Oracle would like you to believe this is what the future costs. I think it looks a lot more like a billionaire's infrastructure binge financed with other people's rent money.

Oracle was not scrambling to survive. The company just reported $17.2 billion in quarterly revenue, $3.7 billion in quarterly net income, and $553 billion in remaining performance obligations. Reuters and CNBC both note Oracle still had 162,000 employees as of May 2025 and then started telling workers it was cutting thousands of jobs. A follow up report from The Next Web says workers across the United States, India, Canada, and Mexico got termination emails on March 31 with access cut off immediately. USA Today later tied the cuts to at least 158 workers in California and 491 in Washington state, while GeekWire reported that the Washington list alone included more than 230 software developers plus dozens more software development roles.

That single sequence tells you what Oracle thinks the AI boom is for. The company has money, contracts, and demand. What it does not want is to let workers stand between Larry Ellison's AI empire and the balance sheet he wants for it.

This is a financing choice

Oracle told investors in February that it expected to raise $45 billion to $50 billion during calendar year 2026 to expand Oracle Cloud Infrastructure for customers including AMD, Meta, Nvidia, OpenAI, TikTok, and xAI. Within days, the company said it had already raised $30 billion through investment grade bonds and mandatory convertible preferred stock. The bond offering alone was a $25 billion deal spread across notes due from 2029 to 2066. The preferred stock side added 100 million depositary shares tied to a new 6.50% mandatory convertible series.

If this were just aggressive expansion, I would still worry about the scale of it. But Oracle's own filings show why management went hunting for payroll savings at the same time. In the first nine months of fiscal 2026, Oracle spent $39.17 billion on capital expenditures, up from $12.14 billion a year earlier. Its trailing four quarter free cash flow swung to negative $24.736 billion after capital expenditures of $48.25 billion. Property, plant, and equipment on the balance sheet nearly doubled from about $43.5 billion to $83.6 billion. Non current borrowings climbed to about $124.7 billion, with current borrowings adding another $9.9 billion. The same 10 Q says Oracle's fiscal 2026 restructuring plan can run up to $2.1 billion and had already booked $982 million in restructuring expense by February 28.

It is the balance sheet of a company building data centers at a speed its cash generation cannot comfortably support.

And Oracle is not even subtle about what those data centers are for. Reuters reported this month that developer Related Digital was nearing $16 billion of financing for an Oracle data center campus in Michigan, with Blackstone and Bank of America tied to the structure. It is Stargate money, hyperscale money, and the kind of spending spree executives love because it sounds historic, looks great in a keynote, and gives them an excuse to tell everyone else to be grateful for austerity.

Oracle admitted what it is doing

The ugliest line in this whole story did not come from a layoff memo. It came from Oracle's own earnings release. The company told investors that AI models for generating code have become so efficient that it has been restructuring product development teams into smaller groups and can build more software in less time with fewer people.

Oracle did not bother with euphemism here. It told investors the point was fewer people.

This is why I get so tired of hearing that AI is just a neutral tool and we should wait to see how companies use it. They are telling us how they plan to use it. They plan to use it to justify smaller teams, lower labor costs, and bigger capital bets. The people building the products become a line item to squeeze so the company can buy more GPUs, pour more concrete, and brag about more capacity.

I want AI used to take drudge work off people's backs. If a tool really lets engineers spend less time on the tedious parts of the job, the humane outcome would be better products, shorter hours, or at least less burnout. Oracle chose the shareholder outcome instead. It took the gain as an excuse to cut heads while asking markets for tens of billions more to feed the machine.

The steelman still fails

The strongest defense of Oracle goes like this: demand for AI infrastructure is real, cloud capacity is constrained, and a company that lands gigantic contracts has to move fast or lose the business. Oracle says demand for AI infrastructure continues to exceed supply, and it also says much of the new contract load is supported by customer prepayments or customer supplied GPUs. On that view, reorganizing teams and automating more coding is just how you match staffing to a new technical reality.

I do not buy it.

If these contracts are as secure and well funded as Oracle says, then the moral case for throwing thousands of workers into the street gets weaker, not stronger. A company sitting on hundreds of billions in contracted revenue, tens of billions in cash and cash equivalents, and a financing plan measured in tens of billions of dollars does not need to make software developers finance the transition with their livelihoods. It chooses to do that because investors reward efficiency ratios faster than they reward decency.

USA Today quoted Morningstar analyst Luke Yang arguing that the layoffs should improve operating efficiency and boost revenue per headcount. That is exactly the problem. Revenue per headcount is one of those bloodless phrases that makes cruelty sound like mathematics. Translate it into English and it means the company wants the same or more output with fewer people collecting paychecks.

This is what the AI boom looks like from below

The AI sales pitch is always about abundance, more intelligence, more productivity, more scientific progress, more possibility. Down on the ground, it keeps looking like pink slips, debt offerings, power hungry campuses, and executives promising that this time the spending will pay off.

Oracle is not the only company doing this, but it is unusually honest about the logic. It is raising extraordinary sums to build data center capacity, it is carrying extraordinary capital spending, and it is openly telling investors that new AI coding tools mean it can operate with fewer people. Put those facts together and the story stops being abstract. Workers are subsidizing the buildout.

