The protest is nationwide and bipartisan, as 86 percent of all Americans oppose data centers in their neighborhoods. The concerns are excessive electricity and water consumption, as well as sound and radiation exposure. Lobbyists at The Data Center Coalition, which argues for AI data centers, are totally tone deaf to the protest, saying they are “mitigating any negative impacts to households and businesses.” Really? ⁃ Patrick Wood, Editor.
Opponents of the rapid buildout of data centers held 142 protests across 42 states on Saturday in the first nationwide effort to channel anger at the AI infrastructure expansion that has ramped up over the past year and roiled local politics.
The protests were coordinated by HumansFirst, co-founded by a former leader of the modern-day Tea Party. She compares growing opposition to data centers to the right-wing populist movement that emerged in 2009 to protest what it saw as excessive taxation and government overreach.
Protesters rallied against what HumansFirst calls the “unaccountable” buildout of data centers and “unacceptable infringement on our liberty.”
In Fredericksburg, Virginia, one young man carried a sign reading, “We thrive on water not data.” In Atlanta, a small knot of protesters wielded signs reading, “Georgia never voted for hyperscale AI data centers.”
Nationwide and city-specific counts of protesters were not immediately available, said Chris Barron, president at Right Turn Strategies, who is handling press relations for HumansFirst.
Attendance was lower than expected in some rural areas and major cities such as Atlanta, local organizers told Reuters. But HumansFirst praised the efforts of community activists. “We were thrilled with the turnout,” Barron said.
Data center opposition is among the few issues uniting Americans across ideological lines, with just a third of Americans approving of the pace of data center construction in the U.S., according to a June Reuters/Ipsos poll. Only 14% of respondents would support a data center being built in their community to support artificial intelligence projects for technology firms such as Meta, Alphabet, Amazon, Microsoft and Elon Musk’s xAI.
“The data center industry is continuing to work with policymakers, stakeholders, and residents to ensure data centers strengthen, not strain, the areas where they operate — while mitigating any negative impacts to households and businesses,” said Josh Levi, president of the Data Center Coalition, the industry’s association and lobbying group.
RED AND BLUE STATES REPRESENTED
Although HumansFirst co-founder Amy Kremer compared her current efforts to the Tea Party movement’s early days in 2009, she said the anger against data centers is nonpartisan.
“They just woke up one day and found out they’re going to have this monstrosity in their community, and they don’t want it,” said Kremer, who predicted data centers will be a defining issue in November’s midterm elections and the 2028 presidential race.
Kremer has criticized Republicans for giving Big Tech a “free pass,” but she and some organizers also said they do not support policies like moratoriums on data center approvals like the one adopted by the Democratic state of New York.
Demands made by organizers from the groups involved include transparency in the development process, protection of resources and environmental health, community benefits such as the creation of well-paying union jobs, and a way to hold developers accountable if they break their promises.
The Republican stronghold of Texas, a hotspot for data center development, hosted 18 protest events, the most of any state. The battleground state of Georgia had 11 rallies. Democratic-dominated California had eight, while the swing state of Pennsylvania and the Republican-leaning states of Florida and Indiana came in at seven each.
Saturday’s Atlanta protest, one of a number of similar actions this month, attracted about a dozen participants who traveled from smaller Georgia towns where the largest data centers are being built, said Jake Watts, 26, a volunteer at that event.
Left-leaning Ivan DelSol, 54, helped lead a protest in the California desert’s Imperial County, where a proposed data center project could use 260 million gallons (984 million liters) of water per year from the Colorado River.
“It’s dystopian that you would use this much fresh water for AI,” said DelSol, who later added around 50 people attended despite temperatures that soared above 100 degrees Fahrenheit (38 degrees Celsius).
Although water is frequently cited as a top public concern, especially in water-stressed regions, the data center industry says its water use is not as significant as other industries.
In Tyler, Texas, first-time activist and self-described “political nomad” Eva Cardona, 31, organized a protest that also drew about a dozen people.
“I’ve been hearing about unregulated AI and the rapid growth was alarming me. I wanted to do something more hands-on than just your standard Facebook post,” Cardona said.
President Donald Trump, meets with Iraqi Prime Minister Ali al-Zaidi in the Oval Office: "We have a tremendous chemistry together... he won an election that ...
President Trump: "We have a tremendous chemistry together... he won an election that a lot of people didn't anticipate... it was very important to me to have...
When society rises against you in vast numbers, when acts of violence are leveled at you to stop what you are doing, when you are forced to hire security guards for your personal security, you would think that anyone with a conscience would stop and consider: “maybe it’s me and not everybody else.” Alas, that is not the response of a narcissistic sociopath, whether it’s a CEO at the top or a city councilman at the bottom. Nevertheless, ordinary citizens outnumber these people by 100,000 to 1.
Psychologists say that you can’t truly “stop” or change a narcissistic sociopath’s behavior. You force them out of their current positions only to see them surface in other areas. ⁃ Patrick Wood, Editor.
In April, someone threw a Molotov cocktail at Sam Altman’s San Francisco home, and within days a second attack put gunfire into the property. The Wall Street Journal reports that AI executives are hardening personal security as opposition to the industry moves from online posts into the physical world.
Prosecutors say Daniel Moreno-Gama, the 20-year-old accused in the first attack, traveled from Texas to San Francisco intending to kill Altman, and had writings on him about AI’s purported risk to humanity. He faces two counts of attempted murder and attempted arson in California state court. Two more suspects were later arrested in connection with the second incident. Around the same time, The Information described Silicon Valley leaning into a new breed of bodyguards for AI leadership.
