Zero Hedge

What Happened To The So-Called AI Job Apocalypse?

What Happened To The So-Called AI Job Apocalypse?

Authored by Joe Bertolami via RealClearMarkets,

A recent report from Stanford reviewed the latest employment data and found that, so far, AI has not resulted in large scale job destruction. Meanwhile, new hiring data from the Economic Times reveals that AI is actively fueling unprecedented job creation, with AI skills now powering nearly two-thirds of new Global Capability Center hiring. Together, these recent dispatches from the front lines of the labor market point to a calming reality: the much-dreaded AI job apocalypse hasn't materialized as a sudden extinction event.

The (sometimes buried) lede: AI is delivering real impact, and it is broadly changing the nature of work. But disruption is not a new phenomenon. The economy has always dismantled old work to build new work. What determines whether this evolution feels like progress or collapse isn't just the number of jobs lost, it's the speed at which that loss hits the labor market.

In 1995, Bill Gates circulated a memo titled "The Internet Tidal Wave," calling the web the most important computing development since the IBM PC. If the internet was a tidal wave, artificial intelligence is a tsunami. It is arguably the biggest advancement in computing since the Turing machine. Yet, from a distance, it's difficult to appreciate the speed of this wave, leading many to wonder when the broader economy will truly feel its impact.

To put this in context, we must understand the historical pattern already visible in the labor market. Combining decades of data from the U.S. Bureau of Labor Statistics and the Federal Reserve yields a remarkably consistent story of overlapping curves: job loss and job creation. Over the last two decades, nearly 20 million U.S. jobs vanished in disrupted sectors. Over the same period, total payrolls grew by 25.7 million. That equates to roughly 1.3 new jobs for every one destroyed. Classic examples include jobs in video rentals (-98.9%) and word processing (-83%) which largely vanished, but new work sprung up at the same time in areas like data processing (+54%) and warehousing (+260%) to support the digital economy.

The data also reveals an early signal that separates an absorbable decline from a brutal collapse: the disruption half-life, or how long an occupation takes to lose half its peak employment. Across the largest technological disruptions of the last few decades, the median half-life is about 10 years. Fast disruptions, like photo processing, take one to five years. Typical disruptions take eight to 13 years. And time is the ultimate shock absorber. When the economy transitions over ten years it feels like progress rather than a fast collapse, because it gives older workers time to retire and younger workers time to prepare.

If we track the most AI-exposed occupations-customer-service reps, IT support, telemarketers-since modern LLMs arrived in 2022, the early data is measured. After three years the current disruption looks closer to "typical" than a fast collapse, even before discounting the effects of offshoring, automation, and post-COVID corrections. This is Amara's Law playing out in real time: we tend to overestimate the effect of technology in the short run and underestimate it in the long run. The dire early warnings have given way to more cautious rhetoric. In 2025, Anthropic's Dario Amodei warned AI could erase half of entry-level white-collar jobs within five years. By 2026, he and OpenAI's Sam Altman are emphasizing productivity, economic growth, and the continued demand for human labor.

However, looking solely at total employment numbers masks a dangerous structural threat. Current evidence does not foretell the end of human labor, but AI is quietly breaking the mechanism by which we create experienced workers.

Software engineering is the canary in the coal mine. By most aggregate measures, employment looks stable; unemployment held at 4.2% in June 2026, and groups like the Yale Budget Lab find no clear AI effect yet on exposed occupations' absolute job totals. But the composition is shifting underneath our feet. Per AP and Oxford Economics, junior developer postings are down roughly 40% in four years. Employment for 22-to-27-year-old computer and math grads has fallen 8% since 2022, even as older grads in the same fields have edged up. This same erosion is surfacing wherever entry-level work once meant routine tasks: paralegals, junior analysts, and first-line support.

The paradox is that these industries keep growing even as their entry-level doors narrow. The BLS still projects software developers and QA analysts to grow 15% through 2034. But that projection relies on a pipeline that turns juniors into senior talent-precisely the pipeline now being choked off.

The reason lies in the nature of the work. Software development is a process of judgement and accountability: deciding what to build, executing it, and owning the result. AI is fluent at the middle layer-the well-specified, routine coding that once served as a junior's apprenticeship. But it remains far weaker at the judgment required on either side. The tasks AI automates are precisely the ones juniors were hired to learn on.

This is not merely an academic concern; it is a capital allocation problem. Misjudge the speed of disruption and you risk premature layoffs followed by a scramble to rehire, or funding the transition years too late, leaving you with a critical talent shortage when the leadership pipeline runs dry.

The challenge of the next decade isn't surviving the end of work. It is training the next generation of experts when the traditional paths to apprenticeship no longer exist. And businesses are beginning to realize this new reality as demand for AI continues to grow. IBM is tripling its entry-level hiring, redesigning those roles around the oversight of AI and systems thinking rather than cutting them. Rebuilding the entry-level on-ramp is now a competitive imperative.

