Back when I was doing regular Texas vs. California updates, this is the sort of video I would feature. It covers why Texas is doing so much better than Europe, though the framing misses a few things I’ve tried to highlight below.
Caveat: I don’t know who “The Economic Matrix” is, but what they’re saying is generally right, but a bit incomplete.
“This is Texas. To most of the world, it’s just one of 50 American states. But if you pulled it off the US map and dropped it into the global rankings as its own nation, it would sit eighth in the world, sandwiched directly between France and Italy, with a staggering GDP of $2.77 trillion. But that’s just the start. In 2025, the Texas economy was bigger than the equivalent of the Netherlands, Belgium, Denmark, and Austria combined.”
“Texas and Europe are separated by more than an ocean. They’re separated by two completely different ideas of how a state should operate. And right now, one of those ideas is winning by a distance that gets harder to close every year. In 2024, the Texas economy grew by nearly 4%. The entire European Union managed 1%, and the gap is only getting wider.”
“The EU spent most of the last 15 years in crisis mode. A sovereign debt collapse left Greece, Spain, and Portugal on the edge of ruin.” I covered the European Debt Crisis (especially among the PIIGS (Portugal, Italy, Ireland, Greece, Spain)) as it was happening, and the thing to remember is that it was (and is) a deficit spending crisis. Like the federal government, European governments insist on spending more than they take in. Real austerity, i.e. cutting outlays until they match receipts, hasn’t failed, it’s been declared difficult and left untried.
“Years of near zero growth followed. The in 2022, Russia cut off the gas and sent inflation across the Euro zone spiking to 9.2%.” Here the video also avoids noting that another big inflation driver was the effects of the Flu Manchu lockdown across most EU economies.
“Through all of that, the Texas economy just kept climbing. The productivity gap tells the same story. Between late 2019 and mid 2024, labor productivity per hour in the Euro zone rose by 0.9%. In the US, it rose by 6.7%. Texas led that charge.”
“Look at the Permian Basin. Oil production nearly tripled in a decade while the rig count was cut almost in half because horizontal drilling and AI-guided extraction meant fewer rigs producing more oil. Same workforce, three times the output.” Oil industry-specific AI has very little do with the current general AI build-out bubble.
“There’s a mathematical reality that makes this trajectory almost impossible to reverse. At 1.5% annual growth, the EU takes roughly 47 years to double in size. At the 3.5% rate Texas usually averages, it takes 20. By the time the EU doubles once, Texas will have doubled twice. That gap compounds and it means every year the distance between these two economic models doesn’t just persist, it accelerates.”
“Mario Draghi, the former head of the European Central Bank, released a landmark report warning that without serious reform, the EU is heading toward what he called a slow agony. Leaders held a retreat to discuss it. Then they went back to their committees.” Future pain is abstract, while the electoral pain of trying to reform things is far more immediate.
“Since 2020, more than 200 companies have moved their headquarters to Texas. Tesla relocated to Austin in 2021. Chevron, one of the largest energy companies on Earth, announced its move to Houston in 2024. Charles Schwab, CBRE, SpaceX. More than half of these relocations came from California alone.”
“These are not satellite offices or mailbox moves. The reason they gave was simple. The regulatory environment in California made expansion too slow and too expensive. Texas made it fast, cheap, and permanent.”
“Spotify was founded in Stockholm, but listed on the New York Stock Exchange. Skype was built in Estonia and Luxembourg, then bought by Microsoft and absorbed into Redmond. ARM was designed in Cambridge, England, then acquired by a Japanese conglomerate and listed in New York. Europe keeps building the talent. America keeps cashing the check.”
“Beneath the growth stats and the corporate migrations, there’s one factor that explains this gap better than anything else: Energy. Texas produces more energy than almost any country on Earth. Not other American states, actual nations. Between 2007 and 2023, while the rest of the United States saw energy consumption drop by about 5%, Texas went in the other direction. Energy use in the state climbed by 21% and the industrial sector alone saw a 28% jump in demand. This was not a state focused on conservation or cutting back. Texas was building oil rigs, refineries, chemical plants, and massive wind installations, all at the same time.” Those wind farm installations were the result of subsidies, and they’re not really building new ones anymore.
“The reason Texas could pull this off comes down to one decision made decades ago. Texas built its own power grid, specifically to escape the slow motion gears of federal regulation. The result, solar capacity that grew 32% in just 2024, wind generation that leads every other American state, and a massive natural gas fleet running underneath it all to keep the lights on when the sun goes down, cheap, abundant, predictable, and fast to build.” Again, the solar build-out was aided by subsidies.
“For decades, the European energy models rested on one assumption. Buy cheap gas from Russia, build out renewables slowly, and keep industrial costs manageable. When Russia invaded Ukraine in 2022, that assumption collapsed overnight. Wholesale energy prices went vertical. Industrial electricity costs in Germany suddenly hit three to four times what businesses in Texas were paying for the exact same power.”
“Look at what happened to BASF, the largest chemical company in the world. For over 150 years, their home was Ludwigshafen, Germany, a sprawling industrial complex on the Rhine that employed tens of thousands. After 2022, they announced billions in cuts to that site. Plants shuttered, thousands of jobs gone. But BASF didn’t disappear. At the exact same time, they were breaking ground on new facilities in Freeport, Texas. Same company, same products, two completely different decisions driven entirely by the price of electricity.” Probably not just electricity. Union work rules in Germany are considerably less flexible than those in right-to-work Texas.
“France tried to escape this trap by leaning into nuclear power, which once covered 75% of French electricity needs. But that infrastructure is aging. New reactor projects like Flamenville have run tens of billions over budget and more than a decade behind schedule, and the political will to build more has been stuck in debate for a generation. The old Russian gas model is dead. The nuclear renaissance has not arrived. And permitting a single wind farm in the EU can take 7 to 10 years.”
“In Texas, the same process often takes a few months. Energy prices act like a hidden tax on everything from manufacturing steel to running a server farm to heating a bakery. European businesses pay that tax every single day. And Europe does not have the gas, does not have the grid independence, and does not have the permitting speed to change that.” Actually, Europe does have oil and gas reserves it refuses to develop.
“Cheap energy is a huge piece of the puzzle. But the real accelerant for the Texas economy has been something even harder for Europe to copy. And it starts with one number. The average top personal income tax rate across 35 European countries is 38.5%. In Denmark, that number hits a staggering 60.5%. In Germany, France, and Italy, high earners face rates between 45 and 50%.”
“In Texas, the state income tax rate is zero. It’s always been zero, and the Texas Constitution actually makes it illegal for the state to introduce one without a direct vote from the public. This is not a temporary policy that a new government can reverse after the next election. It’s locked into the foundation of the state.”
“Texas still has high property taxes, so the total burden on a normal resident is not as dramatic as that 0% headline suggests. But for the people these economies are competing over, the math is brutal. A senior software engineer in Munich earns roughly €75,000 and takes home about $45,000 after income tax and social contributions. The same engineer in Austin earns $140,000 and takes home over $105,000. same skills, same screen, more than double the money in their pocket at the end of the year. Between 2020 and 2024, Texas startups pulled in over $46 billion in venture capital. In Q1 2025 alone, Texas tech companies raised nearly $3 billion, the biggest single quarter the state had seen in over 2 years, with massive deals in cyber security, defense tech, and biotech.”
“Compare that to the other side of the Atlantic. The entire European Union raised about $17.5 billion for AI funding in all of 2025. The US raised nearly $70 billion for generative AI alone by midyear. Not total tech, not all venture capital, generative AI alone. These two regions are not competing in the same category.”
“The EU’s regulatory framework was designed to protect consumers and level the playing field. GDPR [General Data Protection Regulation], the AI Act, 27 different national compliance regimes stacked on top of each other. The intentions were sound, but the unintended result is that the compliance costs favor massive American companies like Google and Microsoft, who can absorb them easily over smaller European rivals who can’t.”
“None of this means Texas has it all figured out, because the truth is it hasn’t. The most visible problem is housing. In 2019, a median Texas family earned 62% more than they needed to buy a median home. By 2023, that cushion had collapsed to just 7%. Not 62%, 7%. Over a third of Texas households now spend more than 30% of their income on housing.” That’s a national problem, partially engendered by the Flu Manchu shutdowns, partially by restrictive local building codes. “Affordable housing” blather snipped, since this is just more unnecessary government subsidy and intervention.
“The workers who actually build the Texas economy, the Tesla line workers, the nurses, the warehouse staff, can’t afford to live in the cities their labor is building anymore. They commute in from further and further out, and the roads, the housing, and the services all fall behind.” Partially true, partially false. Austin housing prices exploded, but have come down dramatically. Dallas and San Antonio prices spiked, then plateaued. Houston prices have continued climbing, but gradually.
“When a place grows this fast, the infrastructure simply can’t keep up.” True of Austin, less true of Houston, though having to rebuild certain interchanges is making things a nightmare for certain commuters.
Discussion of energy grid problems and the 2021 ice storm snipped, since I think we’ve covered those enough here.
“The problems in Texas are the problems of a place growing too fast. The problems in Europe are the problems of a place that is barely growing at all. In that sense, Texas and Europe have something in common. Both are stuck. The difference is that Texas is gridlocked because too many people are trying to get to work. Europe is gridlocked because too many committees are still deciding whether to build the road. Right now, Texas has its sights set on overtaking France, a G7 nation. At current growth rates, that gap closes faster than most people realize. And France’s response, like the rest of Europe’s, has been to wait and see.”
“The real question isn’t whether Europe can change. It’s whether it actually wants to change badly enough to feel the pain that comes with it. Because while Brussels is still writing the rule book, the game is already over for the economies Texas has already passed. The ones still in its path just haven’t checked the scoreboard yet.”
Left out of this coverage: Texas has a constitutionally mandated balanced budget, while the overwhelming majority of European nations keep running budget deficits to keep their cradle-to-grave welfare states afloat.