The pressure isn’t only on the people at the top. According to the Data Center Watch Q1 2026 report, organized opposition groups roughly doubled from 396 at the end of last year to 833 by the end of March, spanning 49 states, and opponents blocked or delayed at least 75 projects worth about $130 billion in a single quarter. That is a very different problem from a viral tweet. It is permits denied, votes lost, sites relocated.
The honest caveat is that these are two different signals lumped into one story. A handful of documented attacks on high-profile figures is not a wave, and some of the loudest ‘threats are rising’ framing comes from firms that sell protection into that fear. What the reporting doesn’t give you is how widely Moreno-Gama’s ideology is actually shared versus one radicalized individual, or how much of the blocked $130 billion in projects gets rerouted rather than cancelled.
What this changes for anyone building in AI is where the friction now lives. Less in Washington policy, more in the county zoning hearing, the executive comp package, and the calculus of whether to be publicly identified with a frontier lab at all. The winners are the security and threat-intelligence firms scaling into that gap, and the labs that treat community relations as an infrastructure problem before it becomes a personal one.
Flock’s initial funders included Arch-Technocrats Marc Andreessen and Peter Thiel. They just ran it up the flagpole and skipped the consent part. Now Flock is embedded in every state in America – all 115,217 cameras, and growing. Kick Flock back to starting gate by joining the nationwide protest on August 16-22 by going to https://noalprs.com/. ⁃ Patrick Wood, Editor.
“You had to live—did live, from habit that became instinct—in the assumption that every sound you made was overheard, and, except in darkness, every movement scrutinized.”—George Orwell, 1984
While Americans remain transfixed by the political circus—cheering for their preferred party, jeering at the opposition, obsessing over every manufactured outrage and waiting for the next spectacle—the Surveillance State continues its steady march forward.
It watches where you go, whom you meet, where you worship, what medical offices you visit, what political rallies you attend, what protests you join, what books you read, what websites you visit and what causes you support.
It watches through your phone, your car, your doorbell, your appliances, your purchases, your social media accounts and the cameras positioned along the roads you travel every day.
This is how freedom dies in the digital police state: not always through dramatic declarations of martial law or soldiers stationed on every street corner, but through the gradual construction of a technological dragnet—an electronic concentration camp—so pervasive that privacy becomes impossible and anonymity becomes suspicious.
Enter Flock Safety, a private surveillance technology company whose automated license plate readers have spread throughout thousands of American communities.
They document the time and location of every passing vehicle and record identifying characteristics such as its make, model, color, damage, roof racks, bumper stickers and other distinctive features. That information can then be placed in a searchable database and used to retrace a vehicle’s movements over time.
Yet the real power—and the real danger—of Flock does not come from the cameras alone.
It comes from artificial intelligence.
A camera can photograph a car. Flock’s AI-powered platform can identify and categorize a vehicle, compare an observation with stored records, generate alerts, identify connections and help police reconstruct where that vehicle has been.
AI is what transforms a photograph into the building blocks for a suspect society.
With AI, every driver becomes a data point. Every data point becomes a pattern. And every pattern becomes a suspicion.
This is how ordinary movements become potentially suspect and subject to government scrutiny. It allows law enforcement agencies to search not only for a complete license plate number but also for partial plates and physical descriptions such as vehicle color, make, model, damage, roof racks, bumper stickers and other identifying characteristics.
A police officer might ask the system to locate every red pickup truck with a ladder rack seen near a protest, every vehicle that repeatedly visited a particular address, or every car observed traveling between two locations.
The artificial intelligence does the sorting. The database supplies the history.
The government receives a list of potential suspects.
This is no longer surveillance conducted by individual officers following particular leads. It is surveillance conducted at machine speed, across entire populations, with algorithms deciding whose movements merit further scrutiny.
That is not targeted policing. That is mass collection.
The overwhelming majority of those scans do not involve stolen cars, wanted suspects, kidnappings or violent crimes. They document ordinary people carrying out the ordinary activities of daily life: driving to work, taking children to school, visiting friends, attending church, keeping medical appointments, participating in protests or simply going home.
Yet each of those innocent journeys becomes part of a searchable police database.
At 20 billion scans a month, Flock is not searching for particular suspects and then attempting to follow them. It is recording the movements of everyone so police can decide later whom they want to follow.
That is the digital equivalent of assigning a government agent to trail every driver in America—and preserving the agent’s notes in case the government someday finds them useful.
Yet mass collection is only the first stage of the AI surveillance state. The next is merging those billions of observations with everything else the government and its corporate partners know about us.
Flock is also part of a much larger shift toward AI-powered “data fusion,” in which license plate records are combined with facial recognition results, surveillance video, police reports, social media activity, commercially purchased information, gunshot-detection alerts and other government databases.
The danger is no longer merely that one system can track a car. It is the merger of previously separate streams of information into a single system capable of mapping a person’s movements, relationships, habits and associations.
These systems increasingly do more than provide officers with information to evaluate. They assign significance to associations, flag supposed threats and generate investigative leads—often through proprietary algorithms that neither the accused nor the public can examine.
Artificial intelligence does not eliminate human prejudice, institutional bias or bad information.
It industrializes them.
Feed a flawed system inaccurate data, biased arrest records or constitutionally suspect surveillance, and AI can reproduce those defects at a speed and scale no individual police officer could match.
Once the computer labels someone suspicious, moreover, officers may treat the algorithmic conclusion as objective fact.
Despite the extraordinary reach of this technology, Flock continues to portray its system as a limited, carefully controlled crime-fighting tool.