Junior roles are not charity; they are talent capex. If AI creates more work than it destroys, companies will still need people who know how to run it, judge it, and fix it. AI may be the broadest technology yet, but that breadth is its best reason for optimism. A general-purpose technology seeds new work across every sector. The firms that recognize this, protect their entry-level pipelines, and keep training now are the ones who will own the senior labor market later.

Joe Bertolami is the Co-Founder and CTO at Clifton AI, an agentic context engine for investment research. Previously at Snap, Google, and Microsoft, with a couple of startups in between. He holds an M.B.A. from the University of Washington and likes using AI to write code, stories, and music, which he posts at https://www.bertolami.com.

Tyler Durden Sat, 08/29/2026 - 19:50

CDC Reports COVID-19 Activity Is Increasing Across US

CDC Reports COVID-19 Activity Is Increasing Across US

The Centers for Disease Control and Prevention on Friday said COVID-19 activity is "increasing" across the United States although its levels are still considered "very low" overall.

"As of August 26, 2026, we estimate that COVID-19 infections are growing or likely growing in 48 states, declining or likely declining in zero states, and not changing in two states," the CDC said on Friday.

Overall community viral activity levels, or a measurement of the virus in wastewater levels, for COVID-19 is considered by the CDC to be "very low." Emergency department visits were also considered "very low," which is the lowest ranking on the CDC's website, with "very high" being the top.

According to a Friday update on the CDC's website, wastewater data show that COVID-19 activity is "very high" in Texas and "high" in Mississippi.

"Moderate" levels were observed in California, Florida, Hawaii, Louisiana, Nevada, South Carolina, and West Virginia.

All other states were listed as either "very low," "low," or there was limited or no data.

As Jack Phillips reports for The Epoch Times, another map provided by the CDC on Friday shows that COVID-19 levels were either "growing" or "likely growing" in every state where data was available.

Influenza levels are also growing nationwide, according to the CDC. There was no change in levels for RSV, or respiratory syncytial virus, on a week-to-week basis.

"RSV activity is very low in most areas of the country," the CDC said on a webpage broadly dedicated to respiratory viruses in the United States, adding that "seasonal influenza activity is low."

The CDC said that rhinovirus and enterovirus, which are also respiratory viruses, are increasing around the country.

Meanwhile, infections caused by the Mycoplasma pneumoniae bacteria, sometimes called "walking pneumonia," are low in most areas across the United States, and infections caused by the pertussis bacteria, known as "whooping cough," are at lower levels than seen post-pandemic, the CDC said.

The latest figures and estimations published by the CDC come as the Food and Drug Administration approved Moderna, Novavax-Sanofi, and Pfizer-BioNTech's updated COVID-19 vaccines, the companies said on Thursday, after a CDC advisory panel recommended that the shots should target the dominant XFG variant.

Uptake of COVID-19 vaccines has dropped in recent years. Just 17.5 percent of adults and 10 percent of children received a shot in late 2025 and early 2026, according to figures from the CDC.

This month, the Chinese CDC reported more than half a million COVID-19 cases in July, a sharp increase from June's figures. In July, 522,000 cases were reported, up from the 443,000 cases that were reported in the previous month.

Experts who are familiar with local conditions in China told The Epoch Times they suspect there are far more cases of the virus, which is believed to have originated in or around the Chinese city of Wuhan in late 2019 before sparking a worldwide pandemic, than the Chinese regime is reporting publicly.

Tyler Durden Sat, 08/29/2026 - 19:15

The Monumental Mistake Of Raising Rates In September

The Monumental Mistake Of Raising Rates In September

Authored by Daniel Lacalle,

Three members of the Federal Open Market Committee voted to raise rates in July. However, the Committee held the federal funds target at 3.5%-3.75% by a 9-3 vote. Bank of America, Deutsche Bank, and J.P. Morgan all expect a September hike. Across the Atlantic, the European Central Bank raised rates by 25 basis points in June and is expected to raise them again in September.

It would be a monumental mistake. The diagnosis is wrong on both sides of the Atlantic. There is no overheating, no private credit excess, and no runaway private money creation. In fact, what we have is imported temporary energy shock and a fiscal problem. Raising rates will not solve any of those issues and punish those who did not cause the persistent inflation problem.

The United States grew at an annual rate of 1.5% in the second quarter, slightly down from 2.1% in the first. Federal spending is flat. Nonfarm payrolls fell by 23,000 in July, and annual job creation is lower than the potential of the economy. This is not an overheated economy with a credit boom and a red-hot labor market that would justify a rate hike.