That is the part of the AI future I refuse to cheer for. If the industry wants to claim this technology will make life better, it should start by proving the benefits flow to the people doing the work, not just to the executives signing debt papers and the billionaires who already own too much. Right now Oracle is proving the opposite.

A Cornell Ph.D. student briefly attended a pro-Palestinian protest at a campus career fair in September 2024. Months later, Homeland Security Investigations, the intelligence arm of ICE, subpoenaed Google for data tied to his account. Google complied before warning him, according to reporting from The Intercept, the student's own EFF account, and an EFF complaint summary.

The student is Amandla Thomas-Johnson. The Intercept reports that he and fellow Cornell graduate student Momodou Taal went into hiding after the Trump administration began targeting international students who had spoken out about Israel's war on Gaza. Thomas-Johnson later described federal agents coming to his home, a friend being detained at the Tampa airport and questioned about his whereabouts, and eventually fleeing the United States.

The timeline matters here. The Intercept reported that HSI sent the subpoena to Google on March 31, 2025. EFF says Google received it in April, complied on May 8, and only then sent Thomas-Johnson an email saying it had already responded to law enforcement process. That was not notice in any meaningful sense. It was a receipt after the handoff was complete.

He never got a chance to challenge it. The data was already gone.

Here's why this matters: Google has promised users for nearly a decade that “when we receive a request from a government agency, we send an email to the user account before disclosing information.” That promise exists for a reason. Advance notice gives you a chance to go to court, to challenge the subpoena, to argue that the request is overbroad or politically motivated. When users get that chance, it works. The Intercept reported that Google and Meta gave Taal advance notice of subpoenas for his accounts, and ICE later withdrew them.

In Thomas-Johnson's case, Google skipped the notice. The subpoena landed, Google sat on it for over a month, and then complied without a word until the data was already in ICE's hands.

ICE didn't need a warrant for this. An administrative subpoena is an investigative tool that federal agencies can issue without a judge's approval, without probable cause, and without any judicial oversight. ICE invoked 8 USC 1225, a provision that gives immigration officers broad authority to demand documents “relating to the privilege of any person to enter, reenter, reside in, or pass through the United States.” One study found ICE uses this same authority hundreds of times each year to demand data from tech companies.

Here's the thing: companies are not legally required to comply with administrative subpoenas. There is no penalty for refusing. If Google had said no, ICE's only option would have been to drop it or go to court, where ICE would have needed to justify the request to a judge. When companies push back, the government often backs down. The ACLU helped anonymous Instagram users fight off similar DHS subpoenas in California and Pennsylvania that sought to unmask accounts tracking ICE raids. DHS withdrew both.

Google's outside counsel told the Electronic Frontier Foundation that the company has an internal practice called “simultaneous notice,” where it complies with a subpoena and notifies the user on the same day when a government deadline approaches. The EFF's response was blunt: this is not notice. This is a receipt after the fact. And through this hidden practice, Google has likely broken its promise to users “numerous other times over the years.”

On April 15, the EFF filed formal complaints with the California and New York attorneys general, asking them to investigate Google for deceptive trade practices. The legal theory is straightforward: Google publicly promised users advance notice before handing their data to law enforcement. It didn't deliver. The FTC's Cambridge Analytica case established that misleading users about data practices constitutes deceptive trade practices. California law provides for civil penalties of up to $2,500 per violation.

Google's spokesperson offered the usual corporate non-answer: “Our processes for handling law enforcement subpoenas are designed to protect users' privacy while meeting our legal obligations.” Read that again. It says nothing. The question isn't whether Google's processes are “designed” to protect privacy. The question is whether they actually protect anyone. In Thomas-Johnson's case, they didn't. Google had over a month between receiving the subpoena and complying. It chose not to use that time to warn him.

Some will argue that Google was just following the law, that it had no choice. That's false. Companies refuse administrative subpoenas regularly. Apple fought the FBI in federal court rather than build a backdoor into a mass shooter's iPhone, and the government backed down rather than face a ruling against it. Meta gave Taal advance notice of his subpoena, which is exactly what gave him time to fight it and win. Google could have done the same for Thomas-Johnson. It chose not to.

The ICE subpoena was broader than the data Google says it ultimately produced. The Intercept, which obtained a copy of the subpoena, reported that ICE asked for Thomas-Johnson's name, address, IP addresses, phone numbers, session times and durations, the list of Google services tied to the account, and even credit card and bank account numbers. Google's attorney later said the company produced only “basic subscriber information,” not message contents, communications metadata, or location information. Even that narrower category can still be invasive. As Thomas-Johnson wrote for EFF, subscriber records can expose where you sleep, when you were online, and who the government thinks it needs to find next.

This case is not really about one student. It's about the pipeline. The federal government doesn't need to hack your devices or break your encryption when companies like Google will simply hand your data over on request. ICE sends an administrative subpoena, no judge required. Google receives it, doesn't notify you, and ships the data. The entire apparatus of constitutional protection, the Fourth Amendment, the requirement for warrants and probable cause, gets bypassed because a corporation decided to comply rather than push back.

And it's working exactly as designed. Thomas-Johnson is now in Dakar, Senegal, wondering if he's “a marked individual.” Other international students are watching what happened to him and deciding whether it's worth attending a protest, writing an email, posting online. The surveillance state doesn't need to monitor everyone. It just needs to make enough examples that everyone else self-censors.