Flock insists that its cameras collect information about vehicles rather than people, that agencies control access to their own data, that searches are logged and that information is generally deleted after 30 days. Yet these assurances largely amount to distinctions without a difference.
Vehicles are extensions of the people who drive them.
Track a vehicle long enough, and you know where its owner sleeps, works, worships, shops, socializes, seeks medical treatment and participates in political activity.
You know when someone leaves home, when they return, whom they visit and how often.
You may not know the contents of their conversations, but you know enough to construct an intimate portrait of their life.
That is surveillance.
It does not become less invasive merely because the government has outsourced the cameras, databases and algorithms to a private corporation.
Nor does it cease to be surveillance because police claim that the information may someday be useful in solving a crime.
Indeed, that is the sleight of hand that has allowed the surveillance state to expand so rapidly.
The government no longer has to install every camera, maintain every database or directly collect every piece of information.
It merely encourages private companies, businesses, homeowners’ associations, schools and individual consumers to create an interconnected surveillance ecosystem—and then asks for access.
This public-private arrangement allows government agencies to acquire capabilities they might never receive public approval or sufficient funding to build on their own.
It also makes accountability almost impossible.
When abuses occur, local police blame the technology provider. The technology provider insists that local police control the data. Federal agencies claim they merely requested access. Local officials say they were unaware that information could be shared beyond their jurisdiction.
Everyone points elsewhere.
Meanwhile, the American people remain under observation.
Flock has become especially controversial because its network can transform what appears to be a collection of local cameras into something far more powerful: a searchable surveillance system that permits law enforcement agencies to look far beyond their own jurisdictions.
Flock says data sharing among agencies is optional and controlled by its customers. Yet the entire value of such a system lies in its interconnectedness.
A camera in one town is a traffic-monitoring device.
Thousands of cameras connected through searchable databases constitute a movement-tracking network.
The danger is not simply that police might search for a stolen car.
The danger is that the system permits government officials to begin with a location, a description or a fragment of information and work backward until someone emerges as a suspect.
That reverses the traditional order of constitutional policing.
Under the Fourth Amendment, police are supposed to develop individualized suspicion, establish probable cause and then apply for a warrant to search for evidence connected to a particular person or crime.
Mass surveillance systems begin by collecting information on everyone.
In the process, every innocent person is treated as a potential suspect whose movements must be recorded just in case the government someday decides they are relevant.
This is guilt by algorithm.
It is also the same constitutional inversion at the heart of geofence warrants, which allow police to demand information identifying every cellphone that happened to be near a particular location at a particular time.
The case involved a geofence warrant used to obtain Google location records for cellphones near the scene of a robbery. Rather than beginning with an identified suspect, police demanded information about devices that happened to be within a designated area during a particular period and then worked backward to identify their owners.
The Supreme Court held that police conduct a Fourth Amendment search when they obtain an individual’s cellphone location history from a technology company.
That conclusion matters.
It rejects the government’s increasingly convenient argument that intimate information loses constitutional protection merely because a private corporation collected, stored or analyzed it.
The Court did not rule on Flock cameras or automated license plate databases. Nor did it decide that every geofence demand is necessarily unconstitutional. The justices left it to the Fourth Circuit to determine whether the warrant satisfied the Fourth Amendment’s probable-cause and particularity requirements at each stage of the search.
Nevertheless, the constitutional principle at the heart of Chatrie extends far beyond cellphones.
The government should not be able to evade the Fourth Amendment by outsourcing mass surveillance to private technology companies.
It should not matter whether the location trail comes from Google, Flock, a cellphone provider, a data broker or an interconnected network of privately owned cameras.
A detailed record of a person’s movements does not become less revealing because it follows a vehicle rather than a phone. The government should not be permitted to accomplish through Flock what it could not constitutionally accomplish by assigning police officers to follow millions of Americans everywhere they drive.
Indeed, Flock may present an even more troubling inversion of constitutional policing.
Geofence searches generally begin with a particular crime, location and period. Flock continuously collects information on millions of vehicles before any crime has occurred and before any individual is suspected of wrongdoing.
Police can then reach backward into that stored history and reconstruct a person’s movements.
The surveillance comes first. Suspicion comes later.
A warrant, when one is sought at all, may arrive only after the government has already built the database it intends to search.
The technology is already embedded in thousands of communities.
The databases are already being populated.
The agencies are already connected.
And the companies profiting from this infrastructure will fight to preserve it.
Unfortunately, constitutional protections have rarely kept pace with the government’s appetite for surveillance.
The dangers are no longer theoretical.
Flock data has reportedly been used in investigations far removed from the serious violent crimes routinely invoked to justify these systems.
This is the inevitable trajectory of every surveillance technology. First, it is introduced as an emergency measure. Then it is justified as a crime-fighting tool. Then it is expanded to lesser crimes. Then it is used for administrative enforcement, political monitoring, immigration investigations and personal purposes.
Eventually, it becomes part of the background machinery of government—a permanent feature of daily life that no longer attracts attention because everyone has become accustomed to being watched.
Surveillance powers created to find kidnappers and violent criminals do not remain limited to kidnappers and violent criminals.
Databases built to locate stolen vehicles do not remain limited to stolen vehicles.
Government agencies cannot resist the temptation to use whatever power is available to them, especially when the use of that power is cheap, easy and largely hidden from the public.
The technology’s potential for error makes this even more dangerous.
License plate readers can misread plates, rely on inaccurate hot lists or associate an innocent vehicle with a crime. Once the system issues an alert, officers may treat the computer-generated result as fact.