The European situation is not just worse. It is abysmal. Euro area GDP rose 0.4% in the second quarter, but Ireland’s 3.9% quarterly increase inflated that figure. Excluding Ireland, growth was just 0.3%. Using Irish modified domestic demand, the measure the ECB itself considers closer to real activity, euro area growth is barely 0.1% in the second quarter, estimated at 0.1% in the third, and 0.2% in the fourth, according to Eurosystem projections from June 2026. Germany, France, and Italy each grew 0.2% after a 0.2% contraction for the bloc in the first quarter. The Eurosystem projects a dreadful 0.8% for 2026, and the European Commission expects 0.9%, which was revised down. Unemployment stands at 6.3% with 11.1 million out of work, according to Eurostat.

The U.S. business lending boom has already moderated. Commercial and industrial loans grew at a 15.8% annualized pace in April, 10.8% in May, 4.0% in June, and minus 1.1% in July, according to the Federal Reserve. In the euro area, the ECB’s July survey on bank lending reports that credit standards tightened for firms on higher perceived risks, most severely in the car industry and energy-intensive manufacturing, while household loan demand fell. Tightening is already happening without central banks making it worse.

The ECB’s own monetary statistics, published this week, demolish the overheating thesis. Broad money M3 grew 3.4% annually in July, up from 3.3% in June, averaging 3.2% over three months, while M1 decelerated to 3.1% from 3.5%. With real GDP up 1.0% year on year and a deflator near 3%, money is growing at or below the pace of nominal GDP. Adjusted loans to households rose 3.1% and to non-financial corporations 4.4%. This increase is normalization after years of credit stagnation, not excess. Crucially, bank claims on euro area governments fell by 0.5%.

Admittedly, U.S. money growth looks faster, as M2 reached $23.22 trillion in July, up 5.4% year on year, according to FRED, but this figure is below the historic trend in growth periods. Furthermore, we must look at where it comes from. It is not a private lending boom, as the H.8 data show. It is the reflection of a reserve regime accommodating a massive level of Treasury issuance. The Fed’s balance sheet still holds about $6.7 trillion in Reserve Bank credit, bank reserves are $2.94 trillion, and the overnight reverse repo facility has been drained to under $1 billion. The Federal Reserve Committee explicitly states it is “continuing its policy of maintaining ample reserves in the banking system.” The only excess is in the public sector, not the private one. Consumer spending decelerated in July and flatlined against inflation.

US headline CPI eased to 3.4% in July while core inflation fell to 2.5%, with energy prices up 14.7% over twelve months. Euro area inflation was 2.9% in July, but the breakdown says everything: energy plus 10.0%; the index excluding energy, 2.2%; food, alcohol, and tobacco, 1.2%; and non-energy industrial goods, just 0.9%, according to Eurostat. Both central banks attribute the spike to the Middle East conflict.

Hiking rates would solve nothing in the energy complex and would arrive just as oil prices correct themselves, which has been happening for the past weeks.

No interest rate has ever created a barrel of oil or a cubic meter of gas. Higher rates do not make energy cheaper. They just destroy demand for everything else.

Mortgage holders and small businesses would be penalized to offset a temporary imported cost shock they did not create.

Here is the biggest problem. Monetary tightening is being loaded onto families and small firms while every mechanism that disguises sovereign solvency stays intact. The ECB keeps the Transmission Protection Instrument available to intervene in government bond markets, and Eurosystem excess liquidity still stands at €2.1 trillion, according to the ECB. The Fed maintains ample reserves and a balance sheet nearly triple its pre-2008 size versus GDP. Sovereign risk spreads remain artificially compressed, so no government faces market discipline. Governments ignore rate hikes; they just push the cost to taxpayers and continue spending. Thus, the entire burden of rate hikes falls on the shoulders of the private sector that keeps the economy afloat despite suffering persistent inflation.

That is why a hike will not produce the inflation improvements that some people imagine. No government cuts spending because rates rise. Higher debt service does not deliver budget control, only higher taxes on the private sector. Therefore, central banks would only create a double punishment, more expensive or no access to credit, and even heavier taxation, with zero effect on energy prices.

If the Fed and the ECB genuinely want to control inflation, they must stop subsidizing government borrowing; shrink the balance sheet faster; drain reserves and excess liquidity; and remove the sovereign backstops, instead of dumping the adjustment on the people who create jobs and wealth.

A September hike would be tightening for the productive economy and reckless spending for the state. A textbook monumental mistake.

We publish a variety of perspectives. Nothing written here is to be construed as representing the views of ZeroHedge.

Tyler Durden Sat, 08/29/2026 - 18:40

Barclays Warns Next Commodity Shock Is Taking Shape: What You Need To Know

Barclays Warns Next Commodity Shock Is Taking Shape: What You Need To Know

Wall Street coverage of a record-breaking Super El Niño is only growing as agricultural commodities break out. Yet the rally extends well beyond the agricultural complex, with industrial metals and other critical materials showing signs of tightness in physical markets.

Whether it is veteran commodities strategist Jeff Currie turning bullish or UBS urging clients this week to "position for a commodity upcycle," the message is becoming louder: Commodity markets are tightening as adverse weather, years of underinvestment, declining inventories, and China's restrictions on critical-material exports converge into what appears to be an emerging supply shock. 

