> daily_signal(2026_05_25)

The Pope published the Catholic Church's first moral doctrine on AI today. He chose the 135th anniversary of the encyclical that gave workers the right to organize. Anthropic's co-founder was on stage.

PickBits Daily Signal · Sunday, May 25, 2026

This week traced who gets to write the rules for AI and who pays when those rules are missing. Today the Catholic Church answered the moral question with a formal document, choosing the exact anniversary of the encyclical that spent 135 years shaping labor law around the world. On the same morning: the White House quietly disclosed $9B in surveillance AI for spy agencies, New York's top fiscal officer put a 110,000-job number on what the city has not prepared for, Ohio taxpayers found out their governor killed a legislative fix to a $1.6B data-center tax giveaway, Meta launched an AI-powered Reddit competitor without a press release, and a Chinese lab permanently locked in the cheapest AI inference prices in the world.

Today the Vatican published the first Catholic moral framework for AI on the anniversary of the encyclical that built modern labor law, the White House quietly cleared $9B for surveillance AI, and the week's remaining news confirmed that the cost of this transition is landing on workers, taxpayers, and communities that nobody asked.

1. The Catholic Church published its first moral doctrine on AI today. The Pope chose the exact anniversary of the encyclical that built the modern labor rights movement.

Pope Leo XIV signed Magnifica Humanitas on May 15 and published it today at the Vatican's Synod Hall in Rome. The document, whose title means "magnificent humanity" in Latin, is the Catholic Church's first papal encyclical dedicated to artificial intelligence. It addresses what the Church calls "the protection of the human person in the age of artificial intelligence": worker rights, the ethics of military AI, and digital surveillance tools the document says strip workers of dignity. Presenting the encyclical alongside Leo XIV were Cardinal Victor Manuel Fernandez, prefect of the Dicastery for the Doctrine of the Faith, Cardinal Michael Czerny, theologian Anna Rowlands of Durham University, and Chris Olah, a co-founder of the AI research company Anthropic.

The signing date of May 15 is not coincidental. Pope Leo XIII signed Rerum Novarum on May 15, 1891, exactly 135 years before this document. Rerum Novarum established Catholic social teaching on the rights of workers, just wages, and human dignity in the industrial era. It became the moral foundation for labor law across much of the 20th century, including in countries with no Catholic majority. Magnifica Humanitas makes the same framing explicit: AI is to the 21st century what industrial capitalism was to the 19th, and the Church is publishing its moral framework at the beginning of the transformation. The encyclical calls for "shared standards of social justice" in AI development, frames AI design choices (not just AI use) as moral acts, warns against autonomous military AI that removes humans from lethal decisions, and addresses employer use of AI to surveil and score workers.

vaticannews.va: Pope Leo XIV's first encyclical Magnifica humanitas to be published May 25

2. The White House secretly approved $9B for US spy agencies to buy NVIDIA Grace Blackwell chips for classified AI. Congress still has to vote to fund it.

The White House signed off on a $9 billion secret funding request on May 22, approving a plan to buy NVIDIA Grace Blackwell superchips for the CIA, NSA, and other US intelligence agencies. The funding covers classified data centers capable of running Grace Blackwell hardware, which requires high-density electrical power and liquid cooling that most existing intelligence facility buildings cannot support. The AI systems built on this infrastructure are designed to process intelligence intercepts, signals data, and other classified inputs at scales that current agency capacity cannot handle. The White House did not announce the approval publicly; it was first reported by The Philadelphia Inquirer on May 23.

The $9 billion still requires a congressional appropriations vote. While waiting for that, the administration separately reprogrammed $800 million in existing funds for faster acquisition of computing capacity. Internal intelligence community warnings since early 2026 described a chip shortage causing US spy agencies to fall behind adversaries in testing and deploying AI tools for espionage work. Amazon Web Services is separately expanding its government cloud to IC-eligible zones with a $50 billion infrastructure upgrade. The Grace Blackwell chip is the same silicon NVIDIA sells to commercial cloud providers; this classified procurement competes directly with commercial orders for the same supply.

inquirer.com: White House approves $9B for spy agencies to catch up on AI

3. NYC's top fiscal officer put a number on AI's impact on 110,000 New York City jobs. He called the city "sleepwalking into the age of AI."

NYC Comptroller Mark Levine published a fiscal analysis on May 21 projecting AI's impact on the New York City economy. The report models five scenarios. In the worst-case, AI eliminates roughly 110,000 private-sector jobs by 2027. In the pessimistic "AI Falls Flat" scenario, the city loses 52,500 jobs as early as this year. Levine's framing was direct: "Despite the impossibly high stakes for us as we stare down the radical transformation ahead, New York City is doing almost nothing to prepare. It is sleepwalking into the age of AI."

The report's concern is both labor and fiscal. NYC's municipal budget depends heavily on income and business tax receipts from finance, professional services, and tech: three of the sectors most exposed to AI displacement. Levine called for expanding the city's rainy-day fund to 16 percent of annual tax revenue and hardening city infrastructure against AI-powered cyberattacks on water and transit systems. New York City is home to one in eight American jobs across supply chains. The comptroller's argument is that the city's fiscal solvency is directly tied to whether workers keep their jobs, and AI is now a material fiscal risk the city has no plan to address.

comptroller.nyc.gov: AI and New York City's Fiscal Future

4. Ohio's data-center tax break cost taxpayers $1.6B in 2025, eleven times the original estimate. The governor vetoed the fix that would have paid for income-tax cuts.