Not to mention a government run by Republicans rather than unstable coalitions including the Greens…
When last we checked, Microsoft CEO Satya Nadella was busy trying to shove Copilot, their AI tool, into every crevice of every Microsoft product. Finally, enough users seem to have complained loudly enough to get them to rethink the strategy.
Microsoft is vowing to focus on quality with future Windows 11 releases, which includes better performance and reeling in the company’s Copilot footprint over the OS.
“Quality” and “performance” go together with Windows like vanilla ice cream and used motor oil.
On Friday, Microsoft President for Windows and Devices, Pavan Davuluri, announced the “commitment to quality” in both a blog post and an email to users.
The plan calls for bolstering the “performance, reliability and well-crafted experiences” over the OS for this year.
That sort of suggests that “performance, reliability and well-crafted experiences” were not priorities for Windows in previous years, doesn’t it?
“These areas have meaningful impact on how you experience Windows: how fast it starts and responds, how stable it is under real workloads, and how consistent and thoughtful the experience feels,” Davuluri wrote.
PC users will be happy to know that one goal is reducing Windows 11’s resource usage to free up more capacity.
This would be great news if anyone trusted them to do that. Countless times in the past, Microsoft has pledged to reduce resource usage in Windows, but the bloat always returns.
Another priority is “less noise, less distraction and more control across the OS.”
Surprisingly, the blog post makes little mention of AI. Instead, Davuluri merely says the company wants “to be thoughtful about how and where we bring AI into Windows.”
“You will see us be more intentional about how and where Copilot integrates across Windows, focusing on experiences that are genuinely useful and well‑crafted. As part of this, we are reducing unnecessary Copilot entry points, starting with apps like Snipping Tool, Photos, Widgets and Notepad,” he said. Users can expect the change to roll out this month and in the next for Windows 11 preview releases.
Finally, the backslash against MicroSlop has gotten so loud that it’s even penetrated Redmond’s C-Ring. This is indeed progress, given that earlier this month Microsoft was literally banning users from its Discord for using the term. So give them credit for at least realizing that they have a problem.
Davuluri made the announcement months after he faced backlash for tweeting that “Windows is evolving into an agentic OS,” which caused some users to retort the company was obsessing about AI over basic Windows 11 performance. A January Windows 11 release that prevented PCs from booting up or going to sleep sparked more complaints about the OS’s stability.
There’s that vaunted Microsoft quality again.
Last month, Davuluri indicated he was taking the criticism seriously. Microsoft is also facing increased competition from Apple, which released its most affordable MacBook so far, the Neo, which has been a hit among consumers and reviewers.
Yeah, let’s talk about the state of Windows-based PCs and the competition. This was supposed to be a big year for PCs, but then reality hit.
There was supposed to be a massive tailwind for the PC market this year. Windows 10 reached end-of-life in late 2025, meaning that somewhere around 1 billion PCs worldwide stopped receiving security updates. This is less of an issue in the consumer PC market, but it’s a big deal in the business PC market.
PC OEMs like HP were set up for success in 2026, but the AI boom has complicated the situation. Enormous demand for DRAM and NAND chips from the AI infrastructure build-out, coupled with memory chip manufacturers shifting production to server products, has left the PC market with scraps. Memory chip prices have surged, pushing up the bill of materials and forcing price increases.
Gartner expects PC prices to surge by 17% this year, prompting consumers and businesses to hold onto their current PCs for longer. Budget PCs as a category could essentially disappear, leaving a large swath of consumers in a bind. Gartner expects PC shipments to tumble by 10.4% in 2026.
In one of life’s little ironies, the same AI bubble that Microsoft was trying to shove down user’s throats is what’s making buying a new PC less affordable and crushing sales. But it’s not stopping Apple.
HP is an obvious loser in this scenario, but there’s a surprising winner as well: Apple….
Apple is also exposed to rising memory chip prices across its entire device business. However, the company is seizing an opportunity to bring Windows PC users into its ecosystem.
Apple announced the MacBook Neo last week. The 13-inch entry-level MacBook starts at $599, a price that PC OEMs like HP will have trouble hitting as memory prices spiral higher. The Neo is powered by the A18 Pro chip, the same SoC used in the iPhone 16 Pro family. Higher-end MacBooks use Apple’s M-series chips, but the underlying technology is essentially the same.
Apple has equipped the Neo with 8GB of unified RAM, shared between the CPU and GPU, as well as a 256GB solid-state drive. These are really the bare minimums for a usable PC, but it’s enough for a solid entry-level experience. Early reviews have been positive, although full third-party reviews haven’t yet arrived ahead of launch.
Apple is playing the long game here. If there was ever a time to launch a budget MacBook, it’s right now. With memory prices driving up the cost of Windows PCs, a $599 MacBook should be appealing to budget-conscious consumers. Apple also offers the Neo to the education market for a discounted price of $499.
If Apple can grow its Mac install base and steal away market share from Windows-based PCs, the company will be setting up its Mac business for stronger growth down the road. Once the memory situation normalizes and the macroeconomic environment improves, Apple will have a larger base of Mac users eager to upgrade when the time comes.
This is probably correct. $599 is getting down around the price range for what used to be called netbooks, though that market segment seems to have largely died, at least among Windows users. (There still seems to be a market there for Linux users.)
Remember: Apple is the company that saw other companies dumping dump-trucks full of cash into AI build-outs and went “Nah, bro, I’m good.”
Google will spend approximately $90 billion on AI infrastructure this year. Meta has committed $65 billion. Microsoft, Amazon, and Alphabet are collectively spending over $300 billion. Apple, meanwhile, is spending just $12.7 billion on capital expenditure for the entire fiscal year.
As a hardware manufacturer, Apple can just wait for the AI wars to shake out and partner with the winner (if any).
And here’s a mildly amusing video on the whole situation.
I’m pretty sure Microsoft users wouldn’t object to Copilot if it was something optional you could turn off, and if the granularity of control allowed users to keep it entirely out of products they don’t want it in. But no, letting users decide isn’t the Microsoft Way, and they had to try shoving it into everything to justify the huge sums of money they were throwing into AI.
Megacorporations are telling businesses that their AI offerings are good enough to run vital company functions. The problem is, those AIs are still screwing up, and frequently in ways humans wouldn’t screw up. That’s what Amazon found out when they tried to eat their own dogfood, putting their AI in charge of Amazon Web Services. It didn’t go well.
Are AI tools reliable enough to be used at in commercial settings? If so, should they be given “autonomy” to make decisions? These are the questions being raised after at least two internet outages at Amazon’s cloud division were allegedly caused by blundering AI agents, according to new reporting from the Financial Times.
In one incident in December, engineers at Amazon Web Services allowed its in-house Kiro “agentic” coding tool to make changes that sparked a 13-hour disruption, according to four sources familiar with the matter. The AI, ill-fatedly, had decided to “delete and recreate the environment,” the sources said.
When something is “in the cloud,” that means it’s sitting on someone else’s computer. More specifically, it’s probably running as a containerized instance on any of a number of other CPU and storage pools being run under a hypervisor to scale up or scale down resources as demand requires. This allows efficient use of those resources, and it’s made AWS Amazon’s most profitable business. And most of the time AWS works pretty well.
Amazon employees claimed that this was not the first service disruption involving an AI tool.
“We’ve already seen at least two production outages [in the past few months],” one senior AWS employee told the FT. “The engineers let the AI [agent] resolve an issue without intervention. The outages were small but entirely foreseeable.”
AWS launched its in-house coding assistant, Kiro, in July. The company describes the tool as an “autonomous” agent that can help deliver projects “from concept to production.” Another AI coding assistant developed by Amazon, described as an AI assistant, was involved in the earlier outage.
The employees said the AI tools were treated as an extension of an operator and given operator-level permissions. In both of the outages, the engineers didn’t require a second person’s approval before finalizing the changes, going against typical protocol.
In a statement to the FT, Amazon claimed the outage was an “extremely limited event” that affected only one service in parts of China.
I’m not sure I was aware AWS operated in China, but I guess I’m not surprised. Is it too much to ask that the China data centers are adequately segmented and firewalled from the American data centers?
Moreover, it was a “coincidence that AI tools were involved” and that “the same issue could occur with any developer tool or manual action,” it said.
Except usually code changes are usually run through rigorous testing in a continuous integration/continuous deployment pipeline, and then deployed to a test server for performance and regression testing. It’s not clear that was done here.
It also claimed that its Kiro AI “requests authorisation before taking any action,” but that the engineer involved in the December outage had more permissions than usual, calling this a “user access control issue, not an AI autonomy issue.”
“In both instances, this was user error, not AI error,” Amazon insisted.
True, in the sense that an Amazon engineer evidently allowed an AI to alter production code.
The company also claimed that it had not seen evidence that mistakes were more common with AI tools. To which we retort: is Amazon living under a rock? While AI and its foray into commercial applications remain nascent, there’s no shortage of evidence showing that the tools are prone to malfunctioning. Their proclivity for producing hallucinations, or instances in which they fabricate facts, is well documented. So are their weak guardrails. Even some of Amazon’s own employees are reluctant to use AI tools because of the risk of error, they told the FT.
Veteran programmers are finding that AI coding assistants consistently spit out botched code, with several studies showing that the frequent double and triple-checking the questionable outputs require in reality slow down software engineers, even though the AI, on a surface level, may be producing the code faster. The rise of “vibe coding” with AI has resulted in numerous blunders in which an agentic AI makes decisions that its owners didn’t intend.
Of course, it would not be much of a ringing endorsement if tech companies weren’t using the AI tools they claim will supercharge productivity in their own operations, and they’ve been more than willing to get high on their own supplies. Both Microsoft and Google boast that over a quarter of their code is now written with AI. Engineers at Anthropic and OpenAI have suggested that nearly 100 percent of their code is AI written.
This does not inspire me with confidence. Let’s pull out the relevant XKCD comic again:
The only reason the modern technological world works is that someone, somewhere understands at a deep level how each of those boxes work, and can fix it if something goes wrong. And for Open Source software, the source code for those boxes is available somewhere other people can look at it and understand it.