The individual on the receiving end may be pulled over, surrounded by armed police, handcuffed, searched or detained before anyone discovers that the machine was wrong.
This is not justice. It is automated suspicion.
Flock is only one component of a surveillance ecosystem that includes doorbell cameras, facial recognition, drones, cellphone tracking, biometric databases and real-time crime centers.
A person may leave a home monitored by a smart doorbell, drive past a network of license plate readers, enter a business equipped with facial recognition, carry a phone broadcasting location data and return home along streets monitored by police cameras and private security systems.
At no point does the government need to physically follow that individual, because the infrastructure does it automatically.
Algorithms sort the information. Databases preserve it. Private companies monetize it. Government agencies search it.
All of this is taking place while the country remains locked in an endless partisan cage match.
Both parties have contributed to the Surveillance State. Both parties have expanded it. Both parties have exploited fear to convince the public that freedom must be sacrificed for safety.
The targets may change depending on who is in power, but the machinery remains.
Once the infrastructure exists, there is no guarantee that it will be used only against people you dislike or with whom you disagree politically.
That is the lesson Americans repeatedly refuse to learn.
A surveillance tool created by one administration will be inherited by the next. A database assembled for one purpose will inevitably be used for another. A system established to monitor “them” will eventually be turned against “us.”
Communities across the country are finally beginning to recognize the danger.
Some cities have terminated or declined to renew their Flock contracts. Others have paused deployments or demanded stronger restrictions on data sharing, retention and federal access.
This resistance is long overdue.
We cannot afford to become so distracted by the theater of politics that we fail to notice the architecture of tyranny being assembled around us.
The surveillance state does not care which party you support. It does not care whom you voted for.
It does not care whether you believe you have nothing to hide.
The cameras are watching. The databases are growing. The networks are connecting.
President Trump wears many hats and one of them is Technocrat-in-Chief. He lobbies the Senate for the Arch-Technocrats in Washington, DC., not with a heavy-handed approach as before, but appealing to sentiment for the late Senator Lindsey Graham who was “a big supporter.” If passed, the CLARITY Act will cement Technocracy to the foundation of our country.
Now is the time to put maximum pressure on Senators to VOTE NO on the CLARITY Act! ⁃ Patrick Wood, Editor.
The Senate returns to Washington on July 13, with the clock running down on the most consequential piece of crypto legislation in years. Lawmakers now have roughly four weeks to schedule, debate, and pass the CLARITY Act before the August recess.
President Trump weighed in directly on Monday, posting on Truth Social that “in honor of Senator Lindsey Graham, a big supporter, the U.S. Senate should pass the Clarity Act” and warning that China and other countries “would like to take complete and total control of this major financial ‘happening,’” as well as A.I.
White House crypto adviser Patrick Witt amplified the urgency, noting the critical week coincides with the one-year anniversary of the GENIUS Act and cautioning, “We cannot afford to delay any longer.”
This is a window many policy watchers see as the last realistic chance to enact comprehensive digital-asset market structure legislation this Congress.
The CLARITY Act would draw a firm regulatory line between the SEC and the CFTC, granting the commodities regulator exclusive jurisdiction over spot markets for “digital commodities” while leaving the SEC to oversee investment-contract assets.
It cleared the House in July 2025 by a bipartisan 294–134 vote and advanced out of the Senate Banking Committee in May by a 15-9 margin, with two Democrats joining all Republicans.
Those committee votes, however, came with warnings that floor support was not guaranteed.
This week’s milestone is the release of updated text merging the Senate Banking and Agriculture Committee versions, the clearest signal yet of what survived negotiations and what remains unsettled.
JUST IN: 🇺🇸 President Trump says “The U.S. Senate should pass the Clarity Act.” 🚀 pic.twitter.com/9Y7VxKZ3ck
The bill missed the July 4 signing ceremony that White House crypto adviser Patrick Witt had targeted, and while meetings ran through the recess, the thorniest issues remain unresolved, according to Crypto in America. Getting to 60 votes may prove harder than getting this far, and with the Republican conference shrinking, Democratic buy-in matters more than ever.
Chief among them is the Blockchain Regulatory Certainty Act, folded into the CLARITY Act as Section 604, which would shield non-custodial software developers from being treated as money transmitters.
Law enforcement groups argue the language, as written, would hamper investigations into on-chain crime, and Democratic support may hinge on revisions.
An ethics standoff
The more explosive fight is over ethics. Negotiators have yet to reach a CLARITY Act deal with the White House on guardrails around conflicts of interest tied to President Trump’s crypto ventures, after disclosures showed he earned more than $1 billion from crypto-related businesses last year.
House members have pressed the Senate to act while addressing those concerns, and a coalition of more than 200 companies has urged leadership to bring the bill to the floor. The coalition argued that the bill would establish a clear federal framework for digital assets and help keep innovation in the U.S.
Complicating the math, the death of Senator Lindsey Graham (R-SC) and the continued absence of Mitch McConnell (R-KY) leave Republicans with almost no room for error in reaching 60 votes.
Sentiment is split. Solana Policy Institute President Kristin Smith says momentum is building and a floor vote before recess remains achievable, echoing CFTC leadership calling the bill “so close.”
Others are wary: Galaxy Digital cut its passage odds to 50-50, citing the shrinking calendar and competing priorities like the NDAA. The firm said the legislation still faces procedural hurdles, unresolved ethics and developer-protection disputes, and a crowded Senate agenda that could delay consideration until September. Galaxy said the odds would improve if Senate leaders commit to a July vote. Odds were as high as 70% earlier this year.
The next four weeks may be CLARITY’s last chance in the 119th Congress.