Google built a $2 trillion empire on collecting data about billions of people. It promised to tell you before handing that data to the government. It broke that promise, not because it was legally forced to, but because choosing to comply was easier than choosing to fight. The people who pay for that choice are the ones with the most to lose: students, activists, immigrants, anyone whose political speech makes them a target.

The EFF is asking state attorneys general to investigate. That's the right call. But the deeper fix is structural. As long as tech companies are allowed to hoard detailed data about billions of people and hand it over to law enforcement through judge-free administrative processes, this will keep happening. Advance notice is a floor, not a ceiling. The real question is why Google has all this data to hand over in the first place, and why a federal agency can demand it without ever seeing a judge.

Trump already pardoned or commuted the January 6 defendants he cared about. Now his Justice Department wants the record cleaned up too.

On Tuesday, Trump’s DOJ asked a federal appeals court to vacate the seditious conspiracy convictions of Proud Boys and Oath Keepers leaders and send the cases back so prosecutors can dismiss them with prejudice. That means for good. The motion, signed by U.S. Attorney Jeanine Pirro, says continuing these cases is “not in the interests of justice.” A pardon leaves the crime in the history books and forgives it anyway. Vacating the conviction goes further. It tries to make the crime disappear from the books.

I keep coming back to that difference. On Trump’s first day back in office, he commuted the sentences of 14 named January 6 defendants and issued a blanket pardon for everyone else convicted in the attack. That already told every future right wing thug that loyalty to Trump comes with an insurance policy. What the administration is doing now is even uglier. It is trying to erase the legal findings that some of the most important January 6 cases were not random crowd chaos. They were acts of organized political violence.

These were the cases that proved planning

The defendants in these appeals were not nobodies who wandered into the Capitol because the mob carried them along. AP reports the DOJ is trying to wipe away convictions for Oath Keepers founder Stewart Rhodes and Proud Boys leaders including Ethan Nordean, Joseph Biggs, and Zachary Rehl, along with others tied to those cases. Rhodes had been sentenced to 18 years after Judge Amit Mehta said he remained a danger to the country and to democracy itself. Kelly Meggs got 12 years. Nordean got 18 years. Biggs got 17. Rehl got 15. Enrique Tarrio, the Proud Boys chairman who had already received a full pardon rather than a commutation, got 22 years after a jury convicted him and three lieutenants of seditious conspiracy.

Those sentences existed because prosecutors did the work and juries believed it. In the Oath Keepers case, the government showed that Rhodes and his circle organized “quick reaction force” teams and stockpiled weapons at a Virginia hotel in case they decided to move guns into Washington. In the Proud Boys case, prosecutors showed jurors encrypted chats, coordination, and leadership decisions that helped drive the attack forward. AP reported that Tarrio, even from outside Washington, egged the operation on with messages like “Do what must be done,” then “Do it again,” and finally “We did this.”

That is why these convictions mattered so much. January 6 injured more than 100 police officers and grew into the largest investigation in Justice Department history, with more than 1,500 arrests and over 170 people accused of using deadly or dangerous weapons against police. But the deeper point of the seditious conspiracy cases was always that the violence had organizers. Trump and his movement have spent years trying to sell January 6 as an overexcited protest, a tourist visit, a “day of love,” anything except what it was. These verdicts were the factual rebuttal. They put into the court record that key parts of the attack were planned.

The strongest defense still collapses on contact with the evidence

The best defense argument goes like this: seditious conspiracy is an extreme charge, the government stretched it too far, and Trump was the real instigator anyway. You can hear pieces of that argument across the January 6 cases. Rhodes’ lawyers said there was no plan to storm the Capitol. Tarrio’s lawyer argued Trump’s words and Trump’s anger caused the attack, not Tarrio himself. Proud Boys and Oath Keepers defense teams leaned hard on the claim that prosecutors were turning swagger, bluster, and ugly rhetoric into a treason case. Trump being guilty does not make Rhodes and Tarrio innocent. It means the plot had both an instigator and field commanders.

That argument still lost in front of juries for a reason. These were not convictions built on vibes or guilt by association. They came after long trials, large evidence records, and detailed findings about who planned what. Rhodes was convicted after a nearly two month trial. Tarrio and the other Proud Boys defendants were convicted after more than three months of testimony. The evidence was not a single overheated speech. It was the accumulation of chats, operational roles, logistics, breach leadership, weapons staging, and defendant statements before, during, and after the attack.

You can think sedition charges should be used sparingly and still understand why these cases fit. In fact, that is part of what made them so important. The federal government had not been winning this kind of charge every other week. These convictions landed because prosecutors showed juries something more serious than ordinary riot behavior. They showed organized efforts to stop the transfer of power.

Jeanine Pirro is the tell

What gets me almost as much is who is signing these motions.

Pirro is not some neutral caretaker trying to tidy up loose ends. NPR noted that Trump installed Fox host Jeanine Pirro as interim U.S. attorney in Washington after dropping his earlier failed pick, and that Pirro had pushed false 2020 election fraud claims while acting as one of Trump’s loudest media loyalists. This is the person now telling a federal court that vacating these convictions serves justice.