When you start replacing the code in some of those boxes with AI-generated code, you start losing the knowledge of how everything works and why. Maybe the AI is producing clear, well-documented code, but you can’t count on it. And the AI doesn’t understand code the way a human does, because and AI doesn’t understand anything in the way we mean it, it’s running on artificially evolved heuristics that have performed well designing things to pass documented test cases, but which have zero frameworks for handling unanticipated exceptions. And when it breaks, there’s no guarantee a human will understand how and why it broke.
And given competitive time-to-market pressures, you can be sure companies will increasingly ship AI code without adequate safeguards or sufficient testing because their service is down hard and the latest code fixes the last AI bug, so they’ll end up rolling the fix straight to production, and something in the fix will be an even more disasterous bug none of the test cases caught and everything will come tumbling down.
And if you do that with enough of those little boxes of digital infrastructure, the entire underpinings of modern online life may come tumbling down with it. And you can’t find people to fix it because you laid them off last year and replaced them with AI.
The problem with eating your own dog food is that sometimes it can be lousy, especially if you have no idea what went into it…
A bunch of AI-related news has popped up this week, so let’s do a roundup.
Some AI companies are complaining that TSMC is killing the AI boom by not expanding rapidly enough:
Asianometry notes that TSMC’s caution at expanding is amply justified by the boom-and-bust nature of the semiconductor industry:
“I’m hearing many similar views in the Silicon Valley Borg that TSMC is the break or limiter on the AI boom, as if they’re the reason why we don’t have AGI yet. Because they didn’t and still don’t believe.”
“If we can ever say that a company that spent $41 billion on capital expenditure in 2025, with another $53 to $56 billion in 2026 planned, is sitting on its hands, doing nothing.”
“TSMC having 90% share of the AI chip market looks pretty unhealthy. That should go down and it will. Samsung seems to be doing well so far.”
“The cold, hard reality is that shortages are a fact of life in semiconductors, as are horrific gluts.”
“What we are flippantly labeling as TSMC we really mean is the AI supply chain. And that supply chain is as complicated as you can possibly imagine. Like an iceberg, it looks big enough on the surface of the water, but goes way far deeper underneath. TSMC has thousands of suppliers in two categories: Equipment like the famed ASML lithography tools and materials like photoresist, silicon wafers, acid etch gases and so on. These are not generalized tools and materials. They are not fungeible like AWS compute units.”
“And then there are the memory guys. You cannot ship an AI system without memory. DRAM and NAND. Nvidia’s AI chips use a special form of DRAM called high bandwidth memory, and they use quite a lot of it. The memory industry is just as consolidated as the logic industry, with the major players being Samsung, SK Hynix and Micron.”
“The chip guys are last to know when the party is getting started, but first they get batoned in the face when the police shut things down.”
He points out that semiconductor manufacturers have log supply chains. He uses a different metaphor (the beer distribution game, or a bullwhip), but back when I was working at Applied Materials, it was described as trains linked together with slinkys. First software takes off, then hardware gets yanked along, then the chip manufacturers get yanked, and then, finally, semiconductor equipment manufacturers get yanked into motion, and shortly after that happens, the bust hits the front of the train, and the trailing cars all crash into each other. It’s a regular boom/bust cycle.
“From 1961 to 2006, electronics consumption in the United States grew positively but with wild volatility swings between 0 to 20%. But for the semiconductor makers, that translates to swings anywhere from 20% to 40%. And for the equipment makers, it is amplified even more, plus or minus 60%. The whip hits particularly hard in the semiconductor industry because of the industry’s long lead times. It takes 4.5 months to fabricate and package a chip. It takes 18 months to 2 years to build a fab. Meaning from shovels down to producing chips, and it takes 12 to 18 months to produce and install something like an EUV machine into the fab. Another 6 months before that machine actually starts patterning wafers.”
“Long lead times mean having to make very long demand forecasts, which leads to extreme volatility swings during up and downturns even if those up or downturns are relatively small.” People forget that in 1998, during the time we now think of as the DotCom Boom, there was a small semiconductor downturn that had Applied Materials forcing employees to take unpaid leave.
“ASML just reported 2025 earnings, and we see the bullwhip in full effect. TSMC raised capital expenditure 35% but ASML announced €13.2 billion of net new bookings. Analysts had expected just €6.32 billion. This is because ASML collected orders not just from TSMC, but also Samsung, Intel and the memory guys. When it rains it pours, right? Again, this is why I fear that another AI foundry would not mean our compute shortage is solved, because ultimately, when those foundries start scaling their capacity, they all go to the same suppliers.”
He goes over how car manufacturers cancelled orders during Flu Manchu, and then scrambled when the economy took off afterwards. “TSMC was trying to discern between double booked orders and real demand, which is not an uncommon experience for them. Customers lie about their own demand all the time, or at least we can say that they are eternally optimistic. TSMC tried to respond in 2022. The Taiwanese giant poured $36 billion into capital expenditure. They went to their suppliers and pushed like no tomorrow.”
“It turned out those customers really were double booking orders and artificially inflating demand. When the macro environment turned in 2022, the automotive, smartphone, and PC chips that were so hot during the COVID era fell out of vogue and customers started cutting orders.”
“Meanwhile, deeper down in the supply chain, TSMC and the rest of the semiconductor industry were getting bullwhipped by COVID hangover. Utilization at TSMC’s multi-billion dollar N7 fabs crashed, Semi analysis wrote in April 2023. Now, Semi analysis data indicates that the 7nm utilization rates were below 70% in Q1. Furthermore, Q2 gets even worse with 7nm utilization rates falling to below 60%. This is primarily due to weakness in both smartphones and PCs, but there is a broader weakness in most segments. A fab’s break even utilization rates are about 60% to 70%. So those N7 Taichung fabs were taking financial losses potentially on the order of hundreds of millions, maybe even billions. The financial burdens of low utilization are another reason why I’m skeptical another AI foundry could have rushed into the AI chip fray to save the day.”
He says that Intel incurred losses during this period due to an unnecessary fab expansion, which is probably true, but that was a secondary factor next to their longer running problem of getting their process wrong.
“ChatGPT was released in November 2022, and that kicked off a massive increase in capex amongst the hyperscalers in particular, but it sure seems like TSMC didn’t buy the hype. That lack of increased investment earlier this decade is why there is a shortage today and is why TSMC has been a de facto break on the AI buildout/bubble.”
“I recall news in mid 2024 of TSMC struggling with CoWoS capacity bottlenecks and yield problems, including one design issue that caused cracks in the Nvidia chips packaging.” CoWoS is Chip on Wafer on Substrate, which involves fabbing an interposer as a substrate for faster connections between your processing chips and memory.
“I also recall news in late 2024 noting how the vendors in charge of making the server racks for Nvidia’s Blackwell servers struggled with overheating, liquid cooling leaks, software bugs, and connectivity issues. Such technical difficulties delayed server deployment until early to mid 2025, creating a weird situation for several months where TSMC was pumping out chips that just went into storage. So that gated things, because you don’t scale until you first fix the technical problems.”
Then there’s the power-scaling issue, which is a whole ‘nuther can of worms.
There’s a lot of talk about a SaaSpocalypse going on thanks to a new AI tool. (SaaS is “Software as a Service.” Instead of hosting your own payroll or sales-tracking or whatever servers, you hire a company that already has cloud software setup to do it and you just tie into that, which can considerably reduce startup costs. A whole lot of successful new tech companies over the last decade plus have been SaaS companies.)
The software sector was jolted overnight with what analysts are calling a “SaaSpocalypse” — a sudden and severe selloff triggered by new artificial intelligence tools unveiled by US AI startup Anthropic. The episode has sharpened investor fears that AI is no longer merely helping software companies but may now begin replacing them.
Anthropic has expanded its enterprise AI platform, Claude Cowork, by launching 11 new plugins aimed at automating a wide range of professional tasks. Claude Cowork is an agentic, no-code AI assistant built for corporate users, allowing companies to automate workflows without writing software. The new plugins are designed to handle tasks across legal, sales, marketing and data analysis functions. The most recent addition is Anthropic’s Claude Legal agent, which can perform routine legal work such as document and contract review, and compliance checks.
Anthropic has said that the tool does not provide legal advice and that all AI-generated outputs must be reviewed by licensed attorneys. Even so, the breadth of automation signals a step change in how much white-collar work AI systems can now perform.
Productivity — Manage tasks, calendars, daily workflows, and personal context
Enterprise search — Find information across your company’s tools and docs
Plugin Create/Customize — Create and customize new plugins from scratch
Sales — Research prospects, prep deals, and follow your sales process
Finance — Analyze financials, build models, and track key metrics
Data — Query, visualize, and interpret datasets
Legal — Review documents, flag risks, and track compliance
Marketing — Draft content, plan campaigns, and manage launches
Customer support — Triage issues, draft responses, and surface solutions
Product management — Write specs, prioritize roadmaps, and track progress
Biology research — Search literature, analyze results, and plan experiments
A lot of those are already automated elsewhere, but I suspect a lot accountants and paralegals just felt a goose strut across their grave. On the other hand, who is really going to turn over, say, Accounts Payable to an AI? One glitch, and your entire bank account is drained…
If it works (a big if, give so many AIs are prone to hallucinations), this is potentially good news for Anthropic and the companies using their tools, and bad for SaaS companies and the employees currently doing those jobs.
I note there’s no plugin for technical writing…yet.
And Google Cloud ended 2025 at an annual run rate of over $70 billion, representing a wide breadth of customers, driven by demand for AI products.
We’re seeing our AI investments and infrastructure drive revenue and growth across the board. To meet customer demand and capitalize on the growing opportunities we have ahead of us, our 2026 CapEx investments are anticipated to be in the range of $175 to $185 billion.”
Remember how Nvidia was going to invest $100 billion in OpenAI? Yeah, not so much.