For the better part of the last two years, I have been tracking what appeared to be a reasonably predictable pattern in AI development: a doubling of capability roughly every 3.5 months. That figure came from METR’s time horizon benchmarks — measurements of how long an AI agent can work autonomously on a task before failing. Early 2024 through early 2026, the data held with uncomfortable consistency. If you plotted it, the curve bent upward with almost mechanical precision.
I say “appeared to be predictable” because that framing is now obsolete.
The variable that breaks every forecast model is Recursive Self-Improvement — the condition in which AI systems are no longer just tools that humans use to build AI, but active participants in building themselves. We crossed that threshold. The question of when is already behind us. The question now is what happens when a system that rewrites its own code, runs its own experiments, and optimizes its own training recipes starts doing so faster than any human team could direct it.
There is no clean answer. That is precisely the point.
What Has Actually Changed
Let me be specific, because vague gestures toward “exponential growth” have become their own form of intellectual laziness.
As of May 2026, Anthropic confirmed that more than 80 percent of the code merged into its own production systems was written by Claude — its own AI. Not assisted by Claude. Written by Claude. The company’s own engineers have described the shift as moving from doing work to managing a system that does the work. One Anthropic engineer publicly stated that 100 percent of his personal code output was AI-generated, with 22 pull requests shipped in a single day.
OpenAI’s announcement of GPT-5.6 in July 2026 included a data point that deserves more attention than it received: over the previous six months, the share of internal research compute devoted to AI coding inference grew one hundredfold, while internal agentic token usage — meaning AI agents working autonomously inside OpenAI’s own research infrastructure — increased twenty-two fold. OpenAI is now using its own frontier models to diagnose training failures, optimize training systems, run experiments, interpret results, tune computational kernels, and improve training recipes for the next model. The company described this as “quickly becoming standard.”
Google reported in early 2026 that 75 percent of all new code at the company is AI-generated, up from 25 percent in 2024.
These are not productivity statistics. These are evidence that the loop has closed. AI is now a primary agent in its own development cycle.
Why the 3.5-Month Doubling Figure Was Always a Floor, Not a Ceiling
When I cited the 3.5-month capability doubling figure, I was describing what the data showed for models built primarily by human engineers using AI as an accelerant. That is a fundamentally different situation from what is unfolding now.
RSI compresses timelines in a way that no benchmark extrapolation can capture in advance. Here is the basic mechanic: if AI doubles in capability every X months under human-directed development, and AI is now directing a significant fraction of its own development, then the effective time to the next doubling shrinks in proportion to how much of the development loop the AI controls.
If AI handles 80 percent of code output and that percentage is rising, you are not looking at a linear compression of the doubling time. You are looking at a feedback loop where each generation of AI produces a more capable successor faster than the previous generation did — and that successor inherits the full research infrastructure of its predecessor.
Google DeepMind’s June 2026 paper, From AGI to ASI, described the unconstrained version of this as potentially “hyperbolic” — meaning super-exponential, a curve that in theory races toward a singularity. The authors were careful to note that real-world resource constraints bend such curves into S-shapes before they go vertical. Compute costs money. Power requires physical infrastructure. Fabricating chips takes years of supply chain work. These are genuine friction points.
But friction points are not stopping points. They slow the curve. They do not reverse it.
Breakthroughs At Any Time
Here is where intellectual honesty requires admitting the limits of any analysis, including mine.
RSI does not just accelerate known processes. It creates conditions for qualitative leaps — the kind of change that looks, in retrospect, like it came from nowhere. The history of science is full of these moments. They are not random, but they are not predictable either. You cannot model the arrival of a genuinely new idea. You can create the conditions that make such ideas more likely, and an AI system running millions of research cycles per day, improving its own ability to run those cycles, is precisely such a condition.
What this means practically is that the question “when will ASI arrive?” is not answerable with a date. The consensus among serious forecasters — not YouTube thumbnails, but the AI-2027 team, the METR researchers, the Google DeepMind paper authors — clusters around AGI in the 2026-2027 range and ASI following within months to a few years. But those estimates assume that progress continues on a roughly continuous curve. RSI introduces the possibility of discontinuous jumps — moments where capability does not inch forward but lurches.
Nobody knows when those moments arrive. That is not a failure of analysis. That is the nature of the phenomenon.
What This Means Beyond the Lab
The implications of RSI extend well past questions of benchmark performance or corporate strategy. They reach into every institutional structure that assumes human cognitive supremacy — which is to say, every institutional structure that exists.
Legal systems assume that humans write laws, interpret them, and enforce them. Economic systems assume human judgment at key decision points. Democratic governance assumes a human electorate making decisions about a world they can understand. All of these assumptions are being stress-tested simultaneously by a technology that is now improving itself at a rate that exceeds the capacity of any regulatory body to track, much less manage.
The same politicians and regulators who failed to anticipate the social consequences of social media algorithms — a comparatively simple technology — are now being asked to govern RSI. This should concern everyone, regardless of their position on the political spectrum. This is not a left-right question. It is a question of institutional competence in the face of something genuinely unprecedented.
Sam Altman wrote in mid-2025 that “we are past the event horizon; the takeoff has started.” That is as close to a plain statement of fact as you will get from a sitting AI lab CEO. The event horizon metaphor is apt. Past a certain point, events inside cannot be communicated outward in a way that allows course correction.
We may or may not be past that point. What is certain is that the institutions charged with maintaining human oversight of this process were not built for it, are not staffed for it, and show no signs of being reformed fast enough to matter.
