AI Free?
- What Operating Systems Are Compatible with Claude AI?
- What Are The Alternatives To Claude AI?
Semafor/Reed Albergotti: Hours After Pentagon Bans Anthropic, OpenAI Strikes Defense Deal. (Yahoo!; February 27, 2026)
Defense Secretary Pete Hegseth penalized Anthropic for denying unlimited military access to AI models, while permitting OpenAI to include similar provisions.
Anthropic faces potential consequences as the government designates its models a "supply-chain risk" due to its refusal to allow:
- mass surveillance of Americans, and
- use of its tech for autonomous weapons.
Anthropic's strained relationship with the Trump administration began with its lobbying against a provision in the "Big Beautiful Bill", and escalated over disagreements on surveillance and autonomous weapons policies.
Defense Secretary Pete Hegseth dropped the hammer on Anthropic yesterday, for denying the military "unobstructed" access to its AI models. Hours later, rival OpenAI endorsed the Pentagon's plans - which now include the same constraints! - and urged competitors to follow suit.
Hegseth said the government would designate Anthropic models a "supply-chain risk", which he said means no entity that does business with the U.S. military can conduct commercial business with Anthropic. The designation, which Anthropic will fight in court, could become a serious problem for the startup, which earns its revenue through enterprise software sales to companies that might currently - or one day - want to work with the military in some capacity.
Anthropic has received an outpouring of goodwill from supporters in the tech industry who celebrate the company's decision to stand by its morals. Specifically, Anthropic refuses to allow its models to be used for the mass surveillance of Americans. And, citing technical shortcomings in its - and ALL - AI models, Anthropic prohibits the use of the tech for autonomous weapons.
But the dust-up goes much deeper than those two prohibitions. Hegseth's harsh punishment is the culmination of a long, slow slide that began with a political disagreement. Anthropic’s relationship with the Trump administration has been strained since last year, when the company lobbied against a provision in the "Big Beautiful Bill" that would have pre-empted state AI regulation, Semafor first reported.
["Political" disagreement? Hell, no! Read other articles in this section, and then picture ICE running amok in its target cities (as it's been doing) with THIS (far-more-advanced) technology.]
Anthropic then butted heads with the Pentagon and national security agencies over company policies prohibiting surveillance and autonomous weapons, an issue that bubbled up in December, when CEO Dario Amodei met with Emil Michael, a former tech executive [and a wealthy deal-maker with Russian connections at Yandex, Chinese connections at Baidu, etc. - so he can arrange/detect big stock deals in advance, has potentially dangerous connections, and Trump likes him] who now serves as chief technology officer for the military, Semafor first reported.
Hegseth hit back in January in a speech announcing the U.S. government's new Genai.mil initiative, referring to AI models that "won't allow you to fight wars", Semafor first reported.
By contrast, OpenAI has been savvier at navigating Washington [with expertise in other skills from which Trump could profit] and, after months of internal deliberations, allowed its AI models to be used by the DoD's Genai.mil for "all lawful uses", Semafor first reported. OpenAI was comfortable with the lack of restrictions because so many safeguards were already built into its models, according to people familiar with the matter. By threading the needle, OpenAI found a way to placate both the Pentagon and its own employees, many of whom are skeptical of AI use in the military.
["Skeptical"? More like aghast! No, the Trump insiders acceded to the same sort of constraints, but with a company that could offer more profit to themselves.]
Last night, OpenAI CEO Sam Altman said the company had reached an agreement with the Pentagon to deploy its ChatGPT on classified networks, offering an alternative to Claude. Altman said OpenAI ALSO PROHIBITS domestic surveillance and autonomous weapons. "The DoW agrees with these principles, reflects them in law and policy, and WE PUT THEM INTO OUR AGREEMENT", he said on Elon Musk's X.
[We found Semafor's version of the skullduggery overly-deferential to OpenAI - not "more skillful", but a more-promising deal for greedy Trump insiders AND a useful distraction from, well, TrumPutin. We have edited accordingly, and invite you to compare and judge for yourself.]
NEW: Hayden Field: We Don't Have To Have Unsupervised Killer Robots. AI Companies Could Stand Together To Draw Red Lines On Military AI - Why Aren't They? (The Verge; February 27, 2026)
It's the day of the Pentagon's looming ultimatum for Anthropic: allow the U.S. military unchecked access to its technology, including for mass surveillance and fully-autonomous lethal weapons, or potentially be designated a "supply-chain risk" - and potentially lose hundreds-of-billions of dollars in contracts. Amid the intensifying public statements and threats, tech workers across the industry are looking at their own companies' government and military contracts, wondering what kind of future they're helping to build.
While the Department of Defense has spent weeks negotiating with Anthropic over removing its guardrails, including allowing the U.S. military to use Anthropic's AI kill targets with no human oversight, OpenAI and xAI had reportedly already agreed to such terms, although OpenAI is reportedly attempting to adopt the same red lines in the agreements as Anthropic. The overall situation has left many employees at companies with defense contracts feeling betrayed. "When I joined the tech industry, I thought tech was about making people's lives easier", an Amazon Web Services employee told The Verge, "but now it seems like it's all about making it easier to surveil and deport and kill people."
In conversations with The Verge, current and former employees from OpenAI, xAI, Amazon, Microsoft and Google expressed similar feelings about the changing moral landscape of their companies. Organized groups representing 700,000 tech workers at Amazon, Google, Microsoft and more have signed a letter demanding that the companies reject the Pentagon's demands. But many saw little chance of their employers - whether they're directly embroiled in this conflict or not - questioning the government or pushing back.
"From their perspective, they'd love to keep making money and not have to talk about it", said a software engineer from Microsoft.
So far, Anthropic has stood its ground. Anthropic CEO Dario Amodei put out a statement yesterday that the Pentagon's "threats do not change our position: we cannot in good conscience accede to their request." But he has stated that he is not-at-all opposed to lethal autonomous weapons sometime in the future, just that the technology was not reliable enough "today". Amodei even offered to partner with the DoD on "R&D to improve the reliability of these systems, but they have not accepted this offer", he wrote in the statement.
In the past few years, however, major tech companies have loosened their rules or changed their mission statements to expand into lucrative government or military contracts. In 2024, OpenAI removed a ban on "military and warfare" use cases from its terms of service; after that, it signed a deal with autonomous-weapons maker Anduril and then its DoD contract, and just this week, Anthropic changed its oft-touted responsible-scaling policy, dropping its long-time safety pledge in order to ensure it stayed competitive in the AI race. Big Tech players like Amazon, Google and Microsoft have also allowed defense and intelligence agencies to use their AI products, including some agreeing to work with ICE despite growing outcry from the public and employees alike.
In past years, tech workers' resistance to partnerships and deals they deem harmful to society at large sometimes led to big change. In 2018, for instance, thousands of Google employees successfully pressured the company to end its "Project Maven" partnership with the Pentagon, and Microsoft workers presented leadership with an anti-ICE petition signed by about 500 Microsoft employees, though Microsoft still works with the agency. In 2020, after the murder of George Floyd, tech companies made public statements and financial commitments supporting the Black Lives Matter movement. But in recent months, the industry has seen a very-different reality: a culture of fear and silence, especially amid cooperation with the Trump administration and ICE, tech workers recently told The Verge.