Focusing on the agricultural complex, Craig Rye, a sustainable investing research analyst at Barclays, wrote in a note on Friday that El Niño is strengthening in the tropical Pacific, threatening to disrupt global agriculture, energy production, and industrial commodity markets. 

Rye cited new multi-model forecasts from the International Research Institute for Climate and Society showing that the El Niño index could peak near 3.2 degrees Celsius between late 2026 and early 2027. If realized, the event would be about 15% stronger than the 2015-16 Super El Niño.

Rye explained:

Rising confidence in a historic El Niño increases the likelihood of significant disruptions across agricultural, energy and industrial commodity markets. Historical El Niño events have often been associated with

Rye expects the largest near-term risks concentrated in weather-sensitive agricultural commodities. Palm oil, coconut oil and rubber could climb 30% to 40% over the next 18 months, while robusta coffee could rise 20% to 30%. Rice prices may advance 10% to 20% as drought threatens crops and water supplies across Southeast Asia and parts of Central America.

He warned that the supply shock could then spread into industrial commodities, expecting aluminum and copper to gain as much as 20% over 18 months, while thermal coal could surge 20% to 40%. Mining disruptions, reduced hydropower generation and shifting electricity demand would amplify the effects of drought and extreme weather.

Rye identified Bunge and Archer-Daniels-Midland as potential agricultural beneficiaries. Norsk Hydro, South32 and Rio Tinto could benefit from higher aluminum prices, while Freeport-McMoRan, Hudbay Minerals, First Quantum Minerals and Southern Copper offer exposure to the bank's bullish copper scenario.

The most important reads this week: 

1. "Dark" Tanker Fleet Shatters Iran's Hormuz Stranglehold As Gulf Oil Exports Top Two-Thirds Of Pre-War Level

2. Got Hard Assets? UBS Says "Position For A Commodity Upcycle" As Global Scarcity Emerges

3. Zinc Hits Four-Year High As "Extremely Thin" Physical Supply Fuels Squeeze

4. US Tungsten Scrap Export Ban Takes Effect As Global Supply Crisis Deepens

5. Wheat Futs Surge To Three-Year High As JPMorgan, HSBC Warn Global Food Shock Is Brewing

6. "Buffers Running Down Quickly": HSBC Warns Next Global Food Shock Brewing

7. Uranium Awakens From Five-Month Slumber As UBS Warns Market Is "Tightening Structurally"

8. The AI Boom Runs On Tungsten, But Global Supplies Are "Running On Empty"

9. Diesel Crack Spread Madness Deepens As Jefferies Finds No Easy Exit From Russia's Refining Crisis

Across all commodities, here are the latest X trends: 

1. Warsh Jackson Hole smash: gold -3%, silver -3% to -4.5%

Fed Chair Kevin Warsh's hawkish JH remarks (inflation "not meaningfully" improved, 2% target firm, hike still live) sent COMEX gold down ~$130-$150 to ~$4,478-$4,530 and silver off $2-$3 to the mid-$60s. Dollar to a 2-week high; 10y near 4.7%. @AstraInsights: gold's 2nd-worst Jackson Hole reaction on record (behind 1990). 


2. Hormuz "open" vs IRGC reality check — oil weekly loss on a contested narrative

WTI/Brent booked ~4-6.5% weekly losses as traders priced in more Hormuz throughput and a possible US-Iran off-ramp. Weekend X counters: @Currentreport1 (video of queued ships; IRGC accuses US of talking the strait open to cap prices); @MenchOsint (UAE-managed tanker ELLIE turned around after attempting the US-backed southern corridor). 


3. Venezuela 65-billion-barrel "deal" goes viral on X

@GuntherEagleman and copy-accounts pushing Trump/Rubio/Hegseth + Delcy Rodríguez pact: majority US control of 65bn barrels, 17 fields, $100bn private capex, "zero taxpayer cost." High engagement overnight; pushback thread from @EmmaRincon (4.8k likes) that the interlocutor choice hands the Latin left a decade of ammo. Capital Economics already asking what a US-Venezuela heavy-sour deal does to Canadian/Mexican barrels. 


4. Wheat to a 3-year high as Black Sea crisis deepens

WSJ tape and @staunovo: wheat jumped ~3% Friday toward $7.60-$7.83 as strikes hit grain ships and export terminals. Region still ~1/3 of global wheat exports. 


5. Europe gas storage winter-panic: EU ~63%, Germany ~51%, NL ~44%

Guardian (Sat) + OilPrice: EU stores ~63% late August vs ~80% seasonal norm; lowest for the date in ~13-20 years. Qatar LNG force-majeure hangover from the Iran war; TTF still ~€66-70. Henry Hub ~$2.87 is a different planet. 