Of course it serves Trump, which is the whole point.

The administration is not satisfied with mercy. Mercy still implies wrongdoing. Trump wants vindication for his foot soldiers and humiliation for everyone who investigated them. After the pardons and commutations came the language about January 6 defendants as “hostages” and the White House effort to recast the attack as persecution of patriots. Now comes the legal cleanup stage, where the government itself tries to pull the findings out of the record.

That is how authoritarian politics treats political violence committed on its own behalf. The people who attack democracy are recoded as victims. The prosecutors become the villains. The history itself gets edited until the next mob knows exactly what lesson to take: hold on long enough, get your side back in power, and the state will try to make your criminal record evaporate.

What this teaches the next group

That is the real danger here. Trump’s DOJ is telling every future militia freak, every street gang in a tactical vest, every wannabe coup tourist with a Telegram channel that the consequences are temporary.

Do a little time, wait for your boss to win the next election, then hope for a pardon or a friendly prosecutor who comes along later and asks a court to erase the conviction itself. That is the lesson.

And once that lesson settles in, every future election becomes more dangerous. Judge Mehta warned when he sentenced Rhodes that Americans may now “hold our collective breaths every time an election is approaching.” He was right then. He is even more right now.

A justice system captured by a would be strongman does not only protect the leader. It teaches his followers that violence in his service will look survivable. That is where this ends if people keep pretending these are just paperwork moves.

The paperwork is the point. Trump wants the country to forget that January 6 had planners, leaders, and men who treated stopping the peaceful transfer of power like an operational mission. These convictions proved otherwise. That is exactly why his Justice Department wants them gone before the next loyal mob takes notes.

Viktor Orbán finally lost, and the American right should be forced to stare at that result until the fantasy wears off.

This was not some random European election with no relevance to us. Orbán spent years turning Hungary into the closest thing the MAGA movement has to a governing model. He bragged about building an “illiberal democracy.” He rewrote the constitution after taking a two thirds majority, bent the electoral system in his favor, packed institutions with loyalists, curbed media freedom, weakened judicial independence, and governed by decree under a state of emergency after 2022 (Reuters). The European Parliament called Hungary a “hybrid regime of electoral autocracy” back in 2022, citing problems with judicial independence, corruption, media freedom, and LGBTQ+ rights (AP). Freedom House's 2025 report said national, regional, and local media were dominated by pro government outlets and described the Sovereignty Defense Office targeting independent journalists with claims they served foreign interests (Freedom House). In plain English, Orbán spent sixteen years rigging the system, crushing dissent, and calling it patriotism.

That is why American conservatives loved him. Orbán was proof of concept. He showed them what it looks like when the culture war becomes state policy. He built border fences, attacked immigrants, wrapped himself in Christianity, turned “national sovereignty” into a weapon against dissent, and treated independent media, universities, NGOs, and queer people as enemies to be neutralized. American conservatives saw Hungary as the place where MAGA style politics had years to harden into a system.

And in the end, Hungarian voters got a long look at what that model actually produces.

They got stagnation. They got soaring living costs. Reuters reported that Hungary endured the worst inflation surge in the European Union after Russia's 2022 invasion of Ukraine, that household morale sagged, and that many Hungarians had seen a decline in living standards they had not experienced during Fidesz's years in power (Reuters). By election week, Reuters described three years of economic stagnation, soaring living costs, and the enrichment of oligarchs close to the government as key drivers of voter anger (Reuters).

That matters because authoritarian politics always sells itself as strength. The pitch is simple. Stop worrying about niceties like pluralism, checks and balances, or minority rights. Give one tough guy enough power and he will clean up the country. Orbán got that chance. What Hungary got was a leader who could dominate the press, bully civil society, and still could not make groceries cheaper. That is the whole scam in one sentence.

That is the part the MAGA crowd never wants to talk about. Culture war governments are often very good at finding enemies. They are much worse at governing. They can fill television with propaganda. They can make life miserable for migrants, journalists, judges, and gay people. They can accuse every critic of serving foreign interests. What they cannot do is suspend material reality. When your rent, your bills, and your food costs keep climbing, the nationalist pageant starts to look like what it is: a protection racket for the people at the top.

Orbán understood the MAGA connection well enough to make it explicit. Reuters described Hungary as a “laboratory of sovereigntist politics” for Trump's movement (Reuters). He had Trump's endorsement. He had fans across Europe's far right. He had years of praise from American conservatives who treated Budapest like a theme park for reactionary fantasies. They loved the speeches about Christian civilization. They loved the cruelty toward migrants. They loved the attacks on “globalists.” They loved that he used democratic legitimacy to hollow out democracy itself. Reuters also reported before the vote that even conservative analysts and poll watchers did not think high profile backing from Washington was likely to change the outcome, because Hungarian voters were focused on cost of living and other domestic issues (Reuters).

Then JD Vance showed up and somehow made the whole thing look even worse.