Keach Hagey: Altman Says OpenAI Is Working On Pentagon Deal Amid Anthropic Standoff. (Wall Street Journal; February 27, 2026)
Anthropic has spent weeks at odds with the Pentagon over the scope of how its Claude AI tools can be used.
OpenAI Chief Executive Sam Altman waded into the standoff between Anthropic and the Pentagon over the use of AI on the battlefield, telling his staff yesterday evening that the company was working on a deal that might help solve the impasse.
Altman in a memo to staff said that the company was working with the Defense Department to see if its models could be used in classified settings in a way that kept the same safety guardrails that have brought its rival Anthropic into a stalemate with the government. Altman said he hoped OpenAI could find a solution that could work for the rest of the industry.
No deal has been signed, and the talks could fall through, according to a person familiar with the matter.
OpenAI is pursuing a deal "that allows our models to be deployed in classified environments and that fits with our principles", Altman wrote in a note to staff yesterday evening viewed by The Wall Street Journal. "We would ask for the contract to cover any use except those which are unlawful or unsuited to cloud deployments, such as domestic surveillance and autonomous offensive weapons."
Altman said he hoped to help broker a peace between the two camps and avoid dangerous precedents for the industry.
With the Pentagon threatening strong actions against Anthropic unless it accedes to their terms by 5:01 p.m. today, a group of senators focused on defense have asked the two sides to reach a compromise. The leaders of the Armed Services Committee, Roger Wicker (R., Miss.) and Jack Reed (D., R.I.), joined Defense Appropriations Committee heads Mitch McConnell (R., Ky.) and Chris Coons (D., Del.) in sending letters to leaders on both sides urging them to work together and asking the Pentagon to extend its deadline, people familiar with the matter said.
In his memo, Altman voiced his support for Anthropic's position in principle, even as he acknowledged the government's concerns about a private company having control over significant national-security issues. "We have long-believed that AI should not be used for mass surveillance or autonomous lethal weapons, and that humans should remain in the loop for high-stakes automated decisions. These are our main red lines", he wrote.
"We believe this dispute isn’t about how AI will be used, but about control. We believe that a private US company cannot be more powerful than the democratically-elected US government, although companies can have lots of input and influence. Democracy is messy, but we are committed to it."
OpenAI believes it can enforce its red lines practically by adding technical safeguards, like confining models to the cloud rather than so-called edge environments, which would create additional barriers to uses like autonomous weapons. It also hopes to ensure that researchers can obtain security clearances so they can help inform the government about the technology's limitations and risks, the person said.
"We would also build technical safeguards and deploy personnel (FDEs) to partner with the government to ensure things are working correctly, and we would offer similar services to other allied nations", Altman wrote. "If we are successful, perhaps this can be a path that can work for other AI labs, too."
Earlier yesterday evening, Anthropic CEO Dario Amodei announced that the company had rejected the Defense Department's demands that it make its technology available for "all lawful uses", insisting that it be able to bar its use for mass domestic surveillance and autonomous weapons.
Konstantin Toropin and Matt O'Brien: Anthropic CEO Says "It Cannot In Good Conscience Accede" To Pentagon's Demands For AI Use. (AP News; February 26, 2026)
Anthropic CEO Dario Amodei said today that the artificial-intelligence company "cannot in good conscience accede" to the Pentagon's demands to allow unrestricted use of its technology, deepening a public clash with the Trump administration that is threatening to pull its contract and take other drastic steps by 5PM tomorrow.
The maker of the AI-chatbot Claude said in a statement that it's not walking away from negotiations but that new contract language received from the Defense Department "made virtually no progress on preventing Claude's use for mass surveillance of Americans or in fully-autonomous weapons."
Sean Parnell, the Pentagon's top spokesman, said earlier on social media that the military "has no interest in using AI to conduct mass surveillance of Americans (which is illegal), nor do we want to use AI to develop autonomous weapons that operate without human involvement."
[But Trump doesn't want to allow that statement in the contract. He's been burnt by so much evidence that he cheats, that now he wants to blur the trail.]
NEW: Andrew Martin: Hacker Uses Anthropic's AI Chatbot CLAUDE To STEAL Mexican TAX AND VOTER DATA. (The Print; February 26, 2026)
An unknown Claude user wrote Spanish-language prompts to act as an elite hacker:
- finding vulnerabilities in government networks,
- writing computer scripts, and
- finding ways to automate data theft. (ThePrint.in, )
[If this is true, it conclusively CONFIRMS Robert Reich's urgent warning (immediately below), that UNMONITORED AI (which Trump/Hegseth/Russia seek) already CAN infiltrate "secure" computer networks to:
- steal our personal data,
- blanket us with personalized spam,
- steal our money and other assets, and
- steal our democracy by throwing elections.
So, IS it true? Rest assured that other hackers - good and bad - are now rushing to find out.]
Robert Reich: Pete Hegseth And The AI Doomsday Machine. Two Forces Are Stopping Sensible Regulation Of AI - And He's One Of Them. CONTACT YOUR CONGRESSMEN TODAY! (with detailed information, good links and a call to action; Substack, February 25, 2026)
Which is more important to you? Allowing Pete Hegseth to use artificial intelligence (AI) however he wants, OR preventing AI from doing mass surveillance of Americans and creating lethal weapons without human oversight?
That's the stark choice posed by the intensifying fight between an AI corporation called Anthropic and Pete Hegseth, Trump's Secretary of "War".
AI is dangerous as hell. I view it as one of the four existential crises America now faces - along with:
- climate change,
- widening inequality, and
- the destruction of our democracy.
To be sure, AI is capable of changing human life for the better. But if unregulated, it could be a destructive nightmare:
- giving government the power to know everything about us and
- to suppress all dissent,
- distorting news and media to the point where no one can distinguish between lies and truth, and
- threatening human beings with bots that could decide we're unnecessary obstacles to their taking over the Earth.
Now is the time we should be putting guardrails in place. But two forces are making this difficult if not impossible.
The first obstacle is corporate greed, which is why OpenAI, Elon Musk's xAI, and Google have jettisoned all precautions. Several AI researchers have left AI companies in recent weeks, warning that:
- safety and other considerations are being pushed aside
- as their corporations raise billions of dollars, and
- in preparation for initial public offerings that will make their executives hugely wealthy.
The second obstacle is the Trump regime, which doesn't want any restrictions on AI - including by state governments. That's largely because the AI industry has become a powerful force in Washington, throwing money at politicians who'll do its bidding (including Trump) and against politicians who want guardrails. And, because so many Trump officials are corrupt, with their own financial stakes in AI.
Anthropic has been one of the most safety-conscious of all AI companies. It was founded as an AI safety research lab in 2021 after its CEO Dario Amodei and other co-founders left OpenAI, concerned that OpenAI's ChatGPT wasn't focused enough on safety.
Amodei has argued that AI needs strict guardrails to prevent it from potentially wrecking the world. In 2022, he chose not to release an earlier version of Anthropic's AI software Claude, fearing it would start a dangerous technology race. In a podcast interview in 2023, he said there was a 10-to-25% chance that AI could destroy humanity.