Ohio exempts data center developers from the state's 5.75 percent sales tax on equipment, for facilities costing $100 million or more to build, for up to 15 years. When the exemption was created, the Ohio Department of Taxation estimated the annual cost at $136 million. In 2025, data center operators claimed $1.6 billion in tax benefits under the exemption: eleven times the original estimate. In 2024, the cost was already $555 million, four times the forecast. The AI-driven surge in Ohio data center construction turned a modest economic development incentive into the state's single largest tax expenditure.

Ohio's Republican-controlled legislature voted to repeal the exemption and redirect the revenue to a new round of income-tax cuts. Governor Mike DeWine vetoed the repeal, arguing that ending the exemption would push data center developers to other states. House Speaker Matt Huffman has said he wants to muster the three-fifths supermajority needed to override the veto but says he lacks the votes. The Ohio House has formed a special legislative committee on data centers that will take testimony from the Department of Taxation on how a $136 million estimate became $1.6 billion in two years. The total $1.6B cost represents roughly $136 in lost state revenue for every Ohio household.

signalohio.org: Ohio data center tax break cost $1.4 billion more than expected in 2025

5. Meta quietly launched a Reddit-rival app called Forum with a built-in AI assistant. Reddit stock fell more than 5% on launch day.

Meta released Forum in the US App Store on May 22 with no press release and no launch event. Forum is a standalone iOS application built on Facebook Groups, presenting group content in a thread-style feed separate from the main Facebook interface. It includes two AI features: Ask, a chatbot that synthesizes answers from across a user's Facebook Groups, and an AI moderation assistant for group administrators. It also allows users to post under nicknames rather than their full Facebook identity: a deliberate design choice to replicate one of Reddit's structural advantages, pseudonymous community participation.

Social media analyst Matt Navarra identified and reported the launch. Financial analysts at Truist flagged Forum as a direct competitive threat and cited gradual erosion risk among casual Reddit users with lower platform loyalty. Reddit stock closed down more than 5% on May 22. Reddit had reported $625 million in advertising revenue in Q1 2026, up 74% year over year. Meta's strategic logic rests on the AI Ask feature: if a user can ask a question and receive a synthesized answer from thousands of group posts instead of scrolling threads, the effort gap between Reddit and a Meta alternative narrows in a way that a simple feed redesign never did.

techcrunch.com: Meta quietly launches a new Reddit-like app called Forum

6. DeepSeek made its 75% V4-Pro price cut permanent. Output tokens are now 34.5 times cheaper than GPT-5.5. The AI price war just locked in a new floor. [Returning]

DeepSeek V4 was last covered April 27 when a temporary 75% promotional discount launched. New development: the discount is now permanent.

DeepSeek announced on May 23 that the 75 percent discount on V4-Pro API pricing, which had been scheduled to expire May 31, is now permanent. V4-Pro output tokens are priced at $0.87 per million, down from $3.48. Input tokens are $0.003625 per million, down from $0.0145. GPT-5.5 output tokens are priced at roughly $30 per million at standard pricing. DeepSeek V4-Pro output is now 34.5 times cheaper than GPT-5.5 with a one-million-token context window. The V4 series is optimized for Huawei Ascend 950 chips rather than NVIDIA processors, meaning DeepSeek's production capacity scales independently of US export control restrictions on NVIDIA hardware.

The lock-in establishes a new permanent floor for the global AI inference market. The 75% discount was widely read as a promotional window that would revert to standard pricing on June 1; making it permanent signals that DeepSeek is prioritizing market share over per-unit margin. Any AI product or application built on API calls from a frontier model now has a comparison price against which every other vendor's pricing is measured. The enterprises and developers that cannot use DeepSeek for security or procurement reasons are paying a 10x to 34x premium for that constraint, and that premium is now a permanent line item rather than a temporary price gap.

the-decoder.com: Deepseek makes its 75 percent discount permanent, pricing output tokens at least 34x below GPT-5.5

» What to watch this week

• Magnifica Humanitas response from AI companies. Leo XIII's Rerum Novarum was opposed by industrial employers and took decades to translate into law. Watch whether Anthropic, Google, Microsoft, or Meta issues a formal response to the encyclical this week, and whether any company invokes it in a product policy or labor context. Leo XIV co-presenting with an Anthropic co-founder is an implicit collaboration signal; whether Anthropic explicitly endorses the document's worker-protection and military-AI provisions is the open question.
• Congressional appropriations vote on the $9B IC AI funding. The White House approved the request; Congress still has to fund it. Watch House and Senate intelligence and appropriations committee schedules for a markup or hearing. EFF and ACLU surveillance newsletters are the tracking sources if no public hearing is announced.
• NYC Council response to Levine's 110K report. The comptroller's analysis is advisory; legislation or a budget amendment turns it into enforcement. Watch the NYC Council finance committee calendar for a scheduled hearing.
• Ohio veto override vote. Speaker Huffman needs a three-fifths supermajority and says he does not have it. The special data-center committee testimony from the Department of Taxation is the next scheduled event. Watch the Ohio House calendar for a hearing date.
• Meta Forum expansion. Forum launched iOS-only in the US with no announcement. A Google Play listing or Android build is the signal that this is a full product launch rather than a limited test. Watch Meta's newsroom.
• Samsung NSEU bonus vote close. Voting closes Wednesday May 27 at 10 a.m. Korea time on the $26B AI employee bonus pool covered Saturday. Turnout is the key number: the deal requires more than half of roughly 70,000 eligible members to cast ballots before a majority-of-voters threshold applies.

Tomorrow's signal lands here.

PickBits Daily Signal is a working brief by Mark Pickering. One digest, every day, until I have enough subscribers that I do not have to go back to a corporate job. Twenty years shipping apps and workflow automation; the last stretch in agentic workflows on ServiceNow. I know what the memo looks like before it goes out. If a friend forwarded this to you, you can subscribe at pickbitsai.substack.com.