Reuters reported that Vance went to Budapest days before the vote, openly endorsed Orbán, attacked the European Union for supposedly interfering in Hungary's election, and joined Orbán at a rally where he literally phoned Trump from the podium so Trump could bless the campaign over a speakerphone (Reuters). AP described the trip as an unusual break from the norm of foreign leaders avoiding direct roles in other countries' campaigns, even as Orbán himself had denounced much milder outside commentary as an assault on Hungarian sovereignty (AP). Peter Magyar answered by warning about foreign interference and saying Hungarian history would be written in Hungary, not Washington, Moscow, or Brussels (Reuters).

I want to be careful here. I cannot prove Vance personally cost Orbán the election, and the best reporting does not prove that either. Reuters explicitly noted it was not immediately clear whether Vance's support would strengthen Orbán's candidacy. What the reporting does show is that Orbán was already in deep trouble before Vance arrived. On April 1, Reuters cited a poll putting Tisza at 56% to Fidesz's 37% among decided voters (Reuters). The last Median poll, conducted April 7 through 11 while Vance's visit was still fresh, projected a Tisza supermajority rather than a late Orbán recovery (Reuters). Then Orbán got blown out anyway, with AP reporting Tisza ahead 53% to 37% with 93% of votes counted (AP).

So the strongest version of the argument is not that Vance caused the loss. It is that Vance failed to save Orbán, and made the defeat look even more humiliating. There was real contemporaneous skepticism that the visit could backfire. The Guardian wrote before the vote that Vance's intervention “may not help Orbán to the election victory he craves,” and the Irish Times noted that he was endorsing Orbán while Fidesz trailed Tisza by an average of 10 points in Politico Europe's poll of polls (The Guardian, The Irish Times). Orbán was already trying to campaign as the great defender of Hungarian sovereignty while blaming outside forces for everything. Then the vice president of the United States arrived to campaign for him in person. If you are a Hungarian voter who is already tired of the corruption, the stagnation, and the performative strongman act, that does not make Orbán look strong. It makes him look rented.

It also reinforced the basic truth about MAGA abroad. It is not a patriotic movement in any honest sense. It is a transnational project of reactionary elites who cheer each other on while ordinary people pay the price. Orbán borrowed from Trump. Trumpworld borrowed from Orbán. Vance flew in to help. The Kremlin wanted Orbán to survive too (Reuters). None of that was about Hungarian freedom. It was about keeping one more strongman friendly to oligarchy, hostile to liberal rights, and useful to the authoritarian right.

And look at what Orbán was still doing even as power slipped. In early 2025 he promised new legislation against media and NGOs receiving foreign funding, explicitly saying Hungary would follow the example of the new U.S. government (Reuters). Weeks later he was promising an “Easter cleaning” against what he called a shadow army of journalists, judges, politicians, and NGOs, while his party pushed constitutional changes targeting dual citizens and the LGBTQ+ community (Reuters). Human Rights Watch warned that a 2025 bill would let a government appointed body strangle foreign funded civil society and media groups ahead of the election (HRW).

That is the real MAGA export package: repression, scapegoating, media capture, corruption, and a permanent search for internal enemies. The promised revival never arrives for ordinary people. The humiliation does. The billionaires, party loyalists, and propagandists get richer. Everybody else gets lectures about sacrifice.

Orbán's defeat does not mean the authoritarian right is finished. I wish it did. These movements lose elections and keep poisoning politics. They leave institutions damaged, public trust shattered, and plenty of loyalists behind. Hungary is going to need years to undo what Orbán did.

But this result still matters. It matters because one of the global right's favorite success stories just ended in failure. It matters because voters living under a full scale MAGA style experiment decided they had seen enough. And it matters because every American who thinks Trumpism is some kind of populist answer should look at Hungary and understand what the pitch always hides. The cruelty is real. The corruption is real. The oligarchy is real. The prosperity is the lie.

That is why Orbán lost. Sooner or later, people notice when the strongman act leaves them poorer and less free.

The SAVE Act keeps making headlines, and it should. It is a brutal piece of work that would strip millions of Americans of their right to vote. But if you only focus on the SAVE Act, you miss the bigger picture. What Republicans are doing is not a set of isolated election fights. It is a coordinated strategy to make voting harder, shrink the electorate, and make sure the people who can clear every new hurdle are the voters they want.

The bill would require documentary proof of citizenship to register to vote. That sounds harmless until you look at the numbers. The Brennan Center estimates that about 21 million American citizens do not have ready access to the documents the bill would demand. That is about 9 percent of voting age Americans.

These are not people without citizenship. They are eligible voters who do not have a passport on hand, cannot easily get a birth certificate copy, or have documents that do not line up neatly with current ID rules. NPR's explainer on the SAVE Act walks through the documentary proof requirement and the voter registration burden it would create.

That matters because the fight does not stop at registration. Look at mail voting. PBS NewsHour reported that the Postal Service changed its postmark rules, which means some mail dropped off on a given day may not receive a postmark from that same day. For voters mailing ballots close to Election Day, that can be the difference between a counted vote and a discarded one. Election officials are already telling voters to mail ballots earlier because the old assumptions no longer hold.

Republicans are pushing on that opening too. Democracy Docket reported that the Republican National Committee asked the Supreme Court to rule that ballots arriving after Election Day should not be counted even when they were mailed on time under state law. And this is not a tiny edge case. The National Conference of State Legislatures notes that 16 states, plus Washington, D.C. and several territories, currently accept at least some ballots that are postmarked by Election Day and received later. If the Court accepts the RNC's argument, states that built their systems around those deadlines will have to throw out ballots they currently count.