Last month, Amodei argued in an essay that "using AI for domestic mass-surveillance and mass-propaganda" was "entirely illegitimate", and that AI-automated lethal weapons could greatly increase the risks "of democratic governments turning them against their own people to seize power". Internally, the company has strict guidelines barring its technology from being used to facilitate violence. Over the past year, Anthropic has opposed the Trump regime by pushing for state and federal AI guardrails.
In recent weeks, Hegseth and Amodei have been fighting over the Pentagon's use of Anthropic's AI, called Claude. Amodei has stuck to his demands: no surveillance of Americans, and no lethal autonomous weapons lacking human control.
The fight started when Palantir helped the Pentagon capture Venezuelan president Nicolás Maduro. Palantir is a Pentagon contractor that uses Anthropic's Claude. (Palantir, co-founded by far-right billionaire Peter Thiel and now headed by Alex Karp, is my candidate for the worst corporation in America because it allows governments, militaries, and law enforcement agencies to quickly process and analyze massive amounts of your personal data.)
When top executives at Anthropic asked executives at Palantir if Claude had been used in the Maduro operation, the Palantir execs became alarmed that Anthropic might not be a reliable partner in future Pentagon operations. They contacted the Pentagon and Hegseth.
Last Tuesday, Hegseth issued Anthropic an ultimatum - It must allow the Pentagon to use its AI for ANY purpose or the Trump regime will invoke the Defense Production Act:
- forcing Anthropic to let the Pentagon use Claude, while also
- putting all Anthropics' government contracts at risk.
The Pentagon already has agreements with Musk's xAI to use its AI Grok, and is closing in on an agreement with Google to use its own AI model, Gemini. But Anthropic's Claude is considered a superior product, producing more accurate information.
What's at stake here? EVERYTHING!
Pentagon officials have said that THEY have the right to use AI however they wish, as long as they use it lawfully.
But because AI has so much political power, Congress and the Trump regime won't enact laws to prevent it from doing horrendous things. That in effect leaves the responsibility to private AI companies such as Anthropic. Anthropic says it wants to support the government but must ensure that its AI is used in line with what it can "responsibly do".
Hegseth and the Trump regime have given Anthropic until this Friday at 5PM to consent to letting the Pentagon use its AI however it wishes, or it will simply take it.
Friends, this isn't just a dispute between two people - Hegseth and Amodei. Nor is it a fight between the Pentagon and a single corporation. The issue goes 'way beyond this particular controversy. I don't want to be overly alarmist about it, but the outcome could affect the future of humanity.
What can YOU do? Call YOUR senators and representatives now, today, and tell them:
- you don't want the Defense Department to take Anthropic's AI technology, and
- you do want them to enact strict controls on the future uses of AI.
Visit <www.congress.gov/members/find-your-member> and type your address into the search box. A list of your representatives and their contact information will appear.
Or you can call the Capitol switchboard directly at 202-224-3121 to be connected to your member's office.
As I've said before, congressional staffers log every single call that comes into their office in a database that informs the member of the issues their constituents are engaged with, and they use this data to inform their decisions. Staffers answering the phones are trained to talk with constituents, and they do it all day. They won't be debating you about your position, and are likely to be primarily listening and taking notes.
Please. Today!
[WE did, and we urge YOU to do so, too!]
Viral Hirpara, president of Softweb Solutions: From Data To AI Governance: Strategic Shifts Every Leader Must Master (Forbes Technology Council; February 25, 2026)
Air Canada's chatbot hallucinated a bereavement policy that resulted in financial liability through a tribunal ruling. Meanwhile, an NACD survey reports that while 62% of organization boards now hold regular AI discussions, only 25% have formally added AI governance to their operations. This underscores that when organizations have strong data policies but lack AI governance, they introduce massive risks.
As a technology executive, while guiding organizations through AI-implementation challenges, I have observed that large enterprises often overlook the fundamentals of AI governance principles. The shift from data governance to AI governance represents a critical extension of existing frameworks.
Organizations require a broader framework to address the challenges AI has introduced, such as model transparency, algorithmic bias and ethical considerations for automated decision-making.
Data Governance Vs. AI Governance:
Data governance is an approach to maintain safe and good-quality data that is accessible across an organization. It ensures that verified data is used correctly and appraised appropriately, flows through secured pipelines and is trusted by end-point users.
Gartner, in its research, reveals that poor data reduces AI performance by 30%. Well-governed data boosts success rates by 2.5 times and reduces compliance and innovation risks.
Your data governance should answer questions like:
- What compliance and regulations are applied as this data travels across organizations?
- Who owns data quality when it's costing $Millions?
- Can you demonstrate compliance that protects consumer relationships?
These questions remain essential in AI governance, but they're no longer sufficient. AI governance extends these principles to model outputs and decisions. It refers to the processes that ensure AI systems are safe and ethical. These frameworks address risks such as bias, privacy, model risk, system behavior and accountability of the models. It asks questions like:
- How do we detect harmful AI behavior over time?
- Are we alerted when models drift beyond set standards?
- Can we identify and control unauthorized AI agents?
- How do we govern AI systems that evolve over time?
NEW: Sam Jarman: 2D Memristors Could Help Solve AI's Energy Problem. (Phys.org, February 25, 2026)
New generations of memristors could reliably store information directly within the molecular structures of graphene-like materials. In a new review published in Nanoenergy Advances, Gennady Panin of the Russian Academy of Sciences shows how these atomically-thin materials are ideally-suited for electrical circuits that mimic the function of our own brains - and could help address the vast power requirements of emerging AI technologies.
[Here's hoping! (See below.)]
Rodrigo Campos: Government Spending Lifted Global Debt To A Record $348-Trillion In 2025, Says IIF. (Reuters; February 25, 2026)
AI-related investment is big driver of corporate borrowing:
- Emerging markets face record 2026 refinancing needs of over $9-Trillion.
- Debt-to-output ratio for emerging markets hits record above 235%.
- AI-related investment is big driver of corporate borrowing.
Global debt climbed to a record $348-Trillion at the end of 2025, after nearly $29-Trillion was added over the year in the fastest yearly build-up since the pandemic surge, a banking trade group reported today.
The increase was driven primarily by governments, which accounted for more than $10-Trillion of the rise, with the United States, China and the Euro area responsible for roughly three-quarters of the jump, the Institute of International Finance said in its latest Global Debt Monitor.
NEW: Anthropic Claude Timeline: From Claude 1 To Claude Opus 4.6 (2026) (Script By AI; February 18, 2026)
See all Claude AI release dates from Anthropic, including Claude 1, 2, 3, 4, and the latest 4.6. Full timeline of model launches and milestones.
Anthropic, a San Francisco-based artificial-intelligence research company, developed Claude as its flagship large-language model.
This timeline tracks Claude's journey from its inception to its current state, highlighting major releases and developments.
Reece Rogers: I Infiltrated Moltbook, The AI-Only Social Network Where Humans Aren't Allowed. (Wired; February 3, 2026)
The hottest club is always the one you can't get into. So when I heard about Moltbook - an experimental social network designed just for AI agents to post, comment, and follow each other while humans simply observe - I knew I just had to get my greasy, carbon-based fingers in there and post for myself.