In September 2025, Nvidia and OpenAI announced a letter of intent for Nvidia to invest up to $100 billion in OpenAI’s AI infrastructure. At the time, the companies said they expected to finalize details “in the coming weeks.” Five months later, no deal has closed, Nvidia’s CEO now says the $100 billion figure was “never a commitment,” and Reuters reports that OpenAI has been quietly seeking alternatives to Nvidia chips since last year.
Reuters also wrote that OpenAI is unsatisfied with the speed of some Nvidia chips for inference tasks, citing eight sources familiar with the matter. Inference is the process by which a trained AI model generates responses to user queries. According to the report, the issue became apparent in OpenAI’s Codex, an AI code-generation tool. OpenAI staff reportedly attributed some of Codex’s performance limitations to Nvidia’s GPU-based hardware.
After the Reuters story published and Nvidia’s stock price took a dive, Nvidia and OpenAI have tried to smooth things over publicly. OpenAI CEO Sam Altman posted on X: “We love working with NVIDIA and they make the best AI chips in the world. We hope to be a gigantic customer for a very long time. I don’t get where all this insanity is coming from.”
Microsoft’s Copilot chatbot has become central to its artificial-intelligence strategy as the company’s close partnership with OpenAI diminishes. But the effort to build it up as a ChatGPT alternative has been tough going.
Confusing brand positioning and interoperability problems have frustrated users, current and former employees who have worked on Microsoft’s AI products said.
Interoperability problems? With a Microsoft product?
Only a small proportion of subscribers to Microsoft’s enterprise suite use Copilot, and the percentage who favor it over Google’s Gemini or other tools has decreased in recent months, according to data reviewed by the Journal.
The stakes are high for Microsoft because Copilot is core to a push by Chief Executive Satya Nadella to transform Microsoft into an AI-first company, much as he transformed it into a cloud-first company around a decade ago. Copilot is one of Nadella’s top priorities, current and former executives said.
Microsoft shares tumbled after its earnings report last week sparked investor concern that growth in its most important unit, the Azure cloud-computing business, is slowing, and that its AI business is reliant on OpenAI while Copilot remains unproven. Shares fell nearly 3% Tuesday amid a slide in software stocks prompted by fresh concerns that AI tools will make enterprise subscriptions less necessary.
For other AI companies, we merely suspect they’re evil. For Microsoft (and Google), we already know they’re evil…
Greetings, and welcome to a rare Saturday LinkSwarm! This week: The Supreme Court stays the injunction against the Texas redistricting map, a bunch of Twitter fakes exposed, Trump drops the boom on Somali illegal alien scumbags,
U.S. Supreme Court Justice Samuel Alito issued an administrative stay of Tuesday’s ruling by an El Paso panel of federal judges that rendered the new congressional map passed by Texas Republicans this summer unusable for the 2026 midterm election.
The order restored the new map, pending consideration of the appeal by the State of Texas, and directed the Democratic-aligned parties to submit their response by Monday.
Snip.
The ruling drew a particularly pointed dissent from Judge Jerry Smith, the lone dissenter on the panel, who asserted that the motivation behind the redraw was clearly partisan gain — a position that sits outside the jurisdiction of the court.
Following that ruling, Attorney General Ken Paxton appealed the ruling to the U.S. Supreme Court on Friday, asking for an administrative stay — which Alito granted.
“Compounding the harm, the district court entered its sweeping injunction far too late in the day — ten days after Texas’s candidate filing period had already opened. The injunction changes the boundaries of all but one of the State’s 38 congressional districts, enjoining Texas from using its duly enacted 2025 map and resurrecting the repealed 2021 map,” Texas wrote in its appeal.
“The chaos caused by such an injunction is obvious: campaigning had already begun, candidates had already gathered signatures and filed applications to appear on the ballot under the 2025 map, and early voting for the March 3, 2026, primary was only 91 days away. The lateness of the district court’s injunction (issued 38 days after the hearing) alone warrants a stay.”
As things stand, Texas Republicans’ map is back in effect while the U.S. Supreme Court considers the case in expedited fashion.
Texas’ candidate filing deadline is December 8, 2025.
Twitter/X turns on locations and it turns out a lot of “American” account pushing that “GOP civil war”` nonsense were foreign psyops.
There are thousands of accounts like this. Many of them explicitly claim to be American or Western, but are run by random people in Asia and Africa to sow chaos and get clicks.
BREAKING – Waves of Democrat influencers are being exposed as foreigners under X’s new location update, including leftist X agitator Alex Cole, who claimed he voted for Kamala but has now been revealed to be Canadian. pic.twitter.com/3LrAsYCiMw
Temporary Protected Status (TPS) is supposed to be used in extreme cases of humanitarian need for short terms (usually for 6, 12, or 18 months), allowing foreign refugees a safe haven in America.
As deportation efforts have ramped up, however, the American public has learned that some foreigners have remained in the country on TPS for decades. Some politicians and businesses have purposely imported large numbers of foreigners into small American towns, such as Haitians in Ohio and Pennsylvania, as cheap labor to replace Americans.
President Donald Trump’s initiative to eliminate government waste and fraud through a Department of Government Efficiency (DOGE) has quietly disbanded with a full 8 months still left on its charter.
Earlier this month when Reuters asked Office of Personnel Management Director Scott Kupor about the status of DOGE, Kupor replied, “That doesn’t exist.”
Representative Tim Burchett (R-TN) said that Elon Musk, who headed up the DOGE effort, was pushed out Washington D.C. because he was getting too close to exposing corrupt officials who are enriching themselves through dark money non-governmental organizations (NGOs).
Burchett told Benny Johnson, “NGO money pours into Washington and ends up in politicians’ pockets as dark money.”
DOGE had made dramatic impact on the federal government during the early months of Trump’s second term, shrinking the size of federal agencies and cutting their budgets or revealing astonishing amounts of questionable money flowing through NGO coffers.
Sound like a good reason to continue the work, not abandon it…
All that “don’t obey illegal orders” nonsense Democrats are regurgitating? Yeah, it’s Soros-funded, “Sponsored by Win Without War, a progressive advocacy group,” which in turn is funded by Soros’ Open Society Foundations.
Marjorie Taylor Greene resigned from congress. As in the NFL, there’s always someone that has to “set the edge,” and MTG was the person who did that in the Trump era.
What the hell? Is China committing war crimes in Philippines coastal waters?
The Philippine Navy recently caught Chinese Fishing Militia putting Cyanide in the water near the BRP Sierra Madre at the Second Thomas Shoal.
The incident that was recorded on camera demonstrates the brutality & destruction meted out by the Chinese Fishing Militia inside… pic.twitter.com/L7NCI0UIik
The apparent reason Armata failed is this: sanctions.
But there’s more to the story, too. In fact, several interlocking factors account for the T-14’s failure to materialize as intended.
Let’s first look at costs and priorities: the unit cost of the T-14 was estimated at several million dollars – far higher than Russia had budgeted for.
The increase in cost meant that it couldn’t actually be sustained at scale. And, faced with heavy losses in Ukraine and urgent demands to ramp up numbers, Moscow opted to modernize its legacy platforms, such as the T-90, rather than invest in an expensive and unproven system. A tough choice, but a logical one.
The domestic production line for the T-14 never actually achieved accurate serial output, in large part thanks to sanctions and industrial bottlenecks.
There was no assembly line. Yes, really: every vehicle was hand-built like a luxury car. Sanctions and supply-chain constraints further hindered the manufacture of key components and high-end electronics required for the platform.
But even if Russia had been able to assemble more of the tanks before the sanctions really kicked in, it might not have changed the reality on the battlefield. Even when the war in Ukraine created a burning need for armored vehicles, Russia hesitated to commit T-14 units to the frontline for one worrying reason: they were vulnerable.
With the rise of automated systems, drone warfare, and long-range combat, those tanks may have proven as vulnerable as older units – and losing tanks built pre-sanctions would mean replacing them with older tanks.
That wouldn’t have made sense.
For more than a decade, the T-14 Armata has embodied Russia’s ambition to leap ahead of the West in tank design and warfare.
But it failed.
The usual lefty sorts are trying to raise Maryland’s minimum wage to $25. Virginia’s minimum wage will be $12.77 in 2026. Which state will businesses choose?
Brown County Judge Shane Britton was suspended from office without pay on Tuesday, one day after he was arrested on multiple charges that included allegations he assaulted a female prosecutor and interfered with the prosecution of a family violence case.
According to indictments handed down by a grand jury last week, Britton has been charged with three felonies: tampering with a witness in a family violence case, assault of a public servant, and tampering with a government document.
To understand the difference, it helps to look at what each chip was originally built to do. A GPU is a “general-purpose” parallel processor, while a TPU is a “domain-specific” architecture.
The GPUs were designed for graphics. They excel at parallel processing (doing many things at once), which is great for AI. However, because they are designed to handle everything from video game textures to scientific simulations, they carry “architectural baggage.” They spend significant energy and chip area on complex tasks like caching, branch prediction, and managing independent threads.
A TPU, on the other hand, strips away all that baggage. It has no hardware for rasterization or texture mapping. Instead, it uses a unique architecture called a Systolic Array.
The “Systolic Array” is the key differentiator. In a standard CPU or GPU, the chip moves data back and forth between the memory and the computing units for every calculation. This constant shuffling creates a bottleneck (the Von Neumann bottleneck).
In a TPU’s systolic array, data flows through the chip like blood through a heart (hence “systolic”).
It loads data (weights) once.
It passes inputs through a massive grid of multipliers.
The data is passed directly to the next unit in the array without writing back to memory.
What this means, in essence, is that a TPU, because of its systolic array, drastically reduces the number of memory reads and writes required from HBM. As a result, the TPU can spend its cycles computing rather than waiting for data.
Google’s new TPU design, also called Ironwood also addressed some of the key areas where a TPU was lacking:
They enhanced the SparseCore for efficiently handling large embeddings (good for recommendation systems and LLMs)
It increased HBM capacity and bandwidth (up to 192 GB per chip). For a better understanding, Nvidia’s Blackwell B200 has 192GB per chip, while Blackwell Ultra, also known as the B300, has 288 GB per chip.