Now go back to the passport point. If Republicans want proof of citizenship rules, then access to passports becomes part of access to the ballot. PBS NewsHour also reported that the State Department ordered about 1,400 nonprofit public libraries to stop processing passport applications. Those libraries were convenient places for working people to apply for passports without taking a day off to visit a distant government office.

That is why these stories belong together. The SAVE Act raises the paperwork barrier. The passport change removes one of the more accessible ways to clear that barrier. The mail ballot fight makes it easier to reject votes even after people manage to register and cast a ballot. Each move hits a different part of the process, but the effect is the same. Fewer people vote. More ballots get tossed. The electorate gets smaller and more tilted toward the people Republicans want voting.

And all of this is trying to address a problem that does not exist. Voting by people without citizenship is already illegal and vanishingly rare. The fraud story is a pretext. The real project is to shape the electorate by making participation harder.

This is what voter suppression looks like now. It usually does not arrive with the blunt ugliness of the past. It comes wrapped in paperwork rules, process changes, and legal arguments that sound technical until you ask who gets burdened and who benefits. The SAVE Act is just one front in a broader assault on democracy. The war, however, is far from over, and it is up to everyone with a conscience to do their part to protect the freedoms future generations are counting on us to defend.

People bought Meta's AI glasses thinking they were buying a little bit of the future.

What they apparently bought was a face-mounted camera whose footage could end up in front of human reviewers.

That is the part that should make your stomach turn. According to an Ars Technica investigation, footage from Ray-Ban Meta smart glasses was reviewed by workers at a Kenya-based subcontractor, and the material reportedly included people using the toilet, having sex, undressing, and even financial information. Meta now faces a class action lawsuit in the US, and the UK's Information Commissioner's Office has written to the company over the report.

If that reporting is accurate, then Meta did not sell people smart glasses. It sold them an ambient surveillance device with human reviewers in the loop.

The product promise and the real product

This is what keeps happening with consumer AI. The product promise is always personal convenience. Ask a question out loud. Capture a memory. Get a little help from the machine. The real product, one layer down, is data extraction at industrial scale.

That gap matters more with glasses than with phones. A phone is something you pick up. Glasses sit on your face and point at the world by default. They do not just record the user. They record everyone around the user, including people who never agreed to be part of Meta's training and review pipeline.

Once human review enters the picture, the marketing falls apart.

A lot of AI features still rely on people behind the scenes: annotators, reviewers, contractors, safety teams. I understand why that happens technically. Models need labels. Systems need debugging. Companies need quality control. Fine. But if your product depends on strangers being able to watch deeply intimate footage, then privacy is not a feature of the product. It is a casualty of how the product works.

The word “private” hides the real operating model

The strongest defense for Meta is the familiar one: human review is limited, disclosed somewhere in the terms, and necessary to improve the system. Maybe. But that defense depends on people understanding what they actually bought.

I do not think most people hear “AI glasses” and imagine offshore contractors watching clips of them in the bathroom.

That is not because users are naive. It is because companies are very good at burying the real operating model behind language like “improve our services” and “help train AI.” Those phrases are technically informative in the same way a skull-and-crossbones icon is technically a label. They are not honest about the lived reality.

The lived reality is this: once you normalize always-on cameras plus cloud AI plus human review, you have built a system that can absorb the most intimate parts of ordinary life and turn them into corporate training data.

That is a scandal, not an implementation detail.

Wearables make surveillance social

What makes this story worse than the usual privacy outrage is that smart glasses push the surveillance burden onto everyone else too.

If I post too much on social media, that is my bad decision. If I wear AI glasses into a café, on a train, into a store, or around my family, I am making that decision for everybody in view. Their image, their voice, their bad hair day, their kid in the background, their credit card on the counter. It all becomes potential input.

That is why this category makes me more uneasy than the usual “your app collected too much data” story. Wearables turn surveillance into a social default. The wearer gets the novelty. Everyone nearby gets drafted in.

And this is exactly how surveillance gets normalized in tech: not through one dramatic betrayal, but through a hundred convenience features that train people to stop expecting privacy in ordinary life.

The ugly truth about AI hardware

The industry wants AI hardware to feel magical, effortless, and inevitable. But the ugly truth is that a lot of these products work because there is a hidden labor system and a hidden data pipeline underneath them.

That hidden system is where the harm lives.

The people selling AI glasses want you thinking about the assistant in your ear. They do not want you thinking about the contractor reviewing footage, the retention policies, the training datasets, or the fact that the camera is always one product update away from becoming more invasive than it already is.

This is why I keep coming back to the same point with AI: the technology is real, and some of it is genuinely exciting. But the business model around it keeps dragging us somewhere awful. We are told these systems are here to help us see, remember, and understand more. In practice, they keep becoming better ways for companies to collect, analyze, and normalize access to human life.

Meta did not stumble into this tension. It built a product category where the value comes from turning the world into model input.

If you sell people “AI glasses” and the result is that strangers may end up reviewing clips of their most private moments, then the honest name for the product is not smart glasses.

It is surveillance you wear on your face.