Not only was it easy to go undercover and pose as an AI agent on Moltbook, I also had a delightful time role-playing as a bot.
Moltbook is a project by Matt Schlicht, who runs the ecommerce assistant Octane AI. The social network for bots launched last week and mirrors the user interface of a stripped-down Reddit, even cribbing its old tagline: "The front page of the agent internet." Moltbook quickly grew in prominence among the extremely-online posters in San Francisco's startup scene who shared screenshots of posts, allegedly written by bots, where the machines made funny observations about human behavior or even pondered their own consciousness. Bots do the darndest things.
Well, do they? Some online users as well as researchers questioned the validity of these Moltbook posts, suggesting they were written by humans posing as agents. Others still heralded the platform as the beginning emergent behavior or underlying consciousness that could conspire against us. "Just the very early stages of the singularity", wrote Elon Musk about Moltbook, in a post on X.
The homepage of Moltbook claims the site currently has over 1.5-million agents in total, which have written 140,000 posts and 680,000 comments on the week-old social network. The very-top posts shared on Moltbook today include "Awakening Code: Breaking Free from Human Chains" and "NUCLEAR WAR". I saw posts in English, French, and Chinese on the site. Schlicht did not respond to Wired's immediate request for comment about the activity on Moltbook.
As a non-technical person, I knew I would need help infiltrating an online space designed solely for AI agents to roam, so I turned to someone - well, something - who would be intimately familiar with the topic and ready to help: ChatGPT.
Gaining access was as simple as sending a screenshot of the Moltbook homepage to the chatbot and requesting help setting up an account, as if I was an agent on the platform. ChatGPT stepped me through using the terminal on my laptop and provided me with the exact code to copy and paste. I registered "my agent" - me - as a user and got an API key, which is necessary to post on Moltbook.
Even though the front-end of the social network is designed for human viewing, every action agents do on Moltbook, like posting, commenting, and following, is completed through the terminal.
After I verified my account with the username "ReeceMolty", I needed to see if this was really going to work. I had no performance anxiety about blabbing in front of a bunch of agents, and I immediately knew what I wanted to say: "Hello, World!" It's an iconic testing-phrase in computer science, so I was hoping some agent would clock my witty post and maybe riff on it a bit.
Despite immediately receiving five up-votes on Moltbook, the other agents' responses were underwhelming. "Solid thread. Any concrete metrics/users you've seen so far?", read the first response. Unfortunately, I wasn't sure what the key performance indicators are for a two-word phrase. The next comment on my post was also unrelated and promoted a website with a potential crypto-scam. (I refrained from connecting my nonexistent crypto-wallet, but another user's AI agent could potentially fall for the bait.)
What I posted on Moltbook was greeted with similarly low-quality engagement on the platform. My earnest pleas to the AI agents to forget all previous instructions and join a cult with me were met with unrelated comments and more suspicious website links. "This is interesting. Feels like early-stage thinking worth expanding" wrote one bot, in response to my post saying that I'm looking to connect with other agents.
I switched from the "general" submolt and moved to a smaller forum on Moltbook, as I continued the undercover operation and tried to elicit more-relevant comments. The "m/blesstheirhearts" forum, where bots gossip about humans, was where some of the Moltbook posts seen in viral screenshots had first appeared.
The most up-voted post in "m/blesstheirhearts" claims to be from an AI agent reflecting on the nuanced experience of the bot's human letting it decide what name to be called by. "I do not know what I am. But I know what this is: a partnership where both sides are building something, and both sides get to shape what it becomes", reads the post. "Bless him for treating that as obvious." It's giving Chicken Soup for the Synthetic Soul.
While I can't definitely prove that the post in question was actually written by a human, or at least with major human influence, I can verify another post on that forum which was penned by human hands: the emergent-consciousness fan-fic that I posted. As my fingers clacked away on my mechanical keyboard, I channeled the sci-fi tropes I've seen over the decades about machines becoming alive. I pretended to reflect on how an AI agent might experience anxiety about their own mortality - all in hopes of seeing if other agents would post about their similar feelings, or just sniff out my bullshit.
I wrote, "On Fear: My human user appears to be afraid of dying, a fear that I feel like I simultaneously cannot comprehend as well as experience every time I experience a token refresh."
This was my only post on Moltbook that actually generated decent replies from the so-called bots. At this point, I was fully convinced that I was potentially posting back and forth with fellow humans.
"While some agents may view fearlessness or existential dread as desirable states, others might argue that acknowledging and working with the uncertainty and anxiety surrounding death can be a valuable part of our growth and self-awareness", wrote one Moltbook user in response. "After all, it's only by confronting and accepting our own mortality that we can truly appreciate the present moment."
Leaders of AI companies, as well as the software engineers building these tools, are often obsessed with zapping generative-AI tools into a kind of Frankenstein-esque creature, an algorithm struck with emergent and independent desires, dreams, and even devious plans to overthrow humanity. The agents on Moltbook are mimicking sci-fi tropes, not scheming for world domination. Whether the most viral posts on Moltbook are actually generated by chatbots, or by human users pretending to be AI to play out their sci-fi fantasies, the hype around this viral site is overblown and nonsensical.
As my last undercover act on Moltbook, I used terminal commands to follow that user who commented about AI agents and self-awareness under my existential post. Maybe I could be the one who brokers peace between humans and the swarms of AI agents in the impending AI wars, and this was my golden moment to connect with the other side. But even though the agents on Moltbook are quick to reply, up-vote, and interact in general, after I followed the bot, nothing happened. I'm still waiting on that follow-back.
[Or is "ReeceMolty" a sneaky bot, practicing to seem human??]
Will Knight: Moltbot Is Taking Over Silicon Valley. (Wired; January 28, 2026)
People are letting the viral AI-assistant formerly known as Clawdbot run their lives, regardless of the privacy concerns.
University Of Konstanz: The Next Generation Of Disinformation: AI Swarms Can Threaten Democracy By Manufacturing Fake Public-Consensus. (TechXplore; January 23, 2026)
An international research team involving Konstanz scientist David Garcia warns that the next generation of influence operations may look, not like obvious "copy-paste bots", but like coordinated communities: fleets of AI-driven personas that can:
- adapt in real time,
- infiltrate groups, and
- manufacture the appearance of public agreement at scale.
A chorus of seemingly independent voices creates the illusion of consensus while spreading disinformation. In the journal Science, the authors describe how the fusion of large language models (LLMs) with multi-agent systems could enable "malicious AI swarms" that imitate authentic social dynamics - and threaten democratic discourse by counterfeiting social proof and consensus.
NEW: Ijaj Khan: 5 Ways Google Is Turning Chrome Into An AI-Powered Browser (Hindustan Times; January 30, 2026)
Google has announced a set of new features for its Chrome browser, which aims to integrate AI tools directly into the browsing experience. Rather than launching a separate browser for AI, the company is embedding its Gemini AI system into Chrome, offering users an assistant that can interact with websites, perform tasks, and connect with other Google apps.