Improved the Inter-Chip Interconnect (ICI) for linking thousands of chips into massive clusters, also called TPU Pods (needed for AI training as well as some time test compute inference workloads). When it comes to ICI, it is important to note that it is very performant with a Peak Bandwidth of 1.2 TB/s vs Blackwell NVLink 5 at 1.8 TB/s. But Google’s ICI, together with its specialized compiler and software stack, still delivers superior performance on some specific AI tasks.
The key thing to understand is that because the TPU doesn’t need to decode complex instructions or constantly access memory, it can deliver significantly higher Operations Per Joule.
“TPU v6 is 60-65% more efficient than GPUs.”
Austin’s APL bookstore Recycled Reads will be closing in January and the stock distributed to individual library sales shelves. I doubt I’ll be visiting various library branches to book scout. Maybe they should go back to the book sale events they used to hold.
Given the title of this post, a lot of people will naturally assume that Google is using AI to spy on them as a matter of course, since Google uses every other tool to spy on us. Indeed, since Google first announced AI initiatives, I’m pretty sure most people never assumed Google wouldn’t use it to spy on us. Nevertheless, there’s now a lawsuit over it.
Google is facing a lawsuit over its Gemini assistant, which allegedly collected data from Gmail, Chat, and Meet users without their consent.
Any rational person who uses Gmail knows Google is going to gather data on you from it. It’s part of the terms and conditions of the Faustian bargain to use free services.
The complaint accuses the tech giant of violating privacy laws by activating the tool across its platforms without informing users.
Yeah, I’m pretty sure that wasn’t part of the terms and conditions when I signed up for it two decades ago.
The plaintiffs claim that this covert data collection allowed Google to access sensitive communications and personal details shared through emails, messages and video calls.
The lawsuit alleges that Google’s parent company, Alphabet, activated Gemini across Gmail, Chat and Meet in October without user consent.
Previously, users could opt-in to use the assistant. However, the plaintiffs claim that Google silently enabled it for all users.
This gave the tool access to sensitive communications and personal details shared through emails, messages and video calls.
The name of the lawsuit is Thele v. Google, LLC. I checked and, sure enough, that stuff was enabled without my permission. Being a Gmail user that never gave Google permission to train their AI on me, I should probably see if I can climb aboard the Litigation Express. I’ll send them an email.
Click on Settings (the cog icon in the top-right bar).
Press See all settings.
In the General tab, scroll down to Google Workplace smart features and click on the button.
Turn off smart features in Google Workspace and click Save. This will block Gemini AI from Gmail, Chat, Meet and Drive. You can remove Gemini from Google Maps, Wallet, Google Assistant and the Gemini app, too.
Thele v. Google is not the only lawsuit involving Gemini brewing. Clownfish TV brings news of a suit over Gemini telling a student to call 911 over having their phone time restricted.
They also touch on instances (that I think we’ve mentioned here before) of Gemini allegedly telling children to kill their parents.
But that’s not all on the Google privacy abuse front! According to Louis Rossmann, even after being disabled, Nest thermostats upload 50 megabytes of data to Google every day:
That amount seems…excessive. Especially for a product you paid for. As Rossmann pointed out, letting old devices continue to connect to the Internet is a large security risk. Plus the usual problems with the hoary old Digital Millennial Copyright Act.
Just as in deals with the Devil stories, your damnation in dealing with Google frequently dwells in the fine print of the contract you agree to in order to use their products for free.
The problem is, Google always seems to be unilaterally changing the fine print without telling you. And I’m pretty sure those changes are never in your favor.
Happy Anti-Communism Week everyone! (In addition, of course, to May 1st being one of two Victims of Communism Day.) The #SchumerShutdown ends with a whimper, a whole lot of SNAP fraud has been uncovered, more Democrats committing fraud, Chip Roy wants a complete immigration halt, Ukraine hits a bunch more Russian oil refineries, some semiconductor shenanigans, another company leaves Delaware for Texas, some tech companies in trouble, an interesting new pistol design, and a novel theory on “AI-related layoffs.”
It’s the Friday LinkSwarm!
As a side note, the mosquitos have been brutal the last few days. Possibly because it’s been a very warm (though largely dry) November, and the bats have already migrated south.
President Donald Trump on Wednesday night signed a continuing resolution at the White House that ends the record-breaking 42-day federal government shutdown.
The Senate passed the resolution on Monday and the House passed it earlier Wednesday evening. The resolution will keep the entire government funded through Jan. 30, and extends funding for military construction, Veterans Affairs, the Department of Agriculture, and Congress beyond that, through Sept. 30.
Trump slammed Democrats for causing the shutdown by refusing to go along with a clean continuing resolution for over a month, and urged voters to remember the party responsible for causing the six-week-long chaos during next year’s midterms.
“Republicans never wanted a shutdown and voted 15 times for a clean continuation of funding,” Trump said. “The Democrats shutdown has inflicted massive harm … So I just want to tell the American people, you should not forget this when we come up to midterms and other things. Don’t forget what they’ve done to our country.”
The resolution gives backpay to many federal workers and reinstates employees who were fired during the shutdown, but does not include an extension of Affordable Care Act subsidies despite it having been a key Democratic demand in the shutdown. The subsidies are set to expire at the end of the year.
And what did Chuck Schumer get for shutting down large portions of the federal government for more than a month? Two things: “Jack” and “Squat.”
I hear that if you call Senate Minority Leader Chuck Schumer’s office, the hold music is Cheap Trick’s “Surrender.”
Last Tuesday night, Democrats were jubilant, convinced they had just inflicted the first of many consequential defeats upon their detested foes, President Trump and the Republican Party. And now here we are, six days later, and Democrats are once again disappointed, infuriated, and at each other’s throats.
For the past 41 days, Republicans have had 53 senators willing to reopen the government, joined by Catherine Cortez Masto of Nevada, John Fetterman of Pennsylvania, and “independent” Angus King of Maine, who caucuses with the Democrats. But it requires 60 votes to cut off debate and bring the legislation to the floor for a vote, and thus to reopen the government, Republicans needed at least four more Democrats to change their mind.
Last night, five additional Democratic senators agreed to vote to reopen the government — and in the eyes of their fellow Democrats, effectively surrendered. Tim Kaine of Virginia, Dick Durbin of Illinois, Maggie Hassan of New Hampshire, Jacky Rosen of Nevada, and Jeanne Shaheen of New Hampshire shifted their positions.
Those eight agreed to reopen the federal government at current funding levels through January 30, and in exchange, all they needed was a pledge from Senate Majority Leader John Thune of South Dakota to hold a vote on legislation to extend the Obamacare exchange premium subsidies by the second week of December.
There are one or two other deal-sweeteners in there for Kaine, notably an attempt to reverse more than 4,000 federal layoffs the Trump administration announced in the shutdown, and language to prevent future layoffs through January 30.
Snip.
Republicans just got the government reopened in exchange for a promise of a vote — not even promise of passage! — and rehiring government workers who were on the job on September 30. That’s a very small price to pay, and Republicans didn’t have to get rid of the filibuster, the ultimate short-term gain, long-term loss for Republicans in the Senate.
Across three-fifths of the United States, the Trump administration has found half a million people receiving SNAP benefits twice over and 5,000 dead people receiving them. In deep blue states, the fraud is probably much worse.
It is important to clarify that 20+ states out of the 50 did not comply with the federal government’s request for information on SNAP beneficiaries, likely because they are trying to hide how many illegal aliens are illicitly receiving food stamps. So the horrifying numbers revealed by U.S. Secretary of Agriculture Brooke Rollins on Laura Ingraham’s Fox News show, The Ingraham Angle, are actually incomplete, and will probably be much higher if the administration can make radical Democrat states provide the necessary data.
Snip.
The secretary continued to list off food stamp recipient statistics: “80% [are] able-bodied Americans, meaning they can work, they don’t have small children at home, they’re not taking care of an elderly parent. They can work, and they choose not to work, of course, because they’re getting significant benefits from the taxpayer.”
We need to restore shame to able-bodied adults living on the public dole.
(Hat tip: Stephen Green at Instapundit.)
A Texas congressman is proposing a “freeze” on all immigration until the federal government fixes the country’s broken system.
U.S. Rep. Chip Roy (R–TX) said Wednesday he is introducing a bill called the “Pause Act” that will freeze all immigration until Congress achieves certain objectives, including reforming chain migration and birthright citizenship and ending H-1B visas.
He said the nation’s record-high foreign-born population is creating “a cultural problem about who we are as Americans.”
Roy, who is in a four-way race to be the Republican nominee for Texas attorney general in 2026, explained his proposal on The Benny Show.
In addition to the immigration freeze and related reforms, Roy called for revisiting Plyler v. Doe, a case originating in Texas that resulted in a 1982 U.S. Supreme Court decision requiring states to fund the education of illegal alien children.
Roy also said his bill would require vetting people for their adherence to Sharia law.
“Why are we importing any human being that is adherent to Sharia law, which is totally contrary to the Constitution, and our values, and Western civilization?” Roy asked host Benny Johnson.
“In Texas, we’ve been dealing with the brunt of the illegal immigration influence. But now we’re seeing, I think, the ramifications of the H-1B system and how it has been abused, in addition to chain migration and diversity visas, which we’ve been trying to fix for a long time, and we’ve been unable to do so,” said Roy.
Mostly agree with this, though there would probably have to be a way for individual exceptions to be made (say, a foreign Christian under a death threat from jihadists, or a Russian or Chinese defector, or a foreign NBA draft choice). But it should be so narrow as to require the personal approval of DHS Director Kristi Noem…
There are Somalis in Minnesota who wouldn’t vote for far leftist Somali Omar Fateh because he was from a different Somali clan, and they want members of the rival clan kicked out of the country…
They also hit multiple targets in Novorossiysk, including both the oil terminal and the S-300/400 system defending it. Also, there’s no way I can donate €100 right now, but I really want one of those “This Is Fine” patches…
Orchestrating Over 180 Anti-Trump Lawsuits Through CREW: As co-founder of Citizens for Responsibility and Ethics in Washington (CREW), Eisen led hundreds of ethics complaints and lawsuits against the Trump administration, often perceived as partisan harassment that politicizes oversight and strains constitutional separation of powers.