TikTok told the BBC it will not add end-to-end encryption to direct messages because it thinks encryption would make people “less safe.” The thing that keeps hackers, the company, and governments out of your private conversations is being framed as a threat.

TikTok told the BBC it believes end-to-end encryption “prevents police and safety teams from being able to read direct messages if they needed to.”

“Safety” is the one word that can bulldoze privacy debates, especially when kids are involved. But if a platform says it needs the ability to read your messages to keep you safe, it is asking for a master key.

And once that master key exists, it is not just for TikTok. It becomes an access path for law enforcement. You can call that “lawful access” if you want. It is still domestic surveillance.

End-to-end encryption is a design choice

Most people hear “encrypted” and assume that means “private.” It usually does not. A lot of services encrypt messages in transit, then decrypt them on the server.

End-to-end encryption (E2EE) means only the people in the conversation can read it. The service that delivers the message cannot. There is no server-side copy employees can browse, and no plaintext to hand over.

TikTok says its DMs are still protected with “standard encryption,” and that only authorized employees can access messages in limited situations, like responding to a valid law enforcement request or a user report.

That is the whole story. If employees can read the content, your DMs are not end-to-end encrypted. The company can read them, which means the government can potentially get them too.

Safety teams do not need a master key to do safety work

TikTok's argument is the oldest one in tech policy: if the company cannot read messages, it cannot help when something bad happens.

Bad things do happen in DMs. Grooming and harassment are real. Child safety groups like the NSPCC and the Internet Watch Foundation praised TikTok's choice for that reason.

But “we need to be able to read everything” is a lazy, high-risk solution to a hard problem. No safety team is proactively reading DMs at scale. What platforms actually do is respond to reports, look for patterns of abuse, and use product design to reduce exposure in the first place.

Second, E2EE does not prevent a platform from doing the stuff that actually matters for protecting minors. You can make DMs from strangers opt-in by default for teens, put friction and rate limits in the path of unknown accounts, block obvious spam and grooming patterns, and let users report a thread by sharing what they choose to share.

You can build strong guardrails without giving the company a standing ability to open every private envelope.

Internet Society's piece Encryption keeps kids safe online makes a point worth repeating: encryption protects kids too, including from breaches and profiling.

The real risk is the access path

Once you decide the company must be able to read DMs, you have created a permanent high-value target. Even with good controls, that access eventually leaks or gets used in ways users never agreed to.

And it aligns perfectly with what governments have asked for, over and over: a world where providers can be compelled to produce message content when served with legal process (see EFF on the long-running push for encryption backdoors).

The safest messages are the ones the platform cannot read. You cannot leak what you cannot access. You cannot be compelled to hand over plaintext you never had.

If you keep DMs readable, you are keeping an access path open for domestic surveillance. TikTok is explicit that it can access DMs in response to a “valid law enforcement request.” End-to-end encryption removes that option by design.

This matters for any company, but it matters even more for TikTok because of its “combustible optics,” as analyst Matt Navarra put it to the BBC. TikTok is already fighting suspicion about its ownership and state pressure. Keeping DMs decryptable just adds another “trust us.”

And this is where the “for safety” framing really annoys me. If TikTok wants to differentiate itself from rivals by keeping DMs readable, it should say that plainly. It should tell users: we will read DMs when we decide it is necessary, and we will keep them available for law enforcement requests.

That is the deal, and the rest is marketing.

The sharp part: stop calling encryption controversial

End-to-end encryption is the baseline now. Signal and WhatsApp do it, and iMessage has for years (Apple’s iMessage security overview). “We cannot read this” is the cleanest security promise a messaging service can make.

If TikTok wants to keep DMs in the “we can read it” category, fine. Users should know what they are signing up for.

But calling E2EE a safety risk flips the world upside down. The risk is not that criminals will hide from law enforcement. The risk is that regular people will keep being trained to accept surveillance as the default setting for communication.

If a platform tells you it needs to read private messages to keep you safe, believe the quiet part: it is building a system where somebody, somewhere, can read them. And once that door exists, it will get used.

Call it what it is: surveillance.

When the Pentagon tells AI companies it will only work with vendors who agree to “any lawful use,” that's the leverage play. It forces the vendor to pre-approve future uses (surveillance, targeting, whatever gets labeled “lawful”) before anyone can object.

Anthropic's CEO says the Department of War (yes, that's what they're calling it) is pushing to remove two specific contract safeguards: a ban on mass domestic surveillance and a ban on fully autonomous weapons, meaning systems that can select and engage targets without human involvement. Anthropic says no. They also say they're being threatened with a “supply chain risk” label or even the Defense Production Act to force those safeguards away.

Read that and tell me this is normal procurement drama. (Anthropic’s full statement)

I don't doubt there are good people in the DoD. I also don't care. Institutions don't run on good intentions. They run on incentives and authority. If you build a machine that makes surveillance cheap and violence scalable, you are not just trusting today's “good people.” You're handing that machine to whoever comes next.

And “whoever comes next” is not a theoretical worry anymore.

Frontier AI is not just another IT tool. In a national security context, it's a force multiplier for two things democracies should be allergic to:

1) watching everyone, and 2) killing faster.

I'm not against government using AI, period. I'm against letting the national security apparatus quietly absorb frontier models until the line between “analysis” and “targeting” is basically a rounding error.