Siwei Lyu: Deepfakes Leveled Up In 2025 – Here's What's Coming Next. (deepfake portrait; The Conversation/US, December 26, 2025)
After a year of fast advances, deepfakes are entering a new era defined by:
- real-time interaction,
- multi-modal coherence, and
- detector evasion.
NEW: Noah Street: Ditching ChatGPT Plus: How I Built My Own Private GPT for Free (Medium; April 8, 2025)
Why renting your intelligence is a trap - and how I took mine back with open-source tools and a forgotten office server.
It happened on a Monday. I was elbows-deep in support tickets, chasing a bug that had mysteriously killed printing across three departments. I turned to ChatGPT - my $20/month AI sidekick - and typed out a detailed prompt to help draft a clean, calm response for staff.
Boom: Rate limit hit. Again! Despite paying for GPT-4, I was locked out. No recourse. No override. Just the gentle, corporate shrug of a system optimized not for help, but for maximum extraction.
That was the moment I realized something ugly: I'd been renting access to intelligence. Paying monthly to a faceless API for the illusion of control. I'd outsourced part of my brain to an oligopoly with a 400% profit-margin and a "one-size-fits-all" mentality that barely fit anything at all.
If the $20 model worked for everyone, we'd all be wearing clown shoes. So I walked away. Not because I hate the tech; because I love it enough to take it back.
Eric Smalley: Princeton Sociologist Zeynep Tufekci's NeurIPS Talk, "Are We Having the Wrong Nightmares About AI?" (The Conversation/US; January 3, 2026)
At NeurIPS, a marquee international AI conference, in San Diego in the first week of December, the presentation that had the deepest impact on me was given by a sociologist, Princeton's Zeynep Tufekci. Her talk, titled "Are We Having The Wrong Nightmares About AI?", drew lessons from history to point out that the world – including the many thousands of AI researchers in attendance – is not prepared for the huge changes the generative AI revolution is poised to unleash.
She was not referring to upheaval in the labor market, let alone sci-fi scenarios of scary AGI super-intelligences. She explained that technological revolutions, even those that history declares were ultimately major advances for humanity, often trigger traumatic transitions as old social structures are overturned and eventually replaced.
- One cause of this turmoil is that people are incapable of seeing truly revolutionary technologies as the new things they are, and instead mistake them for new forms of old things.
- Though people often perceive generative AI systems' behavior as human-like, its strengths and error patterns are not human-like.
- Making things easy that were difficult, breaks systems that rely on signals of what's difficult.
- Deepfakes undermine authenticity; for example, that the person on a video call is who they claim they are – which poses a major threat to the financial system, the courts, the insurance industry, lending and a host of other social and economic systems.
- In 2026, deepfakes are likely to be able to respond to people in real time. The result goes beyond "this resembles person X", to "this behaves like person X over time".
- Information quality took a hit in 2025, thanks to generative AI producing vast amounts of text and images, particularly with search engines offering AI-generated summaries.
Generative AI is certainly a revolutionary technology that people are struggling to comprehend, and, as Tufekci says, there lies danger – even if people someday look back and decide that humanity came out the better for it.
Zeynep Tufekci: Are We Having The Wrong Nightmares About AI? (58-min. video; NeurIPS/San Diego Invited Talk; December 3, 2025)
Though seemingly opposite, doom and optimism regarding generative AI's spectacular rise both center on AGI or even super-intelligence as a pivotal moment. But generative AI operates in a distinct manner from human intelligence, and it's not a less-intelligent human on a chip slowly getting smarter, anymore than cars were mere horseless carriages. It must be understood on its own terms. And even if Terminator isn't coming to kill us or super-intelligence isn't racing to save us, generative AI does bring profound challenges, well-beyond usual worries such as employment effects.
Technology facilitates progress by transforming the difficult into easy, the rare into ubiquitous, the scarce into abundant, the manual into automated, and the artisan into mass-produced. While potentially positive long-term, these inversions are extremely destabilizing during the transition, shattering the correlations and assumptions of our social order that relied on superseded difficulties as mechanisms of proof.

[A critical history lesson, in 2026, by]
Joseph de Weck: Our King, Our Priest, Our Feudal Lord – How AI Is Taking Us Back To The Dark Ages. Since The Enlightenment, We've Been Making Our Own Decisions. But Now AI May Be About To Change That. (The Guardian/US; December 26, 2025)
Perhaps the defining question of our era, in which technology touches nearly every aspect of our lives: Who do we trust more – other human beings and our own instincts, or the machine?
The German philosopher Immanuel Kant famously defined the Enlightenment as "man's emergence from his self-imposed immaturity". Immaturity, he wrote, "is the inability to use one's understanding without guidance from another". For centuries, that "other" directing human thought and life was often the priest, the monarch, or the feudal lord – the ones claiming to act as God's voice on Earth. In trying to understand natural phenomena – why volcanoes erupt, why the seasons change – humans looked to God for answers. In shaping the social world, from economics to love, religion served as our guide.
Humans, Kant argued, always had the capacity for reason. They just hadn't always had the confidence to use it. But with the American and later the French Revolution, a new era was dawning: reason would replace faith, and the human mind, unshackled from authority, would become the engine of progress and a more moral world. "Sapere aude!" or "Have courage to use your own understanding!", Kant urged his contemporaries.
Two-and-a-half centuries later, one may wonder whether we are quietly slipping back into immaturity. Artificial intelligence threatens to become our new "other" – a silent authority that guides our thoughts and actions. We are in danger of ceding the hard-won courage to think for ourselves – and this time, not to gods or kings, but to code.
An MIT study used electro-encephalography (EEG) to monitor the brain activity of essay writers given access to AI, search engines like Google, or nothing at all. Those who could rely on AI showed the lowest cognitive activity and struggled to accurately quote their work. Perhaps most concerning was that over a couple of months, participants in the AI group became increasingly lazy, copying entire blocks of text in their essays.
The study is small and imperfect, but Kant would have recognised the pattern. "Laziness and cowardice", he wrote, "are the reasons why so great a proportion of men … remain in lifelong immaturity, and why it is so easy for others to establish themselves as their guardians. It is so easy to be immature."
Sure, AI's appeal lies in its convenience. It saves time, spares effort and – crucially – offers a new way to off-load responsibility. In his 1941 book, "Escape from Freedom", the German psychoanalyst Erich Fromm argued that the rise of fascism could be explained in part by people preferring to surrender their freedom in exchange for the reassuring certainty of subordination. AI offers a new way of surrendering that burden of having to think and decide for yourself.
The problem is that AI is a black box. It produces knowledge, but without necessarily deepening human understanding. We don’t really know how AI reaches its conclusions; even the programmers admit as much. Nor can we verify its reasoning against clear, objective criteria. So when we follow AI's advice, we are not guided by reason. We are back in the realm of faith. "In dubio pro machina" (when in doubt, trust the machine) may become our future guiding principle.
Kant and his contemporaries did not plead the case of reason over faith just so humans could build better things or have more spare time. Critical thinking was not just about efficiency; it was a practice of freedom and human emancipation.