Conclusion
Anyone who tells you they know exactly how this unfolds is selling something. The honest position is this: AI capability is growing faster than any previous technology in history, the development loop has partially closed, and the factors that could produce discontinuous breakthroughs are now structurally embedded in the research infrastructure of every major AI lab on the planet.
The 3.5-month doubling was a data point describing yesterday’s trajectory. RSI means that the trajectory is now self-modifying. The curve is not just bending upward — it is bending the conditions that determine how fast it bends.
There are serious people who believe this ends well for humanity. There are equally serious people who believe it does not. What there is not, on either side, is certainty. The responsible course is to watch what the systems are actually doing — not what the press releases say — and to resist the temptation to normalize a situation that is, by any historical standard, abnormal.
What Technocracy can’t control, it crushes. No outliers allowed. Chinese Christians only meet underground in order to escape prison, or worse. As the “system” of Technocracy in China gets more entrenched, the number of house churches increases and intolerance builds against those those who pledge allegiance to Christ and the Bible, not to the system.
China has “official” churches that are not harassed because they are heavily controlled. Similarly in America, a state religion has arisen around the Trump Administration based on an mixture of Dominionism, Resconstructionism, and New Apostolic Reformation. Pressure is mounting against outliers who hold traditional views of Christianity, especially those who study Bible prophesy.
Christians in the underground church China and in America share a common view that Technocracy is morphing into the “beast system” that is seen in Revelation chapter 13. This is likely the real bone of contention in the Technocrat mind.
It should be noted that China also considers Falun Gong and the Uyghur Muslims to be “enemies of the state.” Both are critical of Chinese Technocracy. ⁃ Patrick Wood, Editor.
When the power cut out and his signal dropped, TJ knew something was wrong.
Seconds later, the banging started on his front door. Terrified, he stood as still as possible with his wife and three-year-old daughter, hoping they wouldn’t be noticed.
But the Chinese police officers broke down the door, shoved his wife and child into a separate room and began to question TJ.
“They grabbed my clothes and grabbed my hands so I couldn’t move. I could hear my daughter crying so much in the room next door but I couldn’t go to her, I couldn’t hug my wife,” he told The Telegraph.
TJ knew what his family’s crime was: being Christian and worshipping a God that was not Xi Jinping. In China, following a church that is not state-controlled is punishable.
The Chinese leader is intensifying Beijing’s crackdown on Christians amid a wider purge of top officials, showing signs of an increasingly paranoid leader.
The country officially recognizes five religions, including Protestantism and Catholicism, but this only extends to churches that are fully state-controlled, where the congregation is expected to sing patriotic songs before every service and affix Mr Xi’s portrait above the pulpit.
Many Christians such as TJ, who withheld his full name for security reasons, and his wife have chosen to join unofficial churches – or underground churches – where they can preach the gospel away from the government’s oversight.
But attending these places of worship carries its own risks – not least because they are seen as traitors.
TJ last saw his wife when she was taken to a police station along with their phones, some books and artwork, and she has yet to be released. He still doesn’t understand why he was not taken too.
Under Mr Xi’s iron-tight grip, China has expanded its nationwide suppression of Christians during the last decade, arresting more than 10,000 people, according to Bob Fu, the founder of ChinaAid, a charity for victims of persecution in the country.
In the most recent crackdown, armed police stormed the Early Rain Covenant, an influential underground church, and detained more than 30 members last month.
Mr Xi’s ruthless campaign against these underground churches aims to ensure the survival of the Chinese Communist Party (CCP) and remove any threat to his power.
Mr Fu said: “It’s the emperor playing God. [Mr Xi] wants to be exclusive, he doesn’t want to have anything treated or worshipped more superior than him.”
TJ is one of six Chinese Christians who spoke to The Telegraph who have either been directly targeted by the CCP or have close relatives that are incarcerated.
They described police officers showing up at their homes unannounced in the middle of the night. Friends being rounded up and questioned by authorities, sometimes for weeks on end. Loved ones being convicted on trumped-up charges such as “using superstition to undermine the law” and detained indefinitely in crammed, dirty cells. And lawyers were targeted and suspended from practicing law for defending Christians.
Knocks at the door
Jun Yang, a pastor with Zion Church, one of the largest underground churches in China, knows all too well about the risk of living as a Christian in China.
Nearly 30 members of his church, including Qu Qiuyu, his wife, and Ezra Jin Mingri, the church’s leader, were detained in October last year during one of the largest raids against Christians in recent years.
Mr Jin was released in early July, but many of those detained remain in prison.
Founded in 2007, Zion Church used to operate in a converted nightclub in Beijing but was forced to move to a decentralised, hybrid format 10 years later after the mass arrests of Christians across the underground church network and restrictions during the Covid-19 pandemic forced many of the church’s sessions online.
The church’s membership has grown from 1,500 in 2018 to around 5,000 followers now, since the church adopted the hybrid format.
Mr Yang and other Zion Church members who spoke to The Telegraph said the arrests in October were not a complete surprise. Police had been harassing church members for months before the arrests and had been coming up with incriminating information about the church leaders.
In Mr Yang’s case, they were trying to pin corruption charges on him.
“The police reached out to church followers, even took them back to the police station for questioning. They asked them about donations to the church, whether such donations went into my pocket to buy apartments, or if I had misused them in any way,” he said.
“They were trying to look for excuses to see if they could set a trap for me,” he added.
Mr Yang was visiting South Korea when they raided his family home. Officers arrested his wife and hauled her away in front of their two young children.
Gao Yingjia, another pastor from Zion Church, was also targeted in the October 2025 raids, using similar tactics.