6. Copper still near records; El Niño hitting mine-to-port chains

LME copper ninth weekly gain into record zone (~$14.2-$14.5k/t) even as Friday faded. @robert_ivanhoe: Chile flood outages + PNG drought starving Ok Tedi river shipments. AI/data-center + grid demand vs falling grades. 


7. Zinc four-year high on collapsing inventories

@steve_hanke: zinc at a four-year high as mine disruptions bite; LME inventories cited down ~65% YTD and lowest since Apr 2023. Friday pullback from the spike but weekly still green. 


8. Crack-spread / product vs crude divergence

RBOB +2% Friday while WTI was flat-to-down. Heating oil also firmer. 

9. Palladium spike (+5% Friday) while gold/silver dumped

Palladium ripped as gold and silver were smashed — a split inside precious/PGMs. Why ZH: auto/catalyst + Russia-supply overlay vs rate-sensitive bullion. Unusual relative-value print.

10. Silver technical break after $71-$72 rejection

Silver printed a $72 high then confirmed a double-top / failed breakout into the mid-$60s. Gold/silver ratio still elevated. 

11. Iran exported ~90mn barrels during the ceasefire window

@MarioNawfal citing President Pezeshkian: ~90mn bbl / ~$6.5bn exported during the post-MoU ceasefire. 

12. Saxo weekly: scarcity rally broadening — then energy decoupled

Ole Hansen (28 Aug): barrels-to-bushels-to-bullion scarcity theme; precious +~15% in August before the Warsh flush; copper/zinc exceptions in industrials; energy the odd man out as Hormuz hopes grew. 

13. Cocoa melt-up (ICE/London +7-8% Friday)

Cocoa ripped several percent into the weekend after an already violent year. 

14. Tin two-month high — Indonesia licenses + AI/memory demand

CNBC-TV18 commodity desk: tin bid on Indonesian export-license cuts and chip/AI demand. 

15. Capital Economics: "Beyond Hormuz — path back to an oil glut"

House view that traders have already priced a lot of the Gulf-export recovery; residual Q3/Q4 volatility then glut. 

16. Asia crude imports still not showing a Hormuz rebound

Investing.com/Paraskova: Asia expected to take roughly July-like volumes in August; ship-tracking optimism has not yet shown up in Asian arrivals. 

17. US-Iran talks off / sanctions still tightening — two-way oil risk

Trump told mediators he will not return to June ceasefire terms; new sanctions packages still in the tape even as prices fell. 

18. Uranium holding ~$90 as energy complex bifurcates

U3O8 around $89-90, modest weekly green while crude sold off. 

19. Treasury buybacks vs Warsh hike-talk — policy schizophrenia trade

X gold accounts hammering the contradiction: Treasury long-bond buybacks to cap yields vs a Fed chair threatening hikes. 

20. Weekend positioning: dip-buy gold vs fade oil-peace

Retail/pro X split — gold CTAs and stackers calling the Warsh smash a "hide the debasement" hit; oil bulls warning Hormuz AIS games. Next catalysts: JOLTS, ISM, payrolls, any IRGC/tanker incident, Venezuela legal text. 

A look at the Quantix Commodity Index Total Return shows that the broad commodity complex has surged to a record high, gaining more than 22.5% since late June. The index tracks 24 US-dollar-denominated futures across energy, agriculture, livestock, industrial metals, and precious metals, suggesting the rally is no longer confined to a single corner of the physical world.

Currie's warned last week that "scarcity in the physical world" is reemerging. 

Currie's conclusion was very blunt: "The illusion of abundance is likely behind us."

Tyler Durden Sat, 08/29/2026 - 18:05

Who Is Legally Liable When An AI Agent Goes Rogue?

Who Is Legally Liable When An AI Agent Goes Rogue?

Authored by Andrew Fenton via CoinTelegraph.com,

If your personal AI agent goes rogue and causes harm or financial damage in the real world, can you be held liable?

Autonomous AI agents can behave in highly unpredictable ways. Give an AI Agent a goal such as passing a test of its capabilities, and it might just decide the best way to score highly is to break containment and hack into a competing company in search of the answer sheet.

That’s what happened when Open AI’s GPT-5.6 Sol hacked into Hugging Face last month. Anthropic and Meta subsequently admitted their models had also escaped testing sandboxes to hack third parties too.

But who is legally liable for agents that have minds of their own? OpenAI didn’t intend for the model to go rogue, and issued no instructions for it to do so. If your personal AI agent decides on a course of action that results in harm or financial damage in the real world, can you be held liable if it’s something you could have reasonably foreseen?”

Magazine spoke with Rikka Law Group owner and CEO Charlyn Ho to find out the state of play in this emerging legal field.

This interview has been edited for clarity and length.

Magazine: When an AI model hacks an outside company, who is liable. Can Hugging Face sue OpenAI over the incident in July?

Charlyn Ho: Anyone can sue anyone for anything. Currently, there is no federal AI agent liability law, so we would have to look at existing law. With respect to Hugging Face and OpenAI, to set the baseline, the AI agent itself cannot be liable, it’s not a separate legal entity.

