Built for rumors, fast-moving news, and claims that spread faster than the proof.

Latest fact-checks

Supported

SSI Targets August For First Model Release

Claim checked

“SITUATION DETECTED: Ilya Sutskever's Safe Superintelligence is reportedly planning to launch its highly anticipated first model in August 2026, per Gavin Baker in an interview with Patrick O'Shaughnessy.”

The checked claim is that Ilya Sutskever's Safe Superintelligence is reportedly planning to launch its first model in August 2026, per Gavin Baker in an interview with Patrick O'Shaughnessy. The claim is supported: the O'Shaughnessy–Baker conversation did take place in May 2026 (confirmed by the Invest Like The Best podcast and O'Shaughnessy's LinkedIn post), and social media...

Supported

Z.AI Builds Massive Chinese-Chip Data Center

Claim checked

“http://Z.AI has completed construction of a giant data center that houses only Chinese-made chips, a big step in Beijing’s efforts to replace restricted Nvidia silicon for future AI development https://bloomberg.com/news/articles/2026-07-20/z-ai-completes-giant-data-center-with-chinese-chips-to-train-ai?taid=6a5e6806dd3d1b00013b041c&utm_campaign=trueanthem&utm_content=business&utm_medium=social&utm_source=twitter…”

The checked claim states that Z.AI has completed construction of a giant data center housing only Chinese-made chips, framing it as a step in Beijing's effort to replace restricted Nvidia silicon. The claim is supported. Bloomberg's own July 20, 2026 article confirms that Z.AI (formerly Zhipu) has finished building a roughly 1-gigawatt facility and begun partial operations,...

Supported

Google's Gemini 3.5 Pro Hits Another Delay

Claim checked

“Gemini 3.5 Pro is delayed”

The checked claim is that Gemini 3.5 Pro is delayed, posted by X user @zephyr_z9. Multiple current, reliable sources confirm the claim. Business Insider reported on June 24, 2026, that Google pushed the release from June to July, and as of today, outlets including 9to5Google and TechTimes report the model has now missed its July 17 target as well, with Google still citing...

Supported

Tesla's Robotaxi Service Reaches Miami

Claim checked

“Unsupervised @Robotaxi in Miami!!!”

The checked claim — a brief X post declaring "Unsupervised @Robotaxi in Miami!!!" — is supported by available evidence. A news report dated July 3, 2026 confirms that Tesla officially launched its robotaxi service in Miami the same day, quoting Tesla's official robotaxi account announcing "Robotaxi now available in Miami." The "unsupervised" framing in the post is consistent...

Misleading

SpaceX-T-Mobile Deal Speculation Resurfaces Online

Claim checked

“BREAKING: SpaceX could acquire T-Mobile”

A widely shared X post labeled "BREAKING" claims SpaceX could acquire T-Mobile. While financial analysts have indeed floated this possibility, no acquisition has been announced or is in progress. The post omits that the claim is analyst speculation, that T-Mobile's parent company has long resisted selling, and that the speculation has circulated for weeks.

Supported

Anthropic Staff Confirm Fable 5 Still Offline After Government Suspension

Claim checked

“Seeing chatter about Fable 5 being accessible - can say categorically this is false, we are not serving any Fable / Mythos traffic Looking into possibility of UI bug on front-end (e.g. based on historical context), but also very real chance it’s just people shitposting…”

Anthropic staff members are correct: the company is not currently serving Fable 5 or Mythos 5 traffic. Posts from Amol Avasare and Sam McAllister on June 25, 2026, saying exactly that are consistent with the company's mid-June decision to disable both models following a U.S. government export control directive. Independent reporting from CNBC and Anthropic's own public...

Unverified

GPT-5.6 Delay Claim Rests on Single Unverified Post

Claim checked

“GPT-5.6 has been delayed and will no longer release this week. New target is ~mid-July.”