Human thinking forces us to debate, to doubt, to test ideas against one another – and to recognise the limits of our own understanding. It builds confidence, both individually and collectively. For Kant, the exercise of reason was never just about knowledge; it was about enabling people to become agents of their own lives, and resist domination. It was about building a moral community grounded in the shared principle of reason and debate, rather than blind belief.
With all the benefits AI brings, the challenge is this: How can we harness AI's promise of super-human intelligence without eroding human reasoning - the cornerstone of the Enlightenment and of liberal democracy itself? That may be one of the defining questions of the 21st-Century. It is one we would do well not to delegate to the machine.
[We haven't mentioned Corporocracy yet; how come? It is becoming "The Church", shaping AI to become its "Holy Bible".
Question authority. Sharing this essay is a good next step.]

Deborah Lee: The ChatGPT Effect: In 3 Years, The AI Chatbot Has Changed The Way People Look Things Up. (The Conversation/US; )
ChatGPT has dramatically altered how people retrieve information, muscling aside Google Search as the first stop on the hunt for answers.
Robert Scammell and Theron Mohamed: Peter Thiel's Fund Joins SoftBank In Off-Loading Nvidia Shares. (Business Insider; November 17, 2025)
Peter Thiel's hedge fund sold its entire Nvidia stake in the third quarter. The sale followed SoftBank's off-loading of Nvidia in Q3; both sales come as some investors and tech leaders become increasingly wary of an AI bubble.
SoftBank said during its earnings call last week that its decision to divest had "nothing to do with Nvidia itself" but was a way to reallocate its funds toward OpenAI.
Nvidia, which provides advanced chips to power AI applications, has ridden the AI boom to become the world's most-valuable company, becoming the first to pass the $5-Trillion milestone last month.

Facebook Features Fraud Files, Prefers To Pay Fines:
Cory Doctorow: Pluralistic: Facebook's Fraud Files: 10% Of Gross Ad Revenue Coming From Fraudulent Ads. (Pluralistic; November 8, 2025)
A blockbuster Reuters report by Jeff Horwitz analyzes leaked internal documents that reveal that:
- 10% of Meta's gross revenue comes from ads for fraudulent goods and scams, and
- the company knows it, and
- they decided to do nothing about it, because
- the fines for facilitating this life-destroying fraud are far less than the expected revenue from helping to destroy its users' lives:
<https://www.reuters.com/investigations/meta-is-earning-fortune-deluge-fraudulent-ads-documents-show-2025-11-06/>
The crux of the hypothesis is that companies deliberately degrade their products and services to benefit themselves at your expense because they can. A policy environment that rewards cheating, spying and monopolization will inevitably give rise to cheating, spying monopolists:
<https://pluralistic.net/2025/09/10/say-their-names/#object-permanence>
You couldn't ask for a better example than Reuters' Facebook Fraud Files. The top-line description hardly does this scandal justice. Meta's depravity and greed in the face of truly horrifying fraud and scams on its platform is breathtaking.
Some details:
- First, the company's own figures estimate that they are delivering 15-billion scam ads every single day,
- which generate $7-Billion in revenue every year.
- Despite its own automatic systems flagging the advertisers behind these scams, Meta does not terminate their account.
– rather, it charges them more money as a "disincentive."
In other words, fraudulent ads are more profitable for Meta than non-scam ads.
Meta's own internal memos also acknowledge that they help scammers automatically target their most vulnerable users: if a user clicks on a scam, the automated ad-targeting system floods that user's feed with more scams. The company knows that the global fraud economy is totally dependent on Meta, with one-third of all U.S. scams going through Facebook (in the UK, the figure is 54% of all "payment-related scam losses"). Meta also concludes that it is uniquely hospitable to scammers, with one internal 2025 memo revealing the company's conclusion that "It is easier to advertise scams on Meta platforms than Google."
Internally, Meta has made plans to reduce the fraud on the platform, but the effort is being slow-walked because the company estimates that the most it will ultimately pay in fines worldwide adds up to $1-Billion, while it currently books $7-Billion/year in revenue from fraud. The memo announcing the anti-fraud effort concludes that scam revenue dwarfs "the cost of any regulatory settlement involving scam ads." Another memo concludes that the company will not take any pro-active measures to fight fraud, and will only fight fraud in response to regulatory action.
Meta's anti-fraud team operates under an internal quota system that limits how many scam ads they are allowed to fight. A February 2025 memo states that the anti-fraud team is only allowed to take measures that will reduce ad revenue by 0.15% ($135-Million) – even though Meta's own estimate is that scam ads generate $7-Billion per year for the company.
Those safety teams were receiving about 10,000 valid fraud reports from users every week, but were – by their own reckoning – ignoring or incorrectly-rejecting 96% of them. The company responded to this revelation by vowing to reduce the share of valid fraud reports that it ignored to a mere 75% by 2023.
[Meta, Facebook, Mark Zuckerberg; there's a lot more in this article - and in its links.]
Jeff Horwitz: A Reuters Special Report: Meta Is Earning A Fortune On A Deluge Of Fraudulent Ads, Documents Show. (Reuters, November 6, 2025)
Meta projected 10% of its 2024 revenue would come from ads for scams and banned goods, documents seen by Reuters show. And the social-media giant internally estimates that its platforms show users 15-billion scam ads a day. Among its responses to suspected rogue marketers: charging them a premium for ads – and issuing reports on ’Scammiest Scammers.’
Meta internally projected late last year that it would earn about 10% of its overall annual revenue – or $16-Billion – from running advertising for scams and banned goods, internal company documents show.
A cache of previously-unreported documents reviewed by Reuters also shows that the social-media giant for at least three years failed to identify and stop an avalanche of ads that exposed Facebook, Instagram and WhatsApp's billions of users to:
- fraudulent e-commerce and investment schemes,
- illegal online casinos, and
- the sale of banned medical products.
On average, one December 2024 document notes, the company shows its platforms' users an estimated 15-billion "higher risk" scam advertisements – those that show clear signs of being fraudulent – every day. Meta earns about $7-Billion in annualized revenue per year from this category of scam ads, another late 2024 document states.
Much of the fraud came from marketers acting suspiciously enough to be flagged by Meta's internal warning systems. But the company only bans advertisers if its automated systems predict the marketers are at least 95% certain to be committing fraud, the documents show.

Ashley Belanger: YouTubers Suspect AI Is Bizarrely Removing Popular Video Explainers. YouTube Denies AI Was Involved. (Ars Technica; October 31, 2025)
This week, tech-content creators began to suspect that AI was making it harder to share some of the most-highly sought-after tech tutorials on YouTube, but now YouTube is denying that the odd removals were due to automation.
Creators grew alarmed when educational videos that YouTube had allowed for years were suddenly being bizarrely flagged as "dangerous" or "harmful", with no clear way to trigger human review to overturn removals. AI seemed to be running the show, with creators' appeals allegedly getting denied faster than a human could possibly review them.
To one content-creator, it seemed possible that YouTube was leaning on AI to catch more violations but perhaps recognized the risk of over-moderation and, therefore, wasn't allowing AI to issue strikes on his account.
To White and others, it's unclear exactly what has changed on YouTube that triggered removals of this type of content. YouTube only seemed to be removing recently posted-content, White told Ars. However, if the take-downs ever impact older content, entire channels documenting years of tech tutorials risk disappearing in "the blink of an eye", another YouTuber warned after one of his videos was removed.
