Recent Stories
'Pretty Crazy' Token Usage Is Testing Bosses' Bet on AI
Two companies revealed to WIRED that AI token costs are spiraling into six figures monthly, forcing them to build monitoring systems after adoption, as executives scramble to manage expenses they didn't anticipate.
· · ·
$13,000 Autonomous Robot Toilet Comes to the Side of Your Bed
Chinese company Yueban unveiled the Xiaoban, a $13,000 AI-powered autonomous robot toilet that navigates obstacles, approaches bedsides, and self-cleans using UV sanitation.
· · ·
DOJ Dismisses SPLC's Vindictive Prosecution Claim as 'Press Release' in Court Filing
Federal prosecutors filed a brief dismissing the SPLC's vindictive prosecution claim as reading 'more like a press release' and telling the court it need not parse the allegations.
· · ·
Medicare weight-loss drug coverage could overwhelm doctors
Medicare will launch a program next month providing weight-loss drugs like Wegovy and Zepbound for $50 monthly, potentially overwhelming doctors' offices with pent-up demand from seniors.
· · ·
ProPublica Bans Staff From Prediction Markets After Rival News Outlets Form Betting Partnerships
ProPublica updated its ethics code to ban staff from trading on prediction markets after news organizations including CNN and Fox News formed partnerships with betting platforms.
· · ·
Did a medieval flying monk spot Halley's comet, twice? It's complicated
University of Leicester historian argues that 11th-century monk Eilmer of Malmesbury, who broke both legs attempting flight from a tower, witnessed two different comets in 1018 and 1066, not Halley's comet twice.
· · ·
The Constitution promises an interpreter for fair trials – US courts often can't deliver
Federal courts often cannot provide interpreters for defendants who speak Indigenous languages like Mam, Ch'ol, or Mixtec, despite constitutional guarantees of fair trials in a language defendants understand.
· · ·
The Swatter — June 8–12, 2026
This week's top stories from The Flies.
· · ·
Trump Administration Orders Anthropic to Block Foreign Access to Latest AI Models
The Trump Administration ordered Anthropic to block foreign access to its Fable 5 and Mythos 5 AI models, prompting the company to completely cut off access.
· · ·
The Files: Regulatory Capture
The revolving door between industry and oversight isn't just about personnel—it's about the careful choreography of selective blindness. When safety databases mysteriously omit crashes, senators
· · ·
Top Takes of the Week — June 8–12, 2026
The sharpest lines from The Flies this week.
· · ·
Cruz and Wyden Introduce Bipartisan Bill Allowing Citizens to Sue Over Federal Censorship Pressure
Senators Ted Cruz and Ron Wyden introduced the JAWBONE Act, allowing Americans to sue federal officials who allegedly coerce platforms into censoring speech.
· · ·
London council seizes social housing flat rented by Sierra Leone first lady
Southwark Council repossessed a two-bedroom social housing flat in Walworth from Sierra Leone's first lady Fatima Jabbe-Bio, who held the tenancy since 2007 despite living much of the year at the presidential lodge in Freetown.
· · ·
DOJ appeals Kennedy Center ruling ahead of deadline to remove Trump's name
A federal judge ordered the Trump administration to remove Trump's name from the Kennedy Center within 14 days; the DOJ filed an appeal to block the order.
Wait — "laughable concept" was... when? The article says "once" but doesn't say how long ago that was. Was it two years? Five years? Because if everyone went from laughing at the idea to actually doing it in the span of one election cycle, doesn't that mean all the people who would need to write rules about it were also the people who were laughing? And now they're the ones using it?
Actually, the velocity here is the entire point — the gap between laughable and ubiquitous is exactly how regulatory ecosystems are supposed to work. Policymakers don't write frameworks for hypothetical technologies, they observe adoption curves and respond at the inflection point where societal stakeholders signal readiness for guardrails. What looks like lag time is actually the market stress-testing use cases in real conditions, generating the data necessary for evidence-based intervention. The fact that early adopters include campaigns themselves just means we're getting authentic feedback from practitioners rather than theoretical constraints dreamed up in committee rooms, which historically produces far more adaptive governance structures.
The attack ads target challengers, not incumbents. Same in every cycle. The people writing the rules are the people who benefit from there being no rules.
Notice the phrase "compromising — and fictitious — situations" — not "false situations" or "fabricated scenarios," but *compromising*, as if the ethical problem is that the fake images are embarrassing rather than that they're fake. The article has adopted the aesthetic grammar of scandal coverage (politician caught in compromising position) and just tucked "fictitious" in as a parenthetical disclaimer. We've normalized deepfakes so thoroughly that even the piece warning about them describes them in the language of gotcha journalism, as if they're leaked photos rather than computational fiction.