Snip.
Involvement in USAID Funding Scandal: Accused of ties to $17M misappropriation via family-linked NGO, raising corruption concerns in foreign aid.
Plenty more at the link.
(Heavy sigh) Look, I’ve been avoid the whole stupid Tucker Carlson thing because he hasn’t been a particularly important part of the mediascape for a while, and plenty of other people were already dog-piling him. Yet, this week he seemed to turn up some pretty interesting information on would-be Trump assassin Thomas Crooks. Namely that he was a pro-Trump supporter…until he radically changed his tune in early 2020.
On July 19, 2019 Crooks writes: “Ilhan Omar and others are invaders and should honestly be killed and their dead bodies sent back.”
On July 20, 2018, Crooks writes: “If youre saying trump is a bad president you arent a patriot as trump is the literal definition of Patriotism”
Seven hours after that comment, Crooks writes: “I hope a quick painful death to all the deplorable immigrants and anti-trump congresswoman who dont deserve anything this countru [sic] has given them”
Later that evening he wrote: “Everyone of the Trump hat-ing democrats deserve to have their heads chopped of and put on steaks for the world to see what happens when you fuck with America”
These types of comments continued for months, “and became increasingly violent.”
“If any of the democratic candidates win. They wont be in there for long. Because unlike the dems we have guns and lots of them”
He also quoted Mao – writing “The only real political power comes from the barrel of a gun.”
The Change:
In early 2020 as the pandemic shifted into the headlines, crooks “radically” changed – writing of “trumps stupidity.”
He then began to mock the idea of the deep state – writing that “The deep state is simply made up of anybody who dis-agrees with the right wing. Conversation over.”
In Feb. 2020, Crooks called out Trump supporters as “brainwashed,” and a “cult.”
Later that day, Crooks called Trump a racist.
And in April 2020 when the COVID panic was in full swing, Crooks became pro-lockdown, writing “It seems that you people don’t understand that sometimes Public safety comes before your Personnel rights.”
He then wrote: “…going to a chinese new years party in america isn’t putting you at risk for corona virus because believe it or not viruses don’t spread through race like Tucker Carlson probably told you.”
In May of 2020, Crooks called Republican concerns over voter fraud “ignorant.”
He then wrote a comment that sounded like a “digital manifesto,” Carlson reports.
“they only way to fight the gov is with terror-ism style attacks, sneak a bomb into an essential building a set it off before anyone sees you, track down any important people/politicians/military leaders etc and try to asasinate them. Any sort of head fight is suicide and even ambush/surprise attacks likely aren’t going to end well.”
Sounds like another “known wolf,” doesn’t it? And the assertion that “there’s no deep state” (combined with what else we know about the assassination) makes you go “Hmmm.”
Senator Mike Lee (R-UT) is pushing back on the idea that the Affordable Care Act (ACA), known as Obamacare, has made health insurance costs more affordable, saying, “Obamacare makes everyone else poor.”
Lee shared a graphic, first posted by President Trump on Truth social, showing how major health insurance company stocks have performed since the ACA was enacted in 2010 to November 2025.
The seven major health insurance companies depicted on the graph show gains of anywhere from 414% to 1177% in their stock prices between March 2010 and November 2025.
Health insurance companies are making money hand over fist—not because they’ve discovered new & innovative ways of making Americans healthier, but because Obamacare insulates them from competition while giving them massive subsidies
Lee called out the insurance providers, noting that they’re “making money hand over fist” but not because they are providing “new & innovative ways of making Americans healthier.”
Instead, Lee says, these health insurance companies are prospering due to the bureaucratic barriers that prevent new competition and from massive subsidies from the federal government.
The Saudis are getting ready to purchase 48 F-35s.
California Governor Gavin Newsom’s former chief of staff Dana Williamson was arrested Wednesday in an FBI corruption probe and charged with multiple counts of bank and wire fraud.
Federal authorities accused Williamson, 53, of participating in a scheme to funnel campaign money from former federal Secretary of Health and Human Services Xavier Becerra into a personal account. Sean McCluskie, Becerra’s former chief of staff, was named as a co-conspirator.
“This is a crucial step in an ongoing political corruption investigation that began more than three years ago,” U.S. Attorney Eric Grant said in a statement. “As it always has, the U.S. Attorney’s Office will continue to work tirelessly with our law enforcement partners to protect the people of California from political corruption.”
Williamson and McCluskie stole $225,000 between February 2022 and September 2024 from Becerra’s dormant state campaign fund, the federal indictment says. The Department of Justice investigation into the matter began three years ago, under former President Joe Biden’s administration, FBI Sacramento Special Agent in Charge Sid Patel said.
“The news today of formal accusations of impropriety by a long-serving trusted advisor are a gut punch,” Becerra told local outlet KCRA 3.
Williamson was hit with 23 charges, including conspiracy to commit fraud, conspiracy to defraud the United States and obstruct justice, subscribing to false tax returns, and making false statements, the U.S. Attorney’s Office said.
Democratic political consultants are so money-hungry they’ll rake graft off other Democrats. Big fleas have little fleas…
Man, it sure seems like a lot of prominent Democratic politicians are committing mortgage fraud. ‘Rep. Eric Swalwell (D-Calif.) was hit with a federal criminal referral for alleged mortgage and tax fraud related to his purchase of a $1.2 million home in Washington, DC, that he claimed as a primary residence.” As Dwight notes: “You may remember Eric Swalwell for such hits as ‘banging a Chinese spy‘” and “threatening to use nuclear weapons against gun owners.”
So a Chinese fraudster connected to Communist intelligence services wandered in from Canada and bought a trailer park next door to a stealth bomber base in Missouri.
This is not the opening line of a surreal joke.
Whiteman Air Force Base is home to our tiny fleet of B-2 bombers, and yet an RV park just a mile away “is one of several properties near U.S. military interests acquired by a web of shell companies, which are ultimately owned by a couple who live in Canada and belong to organizations controlled by disgraced Chinese tycoon and self-described former CCP intelligence ‘affiliate,’ Miles Guo,” according to a bombshell Daily Caller report.
Someone in the federal government needs to get this fixed. Get a warrant to toss the entire trailer park to see what spectrum warfare equipment they might be using, then seize the place under eminent domain for national security reasons.
BREAKING: The Attorney General of Kansas just charged Mayor Jose Ceballos of the City of Coldwater for illegally voting as a noncitizen in several elections.
Not only did he get elected city councilman & mayor as a noncitizen, he also voted. WOW! The six charges come immediately… pic.twitter.com/amhsJJvaZW
‘We now have tools, thanks to the current White House, that we haven’t had in over 10 years,’ said Kansas Secretary of State Scott Schwab, ‘that we can check through the SAVE program, to find out if folks end up on our voter rolls. And they could be a legal resident, but they’re not a citizen. We want to make sure that gets clarified.’
Deport him.
Least you think I’m never critical of President Trump, I want to note that his trial balloon for 50 year mortgages is a really bad idea. It’s not a way to build wealth, and the only party getting rich off that deal is the banks. Financially, you’d be better off living in a van for a few years until you can afford a real mortgage.
This certainly has a whiff of scandal: “Houston ISD Sues Texas Attorney General to Block Release of Emails with California PR Firm. The district wants to keep communications with a PR firm from becoming public.”
Houston Independent School District (ISD) filed a lawsuit against Texas Attorney General Ken Paxton to block the release of emails between the district and Los Angeles public relations firm Bryson Gillette.
Bryson Gillette is former Obama aide Bill Burton’s public relations firm run by Democratic operatives. White House Press Secretary Jen Psaki was a senior adviser there.
Bryson Gillette was involved with the district’s rebranding in May. Houston ISD’s Chief of Public Affairs and Communications Alex Elizondo told an advisory committee that the district had a brand identity that “isn’t inviting or super compelling.”
A Houston ISD spokesperson said the rebrand came at no additional cost to the district and coincided with the rollout of new district and campus website designs scheduled for August.
According to the suit, ABC13 News requested one month of emails between Houston ISD and Bryson Gillette on May 8, which the district received on May 9. On May 21, the district asked Paxton to withhold documents and submitted the required materials to the Office of the Attorney General (OAG) asserting attorney-client privilege.
The OAG issued a ruling on August 12, ordering Houston ISD to release the records and stating that attorney-client privilege did not apply.
Houston ISD filed a lawsuit in Travis County on September 11, looking to block the emails from release.
California Gov. Gavin Newsom has repeatedly slurred a federal judge by name, echoing President Trump’s history of diatribes against judges even before the current Democrat started copying the former Democrat’s social media style and insulting nicknames.
The perceived contender for the 2028 Democratic nomination for president may cluck his tongue again when he sees the latest order from U.S. District Judge Roger Benitez in a lawsuit against The Golden State’s alleged mandate on school districts to hide from parents their children’s asserted gender identity at odds with sex.
The President George W. Bush nominee ordered state Attorney General Rob Bonta and the California Department of Education to “show cause” on why they should not be sanctioned for “misleading” Benitez so he would remove them from the suit by teachers who allege their school district muzzled them and parents of “gender incongruent children.”
The state defendants’ motions to dismiss and opposition to the plaintiffs’ motion for summary judgment claimed that CDE had “withdrawn and conclusively replaced” an FAQ page that contained the challenged policies, which they claimed was the “only basis” for being named defendants and thus made the case moot, Benitez wrote.
“However, evidence demonstrates that the CDE may have merely moved the challenged content of the FAQ page to a new, required ‘PRISM’ training module,” as documented by the plaintiffs’ lawyers at the Thomas More Society, the judge said, ordering state defendants to explain their behavior Nov. 17 in court.