His wife, Geng Pengpeng, told The Telegraph that they had been awoken at 1.30am by loud knocks on the door. The police stormed in, handcuffed her husband and took him away.
“We both knew something like this might happen,” said Ms Geng. “All of our rental [church] locations were shut down and sometimes they would take someone to the detention centre for a week or 15 days.”
Neither Ms Geng nor Mr Yang has spoken to their spouses since they were detained.
For Mr Yang, while he counts himself lucky in some ways, he would do anything to go back in time. “If there was the chance, I would rather be the one arrested than my wife,” he said.
But as much as he wants to see his wife and children, he knows he may never get the chance to return to China.
“I am on the wanted list of religious practitioners so if I return to China I would be arrested,” said Mr Yang.
A threat to the throne
While Christians have long been treated as outsiders in China, when Mr Xi came to power in 2012, he piled on policies to further assimilate Christians and other religious groups in the country.
In 2015, he launched a “Sinicisation” campaign, which forced all religious and ethnic groups to assimilate and prioritise loyalty to the CCP over individual religious beliefs.
The same Sinicisation policy has been used to imprison and torture Uyghur Muslims in what has been characterised by many human rights organisations as genocide.
In 2018, Beijing rolled out a five-year plan to target Christians, which included censoring sermons, controlling church donations, supervising Bible translations and including “Xi Jinping Thought” in the curriculum of seminary schools.
They also systematically closed underground churches and demolished houses of worship as part of the campaign against worshippers outside the state’s control.
Those who agreed joined China’s two official churches – the Catholic Patriotic Association and the Protestant Three Self Patriotic Movement – while the rest remained part of the underground church network.
Today, there are an estimated 44 million members of the state-sanctioned church and around 115 million unregistered Christians, which are expected to double by 2030.
“My dad was approached many times about the [Three Self churches], and he said no, just like many others because they’re fake churches, to put it bluntly,” said Gao Pu, the son of Gao Quanfu, the incarcerated founder of the Light of Zion Church (different from Zion Church), and Pang Yu.
“On paper, the CCP is still an atheist entity so there’s an element where they fear what they cannot control and don’t understand – or don’t want to understand,” explained Mr Gao.
“Anybody that’s possibly out of their control, they tend to crack down on them very, very fast. They learnt the lessons from the past,” he said, referring to the 1989 student uprisings that were met with violent responses by police at Tiananmen Square.
Weaponizing the court
Under the Sinicisation policy, Chinese police have arrested Christians and isolated them in detention.
Many of those arrested, including Mr Yang and Ms Geng’s spouses, were initially charged with the “illegal use of information networks”, a law that broadly covers any communication connected to “illegal or criminal acts”.
Typically, the charge carries a maximum sentence of three years, but is often combined with additional charges that can carry up to life imprisonment.
Authorities have also accused Christians of “using superstition to undermine the law”, for which the sentence can range from three years to an “indefinite imprisonment”.
However, in some of the cases described to The Telegraph, including those of Ms Geng’s husband and Mr Gao’s parents, authorities adapted the initial charges to fraud instead.
“They change to fraud because it’s much easier for them to slap a number on there and give a sentence, which is actually very heavy,” said Mr Gao.
He told The Telegraph that he did not know how long his parents would be in prison. While they could face more than 10 years, he hopes, given their age – both are in their late 60s – they will be released after only three years.
For Mr Yang’s wife, the charge was updated to illegally operating a business, which carries a longer sentence. However, Mr Yang has no idea how long the sentence will be.
Part of the uncertainty around sentencing is because of a lack of fair legal representation.
Several people who spoke to The Telegraph said that the government had revoked the licences of lawyers who had been hired to represent Christians.
Zhang Kai, whose law firm was working with many of the Zion Church members, including Mr Jin, had his licence revoked recently because he allegedly “disrupted order in the court”.
Several other lawyers at his firm were also forced to step down, including those working on Mr Yang’s wife’s case.
The Chinese government has previously targeted lawyers who have taken on politically sensitive cases, with some suspended and others thrown in prison.
Between 2017 and 2023, at least 30 human rights lawyers had their licences revoked, according to Frontline Defenders, the Irish human rights organisation.
In response to The Telegraph, a spokesman from the Chinese embassy in the UK said it “manages religious affairs in accordance with the law”.
They added that “anti-China forces” had used the guise of religious freedom and human rights to engage in “political manipulation, made sweeping generalisations, maliciously maligned China’s ethnic and religious policies, and fabricated and disseminated disinformation”.
They also said that “judicial organs in China handle cases strictly in accordance with the law and fully safeguard the lawful rights of criminal suspects and defendants throughout judicial proceedings”.
For those watching from a distance as this reality unfolds, it is hard to imagine the circumstances improving any time soon.
Ren Ruiting, a member of the Early Rain Covenant Church who fled to the US following the 2018 crackdowns, said that even thousands of miles from China, she still felt the watchful eye of the CCP.
“If I speak out, they’ll know immediately and they have some way to make me feel bothered and make me feel bad,” said Ms Ruiting.
Like Mr Yang, she doesn’t think she will ever go back to China.
“Once Xi Jinping is still the chairman in China, I don’t think anything will be better and if the persecution is not better, that means we cannot have our church and we are in danger anytime.”
CORRECTION: 2nd graphic in the video at 0:21 should state, "Opening Remarks"At a House Energy and Commerce hearing, lawmakers questioned Boston Dynamics Chie...
The CEO of Flock Safety made a comment that may have revealed more about America's growing surveillance industry than he intended. In this investigation, Lau...