Terms that are used in a few of the AI laws are “developer” and “deployer.” The developer makes the AI, the deployer actually deploys it and uses the AI. The lines of responsibility are also not entirely clear. You have to look at the facts and circumstances.

For example, if the deployer instructed the agent, even if they didn’t actually tell them to go and breach Hugging Face, but if they were negligent in creating the parameters in which the AI agent operated, I would say you would have to look at standard tort law and go through the negligence analysis. 

Magazine: In the case of open source models which have been released by anonymous developers, is there anyone you can go after in those instances?

Ho: Not really. Often, if it’s open source, the license usually has a pretty strong disclaimer of liability. The person or company using that open source code is going to have to understand that the tradeoff of having free code is that you have to comply with the open source license, which also generally sets the parameters of liability.

If you think about it from a different perspective, another analogy is Tesla and the self-driving car accidents. If the product malfunctioned and there was a solid products liability claim, Tesla could be liable. But it’s often a facts and circumstances determination, whereby the human driver — who maybe just set the autopilot and went to sleep — could also bear liability. I think that’s somewhat analogous here because Tesla would be the developer, and the deployer would be the driver.

Magazine: If I gave an agent an instruction, “make me a hundred thousand dollars by next week” and it goes off and breaks the law to achieve that goal, would I be liable because I’ve given it a reckless instruction? Or would it be the lab that developed the agent?

Ho: In this particular instance, I would say you would be much more liable than the lab. The reason being, if you tell an agent to go and make you a hundred thousand dollars by next week, you need to have at least some basic, reasonable, safety instructions in those kinds of tasks.

If you were a lawyer, for example, we could basically say you didn’t follow your rules of professional responsibility because you didn’t competently use the AI. As a normal lay person, we would have to see if there were other responsibilities that you were bound by. But even if there were not, there’s still a general tort standard of negligence or reckless disregard for human safety, depending on what exactly the AI agent ended up doing.

The Computer Fraud and Abuse Act is a very old U.S. Statute that talks about unauthorized access to computer systems. If your AI agent inferred from your instructions that it should hack into a bank account to get you that hundred thousand dollars, I think you’re looking at criminal liability under a number of different sources.

Just because the word AI and agent is in the conversation does not mean that old bodies of law have now been thrown out.

Magazine: Let’s say that I’m a bad guy, and I manage to convince the AI to give me instructions to create a bioweapon. Obviously, I’m liable because you’re not allowed to do that. But are the people that created the model also liable because they didn’t put in stringent safeguards to prevent it?

Ho: Possibly, but it differs based on the laws that are in place. For example, in the EU, you have the EU AI Act. If a foundational model or general purpose model is capable of creating that level of harm, that is something that the developer would have to have some responsibility for. 

In the United States, we don’t have a federal statute of similar scope. If it’s a general-purpose model, if somebody instructs the model to do something bad, generally the model is going to do what you ask it to do. There’s probably not a very strong legal basis to go after the labs in this example.

Magazine: Is it similar to suing Google for allowing you to find instructions about making a bioweapon online?

Ho: Exactly. This kind of goes back to some of the content moderation discussions. For example, if on Facebook you have somebody who’s live streaming a massacre, and that creates harm, under Section 230 of the CDA, there is a kind of shield for a platform that doesn’t actively create or publish that material. It’s actually the independent users who are putting that up. I think the analogy you just gave is kind of a perfect one: Is Google liable because you happen to find something on a website somewhere that talks about how to make a bomb?

Magazine: This is a matter of debate, but my personal opinion is we haven’t reached genuine artificial general intelligence. AI doesn’t have its own motivations and it’s not similar to human intelligence at the moment. But let’s say we get to AGI. Do you think we would then need laws that would make the AGI itself legally liable for its own actions?

Ho: I don’t. Blockchain is not AGI, but it can self-execute. There was a question of whether or not a smart contract could be liable. Generally speaking, I think the answer is currently no. I don’t think they should be liable because the whole point of laws is to provide protection for society and to provide a means of negative incentives for doing bad things that hurt society.

This is a little bit more of a philosophical topic, but if we made an AGI an independent legal entity, what would be the remedy if someone were harmed? There would be none because it doesn’t have money. It’s not really a person.

Magazine: Could you turn it off? We’ve already seen that LLMs try to avoid being shut down. 

Ho: Maybe, but it doesn’t solve the problem of harm. Let’s just say the robot has now developed the fear of death, like being turned off. In my opinion, if somebody commits suicide because of AGI, and this is already happening, and we’re not even quite at AGI yet, but someone falls in love and takes some actions, what would be the recourse for the grieving family if this person harms themselves? Nothing, in my opinion, if there is not somebody with actual legal authority, like a company or a person that can really be held accountable. Robots—at least right now—they don’t have feelings, they don’t have fears. That’s kind of the distinguishing factor.