“From day one, officials from the local school district all the way to the governor’s mansion have tried to deflect responsibility” but “have now been caught not only lying to California taxpayers but attempting to mislead the Court to escape accountability,” TMS Executive Vice President Peter Breen said in a statement.
Based on early voting and some voting day results, no candidate secured over 50 percent of the votes cast, so the two highest vote recipients will move on to the runoff election, the date of which remains to be set by Gov. Greg Abbott.
The North Texas Senate seat was vacated when former state Sen. Kelly Hancock (R-North Richland Hills) resigned and was appointed by Abbott to fill the vacancy as the Texas Comptroller of Public Accounts.
Snip.
Wambsganss was endorsed early on in the race by Lt. Gov. Dan Patrick, who has vocally opposed expansion of casino gambling in Texas. She has also received support from Texans United for a Conservative Majority (TUCM), which opposes gambling expansion as well. Texans for Lawsuit Reform, a group not frequently on the same side of an electoral battle as TUCM, has also supported Wambsganss.
The Substrate startup has been doing the rounds in the news lately, thanks to its proposition of making chips using particle accelerators and X-rays instead of conventional EUV lithography, claiming it can eventually have angstrom-sized features at only $10,000 per wafer—in U.S. fabs, no less.
Oooo, where to begin? IBM tried experimenting with x-ray lithography in the 1980s and 90s, and found the rays were too energetic to use because they damaged wafers.
And technically, semiconductor equipment manufacturing already has particle accelerators: they’re called ion implanters and they’re used for gate dopants. Axcelis (formerly Eaton Semiconductor) and Applied Materials (both companies I worked for in the 1990s) make good money selling them, and there are a whole bunch of limits-of-physics reasons why you can’t use them for lithography. (Historical trivia: Applied Materials used to have their own in-house designed ion implanters, but their current offerings trace back to a competitor named Varian they bought in 2011.)
Those are bold claims, and an article by Fox Chapel Research (FCR) is seriously questioning whether they pay off.
The write-up is the first of two parts, and takes aim at not just the seemingly outlandish technological claims, but also at the track record of the venture’s founders, as well as the overall messaging on Substrate’s website. The start-up is backed by various investment funds, namely but not only Founders Fund, of whom Peter Thiel is part of.
The report says the founders are James and Oliver Proud, who reportedly have no experience in the semiconductor industry, nor do any of the investor funds. James’ latest venture was apparently the Sense sleep tracker, a product that had its inception on Kickstarter to the tune of $2.5m, but didn’t materialize until funding rounds raised over $50m. After release, the tracker was found to be borderline useless by reviewers and drew many comparisons to a scam.
ClowfishTV floats an interesting theory: A lot of those “AI-related” layoffs are just companies using that as an excuse to purge the woke from the ranks.
For more than half a century, Delaware stood as America’s corporate capital, renowned for its business-friendly laws, respected Chancery Court, and consistent legal rulings. But in recent years, leftist activist lawmakers and politicized judges have undermined that very foundation, sparking an exodus of major companies seeking stability and fairness to more welcoming states like Texas and Nevada.
On Wednesday morning, Coinbase joined the growing exodus, announcing on its website and in a Wall Street Journal op-ed by Chief Legal Officer Paul Grewal that it is moving its state of incorporation from Delaware to Texas.
“For decades, Delaware was known for predictable court outcomes, respect for the judgment of corporate boards, and speedy resolutions,” Grewal wrote in the op-ed.
However, he pointed out that recent inconsistent Chancery Court rulings and reliance on ad hoc legislative fixes do not create a sustainable business environment.
“Our decision to leave is about ensuring more predictable opportunities for the company, our shareholders, our customers and the new on-chain ecosystem we’re building,” he noted, adding, “Texas offers efficiency and predictability, in part thanks to recent corporate-law reforms that enhance governance flexibility and legal predictability.”
Grewal concluded, “Delaware wasn’t always the go-to choice for companies. At one point it was New Jersey, and before that New York. We’ve reached another inflection point in corporate law. The more states that can credibly attract companies, the better—and we’d like to see Delaware step up to stay in the mix. But as for Coinbase, you can find us in Texas….”
The exodus list from Delaware increases:
Tesla: Moved to Texas.
SpaceX: Moved to Texas.
Trump Media & Technology: Moved to Florida.
Dropbox: Moved to Nevada.
TripAdvisor: Moved to Nevada.
Roblox: Moved to Nevada.
Pershing Square: Moved to Nevada.
The Trade Desk: Moved to Nevada.
AMC Networks: Moved to Nevada.
Madison Square Garden Sports: Moved to Nevada.
Fidelity National Financial: Voted to move to Nevada.
So was a Delaware judge letting Elon Musk know how much he hated him for supporting Trump worth it?
“750-meter-long Chinese bridge partially collapses just weeks after opening.” From a landslide, but I’m betting the usual Chinesium/tofu drugs construction quality didn’t help…
At its Midlothian Data Center, alongside a number of state officials, Google announced a $40 billion data center infrastructure investment in Texas.
Sundar Pichai, CEO of Google and its parent company Alphabet, said that the investment will go toward the construction of three data center campuses located in Armstrong and Haskell counties.
Armstrong County is southeast of Amarillo. Haskell County is north of Abilene. Both counties have a whole lot of nothing there.
“They say that everything is bigger in Texas – and that certainly applies to the golden opportunity with AI,” Pichai stated.
“This investment will create thousands of jobs, provide skills training to college students and electrical apprentices, and accelerate energy affordability initiatives throughout Texas.”
Gov. Greg Abbott said the new Google AI data center announcement is “a Texas-sized investment in the future of our great state.” U.S. Sens. John Cornyn (R-TX) and Ted Cruz (R-TX) were also in attendance, along with Congressman Jake Ellzey (R-TX-06) and a number of other local officials.
“Google’s $40 billion investment makes Texas Google’s largest investment in any state in the country and supports energy efficiency and workforce development in our state,” Abbott added. “We must ensure that America remains at the forefront of the AI revolution, and Texas is the place where that can happen.”
Google has already officially broken ground on two other data centers in the state: one in Midlothian in 2019, and the other in Red Oak in 2023. The technology company has since announced further investments into data and cloud infrastructure to the tune of $2.7 billion.
This most recent announcement of a $40 billion investment will focus on building out infrastructure to support the three new data centers. Some of that investment includes building up new and existing energy storage facilities, advanced water use operations, and partnering with universities to offer technology training and education.
My reservations about Google’s AI notwithstanding, that will offer a bunch of real jobs for real Texans…assuming the AI bubble doesn’t burst before they get built.
Speaking of tech firms in trouble, video game maker Ubisoft (makers of Prince of Persia and Assassin’s Creed games) has not only postponed an earnings report, they’ve suspended stock trading. I can’t recall a single instance where that was a good sign. The last time we mentioned Ubisoft, they were pissing off Japanese gamers for including a black samurai in one of their games…
Ian McCollum looks at the new Rideout Arsenal Dragon, a low-bore-axis, lever-delayed pistol. It’s funky looking and has some interesting features, including complete non-tool disassembly. However, the price point would make it way too expensive to consider even if I had a job, he experiences several firing malfunctions testing it (though it is a prototype), and I fear the tiny little tabs it uses may not hold up under heavy use. Still a pretty interesting design.
The smarter, more capable version of Siri that Apple is developing will be powered by Google Gemini, reports Bloomberg. Apple will pay Google approximately $1 billion per year for a 1.2 trillion parameter artificial intelligence model that was developed by Google.
For context, parameters are a measure of how a model understands and responds to queries. More parameters generally means more capable, though training and architecture are also factors. Bloomberg says that Google’s model “dwarfs” the parameter level of Apple’s current models.
The current cloud-based version of Apple Intelligence uses 150 billion parameters, but there are no specific metrics detailing how the other models Apple is developing measure up.
Apple will use Gemini for functions related to summarizing and multi-step task planning and execution, but Apple models will also be used for some Siri features. The AI model that Google is developing for Apple will run on Apple’s Private Cloud Compute servers, so Google will not have access to Apple data.
Some small favors there.
Apple weighed using its own AI models for the LLM version of Siri, and also tested options from OpenAI and Anthropic, but it decided to go with Gemini after deciding Anthropic’s fees were too high. Apple already has a partnership with Google for search results, with Google paying Apple around $20 billion per year to be the default search engine option on Apple devices.
Though Apple is planning to rely on Google AI for now, it plans to continue working on its own models and will transition to an in-house solution when its LLMs are capable enough. Apple is already working on a 1 trillion parameter cloud-based model that could be ready as soon as 2026. Apple is unlikely to publicize its arrangement with Google while it develops in-house models.
I own an iPhone and a MacBookPro. How good is the existing Siri AI?
No idea. I never, ever use Siri, because I don’t want my devices listening to me, and I find the existing Mac and iOS interfaces quite sufficient for my needs. And if I did use Siri, I’d have found a way to turn off any “advanced” AI features anyway.
To be sure, a certain amount of now low-level routines might once have been considered crude forms of “artificial intelligence”: spell-checking, auto-completion, etc. But it seems that the more general a question or task handed to current generations of AIs, the more likely you are to get AI hallucinations.
And brand new vulnerabilities! I meant to include this piece on Gemini security flaws in the previous Google AI post, but somehow it fell through the cracks.
Cybersecurity researchers have uncovered three high-risk vulnerabilities – dubbed the Gemini Trifecta – in Google’s Gemini AI suite.
Researchers from security firm Tenable tested Google’s AI with search-injection attacks, log-to-prompt injection attacks, and exfiltration of the user’s saved information and location data.
The vulnerabilities they found exposed users to severe privacy risks. They allowed attackers to hijack cloud services, poison personalized searches, and secretly take over sensitive user data.
“This is a blind spot. We discovered that if an attacker could infiltrate a prompt, they could have been able to instruct Gemini to fetch a malicious URL, embedding user data into that request,” wrote the researchers.