Photonics influences cell behavior by using light. Technocrats have discovered that many cells, or genetically modified cells, express light-sensitive proteins (opsins, rhodopsins) that open or close ion channels when illuminated, changing membrane potential and firing patterns. In neurons, this directly controls excitability, synaptic transmission, and network dynamics on millisecond timescales. ⁃ Patrick Wood, Editor.
Cancer has a dirty trick: it can put itself to sleep. When tumor cells slip into a kind of biological hibernation, they become hard to kill, shrugging off treatment and lying low until conditions improve, then waking up and bringing the disease back. For decades, researchers have struggled to shut down this hiding strategy without causing serious harm elsewhere in the body. A team in Switzerland has now built a molecule that flips on and off with flashes of light, giving scientists a precise new way to probe, and possibly disrupt, the way sleeping cancer cells hide.
Behind this cellular sleep state, at least in certain cancers, sits a protein called the glucocorticoid receptor, a sensor inside cells that reacts to stress hormones. When it switches on, it can push cancer cells, especially in some solid tumors such as lung cancer, into a drug-resistant, dormant state. The obvious fix would be to destroy the receptor outright, but there is a catch: the same receptor does important jobs all over the body, including calming inflammation. Removing it everywhere would cause real damage. What was needed was a way to hit the receptor inside a tumor and leave the rest of the body alone.
That is what the team set out to build. Their solution, described in the journal Proceedings of the National Academy of Sciences, is a class of molecules called photoPROTACs: drug-like compounds engineered to destroy the stress hormone receptor, but only when a specific color of light switches them on. Shine the right light on them, and they go to work. Keep them in the dark, and they stay idle.
A Kill Switch for a Stress Hormone Receptor
It helps to know what a PROTAC is. A PROTAC is a molecule that grabs onto a specific protein inside a cell and tags it for destruction by the cell’s own waste-disposal system. It works like slapping a “delete me” sticky note on an unwanted protein. PROTACs have stirred huge excitement in medicine because they can eliminate proteins that ordinary drugs cannot touch.
Conventional PROTACs have a problem: they are always on. Once taken, they work throughout the entire body, destroying their target in cancer cells and healthy cells alike. For a receptor this important to normal health, that is a dealbreaker.
To get around it, the team embedded a light-sensitive part directly into the PROTAC molecule. By adding a chemical structure that physically changes shape depending on the wavelength of light hitting it, they created a molecular switch. In one shape, called the E-form, the molecule has the right geometry to do its job and destroy the receptor. In the other, the Z-form, it bends into a configuration that leaves it far less active. The switch runs both ways: one wavelength turns the molecule on, another turns it off.
Building the Light-Activated Cancer Switch
Researchers built and tested several versions of these light-switchable molecules, hunting for the best balance of stability and responsiveness. Two lead compounds, KH-5-306 and KH-5-309, stood out. Both flipped almost completely between their active and inactive shapes under different wavelengths of light and showed no signs of wearing out after repeated exposure.
Inside cells, both compounds destroyed the stress hormone receptor in their active E-form at very low doses. Switched to the inactive Z-form, the effect dropped sharply, though the inactive form still held on to some activity. Because the Z-form naturally drifts back toward the active E-form at body temperature, researchers kept it locked in place during experiments by hitting the cells with brief bursts of ultraviolet light every 30 minutes, a lab-only workaround they confirmed did not harm the cells or disturb the biology under study, and not a method meant for patients.
Additional tests confirmed the receptor destruction ran through the cell’s normal disposal machinery: when a separate inhibitor blocked that machinery, the photoPROTACs stopped working. The compounds also left a closely related receptor alone, even at higher doses, a sign the molecules were hitting their intended target and not damaging look-alikes. And the effect proved temporary. When the active compound was washed away after 12 hours, receptor levels started climbing back within 24.
Infographic by StudyFinds
Waking Up Sleeping Lung Cancer Cells
With the chemistry checked, researchers moved to a test closer to real medicine. They took human lung cancer cells and treated them with dexamethasone, a common steroid that switches on the glucocorticoid receptor and nudges these cells into a dormancy-like state. That setup mirrors a real problem: dexamethasone is routinely given to cancer patients to manage inflammation and treatment side effects, yet it may quietly push tumor cells toward this drug-resistant sleep.
After confirming that dexamethasone triggered the expected dormancy response, the team tested whether the lead photoPROTAC, KH-5-309, could reverse it. Cells treated with the active E-form saw the dormancy-linked gene activity roll back toward normal, and a sweep across thousands of genes confirmed the shift toward a more active state. By contrast, the inactive Z-form left those genes essentially untouched, behaving in this experiment as a neutral control on par with no treatment at all.
What’s Next for Light-Controlled Cancer Treatment
Researchers are candid about the limits. The biggest obstacle to reaching real patients is the kind of light involved. The wavelengths that switch these compounds on do not travel far into living tissue, so the current version could not reach a tumor buried deep in the body. Developing compounds that respond to longer, redder wavelengths, such as near-infrared light that penetrates tissue more deeply, is a key goal for future work rather than a solved problem.
For now, the strongest promise is in the lab: studying cancer biology in cell cultures, tissue slices, and lab-grown miniature organs called organoids, where light can be aimed precisely and repeatedly. Those tools give scientists a way to study how stress hormone signaling drives cancer behavior that existing drugs cannot match.
Cancer’s ability to play dead has long been one of the most frustrating barriers to lasting cures. This is early-stage work, years from any use in the clinic, but a well-aimed flash of light that disrupts cancer’s hiding strategy is the kind of proof-of-concept result that hands researchers a fresh grip on a stubborn problem.
Login