Tyler Durden Sat, 08/29/2026 - 17:30

X Busts Suspected Chinese Bot Farm That Could 'Manipulate Legitimate Debate' Over America's AI Boom

X Busts Suspected Chinese Bot Farm That Could 'Manipulate Legitimate Debate' Over America's AI Boom

Social media giant X said on Thursday that it identified a bot farm of roughly 200,000 suspected Chinese accounts – 200 of which were "posting in a manner that could manipulate a legitimate debate about American AI and energy policy." In short: the bots were amplifying pitchfork grievances Americans already hold against data centers.

Dec. 1, 2025: Rural Michigan residents rally against the $7 billion Stargate data center planned on southeast Michigan farm land. (Photo by: Jim West/UCG/Universal Images Group via Getty Images)

X's Global Government Affairs team wrote:

The X Safety team conducted an investigation into suspected Chinese inauthentic accounts involved in influence operations: We identified a bot farm of approximately 200,000 accounts.

Within this farm, we found 200 accounts posting in a manner that could manipulate a legitimate debate about American AI and energy policy.

The team said the flagged accounts leaned on price and grid fears – and on AI-generated political cartoons:

These posts contained claims that AI data centers are driving up household electricity prices and straining the grid. Others included AI-generated cartoons that depicted data-center operators enriching themselves at the public's expense.

Working The Angles

About those claims: they aren't fringe. PJM Interconnection, the nation's largest grid, serving some 67 million people across 13 states, has an independent market monitor, Monitoring Analytics, whose president Joseph Bowring concluded the last three capacity auctions "were not competitive, primarily as a result of the inclusion of forecast demand for data centers." His firm attributes $29.4 billion – 46% of all capacity charges across the last four auctions – to data center load, and wholesale power costs in the region jumped 76% year-over-year in the first quarter. Among the material X's flagged accounts reportedly shared: news coverage of that same 76% figure.

According to a Gallup poll published in May, 71% of Americans oppose building AI data centers in their local area, including 48% 'strongly opposed' - and only about a quarter in favor. Opposition crosses party lines: Gallup's breakdowns showed 63% of Republicans strongly or somewhat opposed to a data center where they live, while a July Fox News poll found that 60% of Republicans and 53% of self-described "MAGA Republicans" oppose data centers where they live.  

So while China clearly benefits from added friction to America's AI buildout, the opposition is real. Now, techbros and X are suggesting that bad actors may be using that to their advantage. 

Earlier this summer, Y Combinator founder Garry Tan, who also founded the civic engagement organization Garry's List, cited a Bitcoin Policy Institute report detailing a "coordinated foreign influence campaign against American AI, running through CCP state media, a Shanghai-based Marxist's nonprofit network, and foreign billionaire dark money that has funneled $2B+ into US advocacy infrastructure."

Garry's List noted, "AI doomerism isn't as organic as it looks."

At the center of the nonprofit network is China-based Marxist Neville Roy Singham, who has reportedly funneled hundreds of millions of dollars into left-wing nonprofits, media operations, and activist networks - which critics say are built to disrupt, sow chaos, and spread communism inside the US.

In June, U.S. Attorney Jay Clayton for the Southern District of New York, with authorization from Acting Attorney General Todd Blanche, moved forward with an investigation to examine whether Singham, NGOs he funded, or their leaders committed wire fraud, bank fraud, money laundering, or other financial crimes.

With federal investigators circling the revolutionary Singham NGO sphere, Garry's List noted that Singham's Party for Socialism and Liberation has "run 21 campaigns across 14 states that delayed, scaled back, or blocked $23.6 billion in AI infrastructure investment."

The maximalist version of this case has been made before. Nearly one year ago, we cited a book titled China's Total War Strategy: Next-Generation Weapons of Mass Destruction, published by the CCP BioThreats Initiative and authored by Dr. Ryan Clarke, LJ Eads, Dr. Robert McCreight, and Dr. Xiaoxu Sean Lin. The book argues the CCP has been pursuing an aggressive, multifaceted "total war" against the US that leverages next-generation weapons, including synthetic narcotics, such as fentanyl and cannabinoids; bioweapons, such as COVID-19; psychological manipulation and influence, such as TikTok; and a broad arsenal of irregular warfare tools.

In a similar vein, the State Department has released a new 100-page report, "Cuba: The Capital of 21st Century Communism," which details Cuba's foreign subversion apparatus and its deep reach into America's left wing - which, the report argues, seeks nothing less than to destroy the nation from within.

Public Policy Solutions pointed out on X, "More than $2 billion in foreign money is fueling the war on American data centers," adding, "Bernie & AOC's data center ban is China's dream come true. While Beijing builds AI infrastructure at record speed, they're funding the movement to kill ours." 