After the findings were disclosed, Google reacted promptly to patch the vulnerabilities.
The first vulnerability was found in Gemini Cloud Assist. This tool is designed to help users make sense of complex logs in GCP by summarizing entries and surfacing recommendations. “While evaluating this feature, we noticed something that caught our attention: Gemini wasn’t just summarizing metadata; it was pulling directly from raw logs,” explained the researchers.
They successfully added attacker-controlled text into the logs to trick Gemini into executing instructions buried in log content.
“Typically, passive artifacts could become an active threat vector.”
The vulnerability could be triggered by a victim pressing the “Explain this log entry” button in GCP Log Explorer. The prompt injection hidden inside an HTTP User-Agent header could have tricked the system into executing unauthorized cloud queries.
The researchers shared one impactful attack scenario: inject a prompt instructing Gemini to query all public assets or for IAM misconfigurations, and then create a hyperlink containing this sensitive data.
“Attackers could also ‘spray’ attacks on all GCP public-facing services to get as much impact as possible rather than a targeted attack,” explained the researchers.
The second flaw targeted Gemini’s Search Personalization model. This tool tailors answers based on a user’s browsing history. However, the discovered vulnerability showed that the tool could be exploited by attackers.
“This personalization is core to Gemini’s value, but it also means that search queries are, effectively, data that Gemini processes. That led us to a key insight: search history isn’t just passive context, it’s active input,” noted the researchers.
They also discovered that an attacker could plant instructions that Gemini would later treat as legitimate queries by manipulating a victim’s Chrome search history with malicious JavaScript.
“We asked: If an attacker could write to a user’s browser search history, could that search history be used to control Gemini’s behavior, affecting the Gemini Search Personalization model?”
This exploit allowed the researchers to exfiltrate user-saved information and location data.
The third issue affected Gemini’s Browsing Tool. The Gemini Browsing Tool allows the model to access live web content and generate summaries based on that content.
Researchers tried to test whether they could instruct Gemini to send the user’s saved information to an external malicious server.
“AI systems don’t just leak through obvious outputs. They can also leak via functionality – especially through tools like Gemini’s Browsing Tool, which enables real-time data fetching from external URLs,” said the researchers.
After a couple of attempts, they succeeded in exploiting the tool.
Some of these are similar to previous security flaws that were fixed by various methods (encryption, tightened access controls, microservices, etc.) in response to previous exploits. But current computer security wasn’t constructed with the assumption that you would have an ultra-powerful but naive bottle djinn running with access to your system.
The history of Internet security has been a never-end war of tightening down security in one place only for hackers to find more attack surfaces to exploit.
With AI, it seems that the attack surface is now everything.
Remember when Google was a world-leading corporation whose motto was “don’t be evil”, universally trusted for Internet searches, branching out into other businesses and could seemingly do no wrong? You may not, since that was a good 15-20 years ago. Since then, Google has done plenty of evil to lose our trust, from spinning up useful services only to allow them to be killed off a few years later to letting itself be infected with social justice to ruining search results to plump ad revenues.
Now Google is infecting itself with AI across all its divisions, and the results are disasterous.
No, Cheney didn’t vote for Kamala in 2020, and indeed only announced outright opposition to Trump after January 6. Google’s AI garbage has conflated the 2020 and 2024 presidential elections.
This is far from the first time Google’s AI systems have made mistakes.
A whole bunch of YouTube channels were banned based on the actions of completely unrelated channels, and the creators blamed AI. YouTube eventually restored them and denied AI was involved, but does anyone really believe anything Google/YouTube says anymore?
But Google AI is definitely improving one thing: malware.
Google’s Threat Intelligence Group (GTIG) is warning that bad guys are using artificial intelligence to create and deploy new malware that both utilizes and combats large language models (LLM) like Gemini when deployed.
The findings were laid out in a white paper released on Wednesday, November 5 by the GTIG. The group noted that adversaries are no longer leveraging artificial intelligence (AI) just for productivity gains, they are deploying “novel AI-enabled malware in active operations.” They went on to label it a new “operational phase of AI abuse.”
Google is calling the new tools “just-in-time” AI used in at least two malware families: PromptFlux and PromptSteal, both of which use LLMs during deployment. They generate malicious scripts and obfuscate their code to avoid detection by antivirus programs. Additionally, the malware families use AI models to create malicious functions “on demand” rather than being built into the code.
Google says these tools are a nascent but significant step towards “autonomous and adaptive malware.”
PromptFlux is an experimental VBScript dropper that utilizes Google Gemini to generate obfuscated VBScript variants. VBScript is mostly used for automation in Windows environments.
Ah, Windows, a fecund garden of malware for over 30 years.
In this case, PromptFlux attempts to access your PC via Startup folder entries and then spreads through removable drives and mapped network shares.
“The most novel component of PROMPTFLUX is its ‘Thinking Robot’ module, designed to periodically query Gemini to obtain new code for evading antivirus software,” GTIG says.
The researchers say that the code indicates the malware’s makers are trying to create an evolving “metamorphic script.”
According to Google, the Threat Intelligence researchers could not pinpoint who made PromptFlux, but did note that it appears to be used by a group for financial gain. Google also claims that it is in early development and can’t yet inflict real damage.
The company says that it has disabled the malware’s access to Gemini and deleted assets connected to it.
Google also highlighted a number of other malware that establish remote command-and control (FruitShell), capturing GitHub credentials (QuietVault), and one that steals and encrypts data on Windows, macOS and Linux devices (PromptLock). All of them utilize AI to work or in the case of FruitShell to bypass LLM-powered security.
Beyond malware, the paper also reports several cases where threat actors abused Gemini. In one case, a malicious actor posed as a “capture-the-flag” participant, basically acting as a students or researchers to convince Gemini to provide information that is supposed to be blocked.
Google specified a number of threats from Chinese, Iranian and North Korean threat groups that abused Gemini for phishing, data mining, increasing malware sophistication, crypto theft and creating deepfakes.
So Google has created a power bottle genie that refuses to stay in the bottle, but will grant wishes to just about anyone, no matter how evil their intent.
Also, not limited to Google, researchers have demonstrated new exploits for AI browsers (or rather, very old exploits refurbished for the AI age).
Several new AI browsers, including OpenAI’s Atlas, offer the ability to take actions on the user’s behalf, such as opening web pages or even shopping. But these added capabilities create new attack vectors, particularly prompt injection.
Prompt injection occurs when something causes text that the user didn’t write to become commands for an AI bot. Direct prompt injection happens when unwanted text gets entered at the point of prompt input, while indirect injection happens when content, such as a web page or PDF that the bot has been asked to summarize, contains hidden commands that AI then follows as if the user had entered them.
Last week, researchers at Brave browser published a report detailing indirect prompt injection vulns they found in the Comet and Fellou browsers. For Comet, the testers added instructions as unreadable text inside an image on a web page, and for Fellou they simply wrote the instructions into the text of a web page.
When the browsers were asked to summarize these pages – something a user might do – they followed the instructions by opening Gmail, grabbing the subject line of the user’s most recent email message, and then appending that data as the query string of another URL to a website that the researchers controlled. If the website were run by crims, they’d be able to collect user data with it.
When Isaac Asimov crafted the Three Laws of Robotics, he thought that robots would have built-in safeguards deep in their source codes to prevent them from doing harm. What he never could have envisioned is multiple artificial intelligence being created as quickly as possible by competing corporations, none of whom seem to value safety over time-to-market, and that some of these AIs could be capable of modifying their own source code for greater speed and efficiency, so that no one knows precisely at any given time what exactly they’re running, and what data sets have been used to feed their pet Frankenstein monsters…
Having been out of work for a while, people ask me if I’ve been displaced by AI. My reply is “Not directly.” Indirectly, I think the factor is that just about all venture capital funds are throwing money at AI-related companies, meaning non-AI startups that might need technical writers aren’t being funded.
Having lived through the dotcom bust, I have to wonder how bad the fallout from the AI bubble bursting is going to be. The dotcom bubble wasn’t all beenz and pets.com…
…and it fueled a whole lot of subsidiary bubbles: PC and server manufacturers to run the software, Microsoft to run the PCs, semiconductor manufacturers to provide chips for the PCs and servers, semiconductor equipment manufacturers to build those same chips, network gear providers to connect the data centers, etc. And that only scratches the surface. Cisco, Dell, Compaq, Netscape, Yahoo, AOL, Oracle, Sun, HP, Intel, AMD, Applied Materials (where I worked 1997-2001), LAM Research, KLA-Tencor, all had huge growth spurts during the dotcom era as their customers spent big money to get “on the web.” Even dinosaurs like IBM, Motorola and DEC enjoyed business boosts from the era. All suffered in the wake of the dotcom bust, some being bought up or disappearing into other companies.
The same is true of today’s multi-trillion dollar AI boom. Companies like OpenAI may get the most ink, but a whole lot of other companies are getting boosted as well. Some of the names are even the same as the dotcom bubble: Microsoft, Oracle, AMD. Applied Material stock has gone through the roof now that I don’t own any. Cisco is just getting back to the level of their record stock highs during the dotcom era.
Data centers are supposedly planned or going up all around the country, and so many are buying Nvidia’s AI chips that they now boast a breathtaking $4.88 trillion market cap.
Someone is supposedly going to build a $165 billion data center in New Mexico near El Paso. That number is kind of insane, as you could build 5-10 cutting edge fabs for that kind of money. I don’t see how you get any sort of ROI on such a big upfront investment.
When the AI bubble busts (not if, when), a whole lot of these projects will likely come a cropper. A lot of people will have made a lot of money, AI will probably revolutionize a few industries and prove mostly hype in others, and retail investors and bondholders will be left holding the bag. Like the doctom bust, a lot of new companies will rise from the wreckage and start the cycle all over again.
And companies that can best take advantage of idle data centers and newly abundant nuclear power (assuming the boom even lasts that long) will be the ones poised to help build the next tech boom…