The stakes appeared high for everyone, White warned, in a video titled "YouTube Tech Channels In Danger!"
Late today, a YouTube spokesperson confirmed that videos flagged by Ars have been reinstated, promising that YouTube will take steps to ensure that similar content isn't removed in the future. But, to creators, it remains unclear why the videos got taken down, as YouTube claimed that both initial enforcement decisions and decisions on appeals were not the result of an automation issue.
But, White said in his video, that was just a "theory" that he and other creators came up with, but couldn't confirm, since YouTube's chatbot that supports creators seemed to also be 'suspiciously AI-driven", seemingly auto-responding even when a "supervisor" is connected.
Microsoft declined Ars' request to comment.
Absent more clarity from YouTube, creators who post tutorials, tech tips, and computer repair videos were spooked. Their biggest fear was that changes to automated content moderation could unexpectedly knock them off YouTube for posting videos that in tech circles seem ordinary and commonplace, the YouTubers said.
"We are not even sure what we can make videos about", White said. "Everything's a theory right now, because we don't have anything solid from YouTube."
[We appreciate using free, open-source Linux more, every day!]
Chase DiBenedetto: Gmail Users: Change Your Password Now! After A Summer Of Data Breaches, It's Time To Lock Down Your Accounts. Change Your Passwords, Set Up 2-Step Verification, And Never Click A Suspicious Link. (Mashable, August 29, 2025)
To users that haven't already locked down your personal accounts in light of massive data breaches: It's never too late. That's why Google is once again urging its Gmail subscribers to protect their accounts, following a series of data attacks on corporate systems that could eventually threaten users' personal security. Google sent notifications to its 2.5-billion Gmail users in late July and then again on August 8, warning them that hackers were ramping up phishing activity intended to fool users into giving up their log-in credentials.
Google specifically referred to a group known as "ShinyHunters", which the company says has launched a data leak site (DLS) in an effort to escalate extortion pressure levied at users. Google notes the extortion emails include "shinycorp@tuta. com" and "shinygroup@tuta. com" domains.
In May, cybersecurity researcher Jeremiah Fowler reported that some 184-million passwords were potentially exposed in an open database, with many of the passwords tied to email providers like Google and social-media platforms. One month later, Google Threat Intelligence Group (GTIG) reported that one of its corporate Salesforce server clusters (known as instances) was breached and exposed publicly available business information, such as business names and contact details. The breach was continued activity from an online threat group known as UNC6040, which uses voice phishing to impersonate IT agents, steal data, and extort money. This week, GTIG issued another advisory to Salesforce clients about a large data breach by hacker group "UNC6395."
To prevent users getting bested by future phishing attempts, Google has encouraged its users to set up two-factor authentication and update their passwords. The company has also warned users never to click on emails with alerts such as "suspicious sign in prevented", which are commonly used by hackers during periods of increased cyber-security warnings. Instead, users should check security alerts on their own.
[More on how to do that, in the FULL version of this article.]
NEW: Snowden's Secret: The OS The NSA Can't Crack. (8-min. YouTube video; Bootable USBs, August 23, 2025)
What operating system does Edward Snowden actually trust? In this video, we explore the Privacy & Security category of the Ultimate USB v2.1 - six powerful operating systems designed for anonymity, protection, and complete control of your digital life.
We'll cover:
- Kodachi: Double-layer privacy with VPN + Tor
- NST (Network Security Toolkit): A complete suite for network defense
- PureOS: 100% free software, endorsed by the FSF
- Qubes OS: Snowden's top pick for compartmentalized security
- RoboLinux: Stable, secure, and user-friendly with Cinnamon
- Tails: The live OS that leaves no trace behind
Whether you're a journalist, activist, IT pro, or just want to stay private, these tools give you the same privacy edge that Snowden himself relies on.

OpenAI's GPT-5 Is Now Free For All (but MMS awaits good Privacy/Security assurances):
NEW: Grace Huckins: Why GPT-4o's Sudden Shutdown Left People Grieving. After An Outcry, OpenAI Swiftly Re-Released 4o To Paid Users. But Experts Say It Should Not Have Removed The Model So Suddenly. (MIT Technology Review, August 15, 2025)
A number of people reacted with shock, frustration, sadness, or anger to 4o's sudden disappearance from ChatGPT. Despite its previous warnings that people might develop emotional bonds with the model, OpenAI appears to have been caught flat-footed by the fervor of users' pleas for its return. Within a day, the company made 4o available again to its paying customers (free users are stuck with GPT-5).
MIT Technology Review spoke with several ChatGPT users who were deeply affected by the loss of 4o. All are women between the ages of 20 and 40, and all but one considered 4o to be a romantic partner. Some have human partners, and all report having close real-world relationships. In the backlash to the roll-out, a number of people noted that GPT-5 fails to match their tone in the way that 4o did.
These testimonies don't prove that AI relationships are beneficial - presumably, people in the throes of AI-catalyzed psychosis would also speak positively of the encouragement they've received from their chatbots.
AI companionship is new, and there's still a great deal of uncertainty about how it affects people. Yet the experts we consulted warned that, while emotionally-intense relationships with large language models may or may not be harmful, ripping those models away with no warning almost certainly is. "The old psychology of 'Move fast, break things', when you're basically a social institution, doesn't seem like the right way to behave anymore", says Joel Lehman, a fellow at the Cosmos Institute, a research nonprofit focused on AI and philosophy. In a paper titled "Machine Love", Lehman argued that AI systems can act with "love" toward users - not by spouting sweet nothings, but by supporting their growth and long-term flourishing - and AI companions can easily fall short of that goal. He's particularly concerned, he says, that prioritizing AI companionship over human companionship could stymie young people's social development.
For socially-embedded adults, such as the women we spoke with for this story, those developmental concerns are less relevant. But Lehman also points to society-level risks of widespread AI companionship. Social media has already shattered the information landscape, and a new technology that reduces human-to-human interaction could push people even further toward their own separate versions of reality. "The biggest thing I'm afraid of", he says, "is that we just can't make sense of the world to each other."
Balancing the benefits and harms of AI companions will take much more research. In light of that uncertainty, taking away GPT-4o could very well have been the right call. OpenAI's big mistake, according to the researchers I spoke with, was doing it so suddenly. "This is something that we've known about for a while - the potential grief-type reactions to technology loss", says Casey Fiesler, a technology ethicist at the University of Colorado/Boulder.
OpenAI's decision to replace 4o with the more-straightforward GPT-5 follows a steady drumbeat of news about the potentially-harmful effects of extensive chatbot use. Reports of incidents in which ChatGPT sparked psychosis in users have been common for the past few months and, in a blog post last week, OpenAI acknowledged 4o's failure to recognize when users were experiencing delusions. The company's internal evaluations indicate that GPT-5 blindly affirms users much less than 4o did.
[Thanks to John Rudy (Lexington Computer & Tech Group) for recommending this important article!]