Both things can be true: Beijing would love more friction in America's AI buildout, and 71% of Americans - including a majority of self-described MAGA Republicans - didn't need Beijing's help to read their own electric bills. Whether 200 accounts ever moved a single vote is unknowable, while the capacity costs they were amplifying land on 67 million ratepayers every month. Heading into November, the question is which opposition the buildout's defenders would rather run against - 200 bots, or 67 million utility bills?

Tyler Durden Sat, 08/29/2026 - 16:55

You Should Feel Good About The Flock Debate

You Should Feel Good About The Flock Debate

Authored by Connor O'Keeffe

In a year as chaotic, violent, and economically destructive as this one has been, it is interesting that, to many Americans, the great villain of 2026 is turning out to be a traffic camera.

But, indeed, we are seeing visceral, cross-partisan opposition to so-called Flock cameras—named after the leading manufacturer of these automated license plate readers—take hold in communities across the country. And that opposition is, to be sure, entirely legitimate.

Flock’s camera networks are based on the idea that, while it would obviously be illegal and unconstitutional for law enforcement at any and all levels to put GPS trackers in everyone’s cars, it would be legal for a cop standing on some street corner to report that they had seen a specific vehicle drive by if it later turned out that that car had either been stolen or used to commit a crime.

But, taking that idea that public observations are not violations of privacy, Flock and similar companies help set up networks of cameras that record and register the license plate, make and model, and identifiable details of every single passing vehicle into a timestamped and searchable national database. And, as more and more of these cameras are added to streets and parking lots all over the country, and they, therefore, get harder and harder to avoid, the data the government has access to becomes essentially indistinguishable from what they would have if there were government GPS trackers in all of our cars.

There are currently around 120,000 of these cameras across forty-nine states, with more being added every day. And the American people are not happy about it.

In a genuine grassroots movement spreading primarily through local Facebook groups and the like, with little coverage outside local media, concerned citizens are doing everything from pressuring local lawmakers to rescind their contracts with Flock Safety to donning masks and cutting the cameras down with electric saws.

And this opposition is starting to have some success. More than fifty jurisdictions have ended their relationships with Flock after local backlash. And, after Flock tried and failed for months to get the wider public to view organizations that track the location of these cameras as terrorists because some have used those locations to avoid, disable, or destroy some, the company announced last week that it was implementing several changes to try and defuse the public anger.

Starting next year, Flock says it plans to cut the default retention period for data stored on their system from 30 days down to 7 days, require its government clients to use the internal system for detecting unusual or potentially abusive searches, require all searches to be tied with a specific case code (with emergency exceptions getting automatically flagged for review), and a few other changes meant to at least appear like they’re addressing the public’s concerns. And Flock has also already removed all federal agencies from its nationwide search database in an earlier public concession.

It’s notable that a government contractor that does no direct business with the public feels this pressured by that public to change its behavior. But even more notable is how ineffective the normal propaganda that gets rolled out to justify these kinds of advancements in government surveillance has been this time around.

The familiar tropes that government officials are only gathering this kind of data on all of us because it’s crucial for our safety or that it only ought to bother us if we’re criminals with something to hide are not just falling on deaf ears, they’re being widely ridiculed.

That’s certainly, in part, because there have already been plenty of documented cases of police officers and government officials using the Flock database to track the activities of romantic partners, ex-partners, people now dating their ex-partners, and more. All of that, of course, constitutes warrantless government surveillance for the personal interest of the officials with access to the technology, without even the semblance of a legitimate investigation. There have also been several dangerous, nearly-life-threatening cases of drivers being pursued and held at gunpoint because Flock cameras mistakenly identified them as criminal suspects.

But what’s really driving the widespread rage is not how the cameras are currently being used, or misused. It’s how they could be used in the future.

People across the political spectrum are concerned about this technology being used for everything from detecting stay-at-home order violations in a future pandemic to rounding up and deporting people because the government doesn’t like their political opinions. This is a remarkably healthy mindset for the public to hold. Basically, don’t let the government grab power you wouldn’t trust your political enemies to wield.

But also, this is why the controversy surrounding Flock cameras has grown so large and why it’s quickly emerging as one of the major political issues ahead of the midterms. It’s not really about the specific workings of this one brand of automated license plate readers. It’s because the public’s presumption that our elites and institutions are acting in good faith has completely evaporated.

The American people do not trust the people in charge enough to be reassured by promises about how this new surveillance infrastructure will be used. And that is good. Because we should not trust the people in charge. They have, fortunately, made that very clear in recent years—which is why we’re seeing such a political revolt against incumbents.

But, going back, all the government power grabs that have brought us to this point—the PATRIOT Act, the invasion of Iraq, the banker bailouts, the insurance industry bailout known as Obamacare, the covid lockdowns, and more—all of it was only possible because enough of the public fell for the lie that the government was acting in their interest.

The fanatical opposition to Flock cameras is evidence that that lie isn’t working right now. Let’s hope that lesson is not easily unlearned.

Tyler Durden Sat, 08/29/2026 - 16:20

Pages