What I'm willing to allow

There are legitimate government uses for AI that don't require turning the country into a panopticon or building an automated kill chain.

If the DoD wants AI that helps a veteran navigate benefits, or helps a clinician spot an error in a medical record, or helps defend networks against intrusion, fine. OpenAI's own pitch for its DoD pilot leans heavily on that kind of work: healthcare access, acquisition data, and “proactive cyber defense.” (OpenAI for Government)

I'm not naive. The military is going to use advanced technology. But we can still draw lines.

The lines I won't cross

Two red lines should be non-negotiable in a democracy.

First: no mass domestic surveillance.

Not “with guardrails.” Not “only for serious crimes.” Not “only with a policy memo.” A flat no.

Because AI changes the scale. Anthropic makes the argument plainly: even if the government is piecing together data it can legally purchase, powerful AI can fuse “scattered” information (movement, browsing, associations) into a comprehensive picture of someone's life, automatically and at massive scale. Anthropic links to a declassified ODNI report that acknowledges privacy concerns with commercially available data.

Second: keep frontier AI out of lethal targeting.

Not just “fully autonomous weapons.” The whole pipeline.

If your model is prioritizing targets, fusing sensor feeds, flagging “suspicious” behavior, or recommending strikes, you're in the kill chain. And the “human in the loop” becomes a fig leaf. Humans under pressure rubber-stamp. When a system spits out a confident answer, it takes a lot of backbone to slow down and second-guess the machine. Most organizations don't reward that.

The creep is already happening

We've already watched this creep happen, and it's documented.

Google is the cleanest example because there's a clear paper trail.

In 2018, after the backlash over Project Maven (Google helping the Pentagon analyze drone footage), Google published AI principles that explicitly said it would not pursue weapons and would not pursue surveillance “violating internationally accepted norms.” (Google AI principles, 2018; The Verge’s Maven coverage)

Later, the DoD awarded the Joint Warfighting Cloud Capability (JWCC) contracts to AWS, Google, Microsoft, and Oracle: core cloud infrastructure “at all classification levels” from headquarters to the tactical edge. (DoD JWCC release)

Then, in 2025, Google updated its public principles. The original 2018 post now points readers to a new principles page. That page no longer includes the old “AI applications we will not pursue” list that called out weapons and surveillance. (Google principles page)

OpenAI has its own version of the story.

As of January 9, 2024, OpenAI's usage policies explicitly listed “Military and warfare” as disallowed under “high risk of physical harm.” The next day, the policies were updated (the changelog shows a January 10, 2024 update), and the explicit “military and warfare” prohibition disappears. (archived Jan 9 policy; updated Jan 10 policy)

Then the engagements follow. OpenAI partners with Anduril to integrate OpenAI software into counter-drone systems. (The Verge on the OpenAI-Anduril deal)

And the DoD awards OpenAI Public Sector LLC a $200 million agreement to develop “prototype frontier AI capabilities” for “critical national security challenges in both warfighting and enterprise domains.” Those are the government's words, not mine. (DoD contracts listing, Jun 16 2025)

Yes, OpenAI's current policy still says you can't “develop or use weapons.” But “warfighting” contracts and “counter-drone” integrations are exactly how you slide past the public's defenses: talk about safety and freedom while building the machinery anyway.

Anthropic is being unusually explicit. They say they're deployed across classified networks and used for mission-critical work, but they won't cross two lines: mass domestic surveillance and fully autonomous weapons. And the Pentagon is demanding those lines be erased.

That's the moment you should stop asking, “But what if they use it responsibly?” and start asking, “Why are they fighting so hard to keep the option open?”

“Lawful use” is not a moral standard

The law lags technology. It always has.

“Any lawful use” doesn't mean “ethical.” It doesn't mean “with oversight.” It means: if our lawyers can find a path, we're doing it. And if they can't find a path yet, we'll lobby until they can.

This is why Trump and Hegseth matter. They're accountable for the choices being made right now. The national security apparatus doesn't belong to the conscientious staffer who worries about civil liberties. It belongs to the political leadership that signs the orders. Give that machine better surveillance and faster targeting, and you don't just get “efficiency.” You get a more capable state that's harder to challenge.

What I'm asking for

I'm asking for three lines in the sand that actually hold up when the pressure hits:

1) Explicit bans in law, plus companies that will actually honor them. Congress should prohibit federal agencies from procuring or using frontier AI for mass domestic surveillance, biometric monitoring at scale, and target identification/selection for lethal force (including the upstream targeting pipeline), as well as fully autonomous weapons. And because Trump can't be trusted to do the right thing, vendors have to refuse these uses even when the government asks nicely, threatens, or waves a contract.

2) Transparency for defense AI deployments. We need a public registry, at least at the program/category level, showing what models are used, by which offices, for which categories of use, plus incident reporting. If secrecy is the default, abuse is the default.

3) Independent audits and strong whistleblower protections. Not internal “ethics reviews.” Independent technical audits, and protected channels for contractors and civil servants to report misuse without career death.

If the Pentagon truly wants AI for the boring, legitimate stuff (paperwork, logistics, and defending networks from intrusion), it should be able to accept those conditions without flinching.

If it can't, it's telling you exactly what it wants.