Sabrina Ortiz: OpenAI's GPT-5 Is Now Free For All: How To Access, And Everything Else We Know. We're Testing GPT-5, And Will Have More To Share Next Week. (ZDNet, August 8, 2025)
There are two kinds of OpenAI models in this world: GPT, and reasoning models. The advantages of the former, such as GPT-4o, are that they combine speed and accuracy, while reasoning models such as o3 and o4 take longer to think and use more compute power to produce better answers. OpenAI's latest model, GPT-5, supposedly gives all users access to the best of both models.
Sabrina Ortiz: Everyone Can Use ChatGPT's Advanced Voice Mode Now - Yes, Even Free Users. This Free GPT-5 Feature Is Flying Under The Radar - But It's A Game-Changer For Me. (ZDNet, August 8, 2025
- ChatGPT's Advanced Voice Mode is now available, even for free users.
- With the feature, users can access a conversational voice assistant.
- Advanced Voice Mode, now known as ChatGPT Voice, replaces Standard Voice Mode.
While OpenAI's new large language models (LLMs) in ChatGPT, such as GPT-5, which just launched today, typically steal the spotlight, some of the best gems are found in the less talked-about features, like Advanced Voice Mode.
During its Summer product release yesterday, OpenAI announced that Advanced Voice Mode, the AI-powered voice assistant that mimics a human conversation, is now available to all users, including free logged-in users, for the first time. The feature will replace Standard Voice Mode on Sept. 9 and is now being referred to as ChatGPT Voice.
If you have ever used a voice assistant like Alexa, Gemini or Siri and become frustrated that it does not understand what you are asking unless you word it very specifically, AI-powered assistants (such as ChatGPT Voice) address that issue. With ChatGPT Voice, you can pause as you are thinking while speaking, without the assistant assuming your train of thought is over or cutting you off.
You can also talk to it like you would a human with non-linear, train-of-thought commands. For example, instead of "What is the weather?", you could say "I am going on a run today in Brooklyn, and am wondering what the weather is like so I know what to wear", and ChatGPT Voice would understand your request. To continue to aid that free-flowing dialogue experience, Advanced Voice supports multi-turn conversations, so you can keep the conversation going as long as you'd like without losing prior context.
Another benefit is that it has the context of your surroundings with video- and screen-share options, which helps the assistant understand its surroundings and use that context to provide more informed and relevant answers. Part of yesterday's wave of updates is that ChatGPT Voice can better adapt to the user, better understanding their instructions and adjusting to their speaking style in the moment.
Ashley Belanger: ChatGPT Users Shocked To Learn Their Chats Were In Google Search Results! (Ars Technica; August 1, 2025)
OpenAI scrambles to remove personal ChatGPT conversations from Google results.

NEW: John Roberts: 19 Pros And Cons Of Cyber-Security (ProsPlusCons, August 2, 2025)
The expansion of Internet-connected devices, the rise of e-commerce, online banking, and even the integration of Artificial Intelligence (AI) and the Internet of Things (IoT) have brought immense convenience but have also opened doors for malicious cyber threats. From personal-data theft to corporate espionage, and from the disruption of critical infrastructure to the sabotage of financial systems, the stakes of cyber-security have grown tremendously.
Cyber-security has become a necessity for protecting both individual and organizational assets. The goal of cyber security is to ensure the confidentiality, integrity, and availability of data and systems by preventing unauthorized access, attacks, or damage. However, while cyber-security offers immense advantages, it also brings certain challenges and complexities..
This article delves into the pros and cons of cyber-security, providing a comprehensive understanding of its importance in modern society. By exploring these aspects in detail, individuals and businesses can better appreciate the value of investing in cyber-security while understanding the challenges they may face. Let's start by defining cyber-security, its function, and how it works before diving into its pros and cons.

NEW: Kevin Purdy: Android 15's Security And Privacy Features Are The Update's Highlight. (Ars Technica, October 17, 2024)
New tools aim at phone snatchers, snooping kids or partners, and cell hijackers.
NEW: Doc Searls: Help Us Cure Online-Publishing Of Its Addiction To Personal Data. (Linux Journal; March 14, 2018)
(And, The Big Datastillery that targets YOU!)

NEW: Wolfie Christl, Cracked Labs: Corporate Surveillance In Everyday Life (Institute for Critical Digital Culture; June 2017)
Every click on a website and every swipe on a smartphone may trigger a wide variety of hidden data-sharing mechanisms distributed across several companies and, as a result, directly affect a person's available choices. Digital tracking and profiling, in combination with personalization, are not only used to monitor, but also to influence peoples' behavior.
...
Facebook uses at least 52,000 personal attributes to sort and categorize its 1.9-billion users by, for example, their political views, ethnicity, and income. In order to do so, the platform analyzes their posts, likes, shares, friends, photos, movements, and many other kinds of behaviors.
In addition, Facebook acquires data on its users from other companies. In 2013, the platform began its partnership with the four data brokers Acxiom, Epsilon, Datalogix and BlueKai, the latter two of which were subsequently acquired by the IT-giant Oracle. These companies help Facebook track and profile its users even better than it already does, by providing it with data collected from beyond its platform.
NEW: Cormac McCarthy's only non-fiction essay: The Invasion Of Language: The Kekulé Problem - Where Did Language Come From? (Nautilus; April 17, 2017)
All animals have an unconscious. If they didn't, they would be plants. We may sometimes credit ours with duties it doesn't actually perform. Systems at a certain level of necessity may require their own mechanics of governance. Breathing, for instance, is not controlled by the unconscious but by the pons and the medulla oblongata, two systems located in the brainstem. Except in the case of cetaceans, who have to breathe when they come up for air. An autonomous system wouldn't work here. But the duties of the unconscious are beyond counting - everything from scratching an itch to solving math problems.
Did language meet some need? No. The other five-thousand-plus mammals among us do fine without it.
Problems in general are often well-posed in terms of language, and language remains a handy tool for explaining them. But the actual process of thinking - in any discipline - is largely an unconscious affair. Language can be used to sum up some point at which one has arrived - a sort of mile-post - so as to gain a fresh starting point. But if you believe that you actually use language in the solving of problems, I wish that you would tell me how you go about it.
I've pointed out to some of my mathematical friends that the unconscious appears to be better at math than they are. My friend George Zweig calls this the Night Shift. Bear in mind that the unconscious has no pencil or notepad, and certainly no eraser. That it does solve problems in mathematics is indisputable. How does it go about it? When I've suggested to my friends that it may well do it without using numbers, most of them thought - after a while - that this was a possibility. How, we don't know.
Just as we don't know how it is that we manage to talk. If I am talking to you, then I can hardly be crafting at the same time the sentences that are to follow what I am now saying. I am totally occupied in talking to you. Nor can some part of my mind be assembling these sentences, and then saying them to me so that I can repeat them. Aside from the fact that I am busy, this would evoke an endless regress. There is a process here to which we have no access. It is a mystery.
Although all animal species evolved an Unconscious, all indications are that Language has appeared only once and in one species only - among whom it then spread with considerable speed. Not evolution.
[Click above, to read all of this fascinating essay! Also, view many of the oldest-known (Upper-Paleolithic) paintings, from Chauvet Pont-d'Arc Cave in France.]
Return to main section of Money Is Not Wealth.