A claim that OpenAI's GPT-5.6 has been delayed from a late-June launch to mid-July 2026 is unverified. The assertion originates from a single post on X by an account with no demonstrated track record of confirmed leaks, and no official OpenAI statement, press coverage, or independent reporting corroborates the delay. Reporting from mid-June 2026 described GPT-5.6 as tracking...

Supported

Betting Markets Cool on GPT-5.6 Launch

Claim checked

“GPT-5.6 release expectations continue to cool Prediction markets now heavily favor "not released by June 28" over a launch this week”

The claim is supported. Polymarket data shows the prediction market now heavily favors GPT-5.6 not launching publicly by June 28, with the probability sitting at 78% as of June 21, 2026. That figure rose sharply from earlier in the month, when markets had priced a June 22–28 release window at around 83%. The reversal reflects a clear cooling of release expectations, consistent...

Supported

Claim About Mythos Breaching NSA Systems in Hours

Claim checked

“Holy Sh*t: that changes the whole Fable 5 story completely: On June 11, the very same day Amazon reportedly uncovered the jailbreak, “Mythos” allegedly breached almost all classified systems belonging to the NSA and U.S. Cyber Command, not over the course of weeks, but within hours. "On June 11th Mark Warner, the vice-chair of the Senate Intelligence Committee, said that General Joshua Rudd, who leads the National Security Agency and the Pentagon’s Cyber Command, had told him that Mythos “broke...”

The core claim is supported by available reporting. According to coverage from The Economist, NSA chief General Joshua Rudd told Senator Mark Warner that Anthropic's AI model Mythos "broke into almost all of our classified systems, not in weeks, but in hours." The quote, the attribution to Rudd, and the involvement of Warner as vice-chair of the Senate Intelligence Committee...

Misleading

Amodei's Bankruptcy Quote Raises Questions

Claim checked

“Anthropic CEO: "If my revenue is not $1 trillion, even $800 billion, there's no force on earth, no hedge on earth, that could stop me from going bankrupt."”

The quote attributed to Anthropic CEO Dario Amodei is substantially accurate but presented without crucial context that changes its meaning. The quote is from a February 2026 podcast interview where Amodei discussed the risk of buying massive amounts of compute. The viral post omits his conditional statement: "if I buy that much compute." Without this phrase, the quote...

Supported

EU AI Gigafactory Plan Slips as Funding, Bidders Dwindle

Claim checked

“In April 2025, the EU called for expressions of interest in building four misleadingly-named "AI gigafactories" (only about 100 MW each, barely a dent). They received 76 applications. Now, over a year later, even this supremely unambitious plan is being heavily scaled back.”

The claim is supported. The EU opened a non-binding call for expressions of interest in AI gigafactories on 9 April 2025 and received 76 submissions covering 60 sites across 16 member states, with respondents proposing facilities powered by roughly 100,000 advanced AI chips each. More than a year later, the plan is being substantially scaled back: only two of the five planned...

Supported

SpaceX to Buy Cursor Parent Anysphere in $60 Billion Deal

Claim checked

“BREAKING - SpaceX to acquire AI coding company Cursor for $60 billion: SEC filing”

The claim that SpaceX is acquiring AI coding company Cursor for $60 billion is supported by an SEC filing and multiple major outlets. A filing with the U.S. Securities and Exchange Commission confirms that SpaceX has entered into a merger agreement with Anysphere, the parent company of Cursor, under which Cursor will become a wholly owned subsidiary. The deal, announced just...

Unverified

Brain-Scale AI Claims Outpace Verified Evidence

Claim checked

“By 2027, AI infrastructure could finally support models approaching brain-scale size, around 100-150T parameters today's frontier models are estimated to be far smaller, roughly 5-6T”

A widely shared post claims that by 2027, AI infrastructure could support models approaching brain-scale size of 100–150 trillion parameters, while today's frontier models sit at roughly 5–6 trillion. The brain-scale framing echoes long-standing industry rhetoric, but the specific parameter counts and the 2027 timeline are not supported by available evidence. Major labs do not...

Supported

Luo Fuli Leads Xiaomi's MiMo After DeepSeek Stint

Claim checked

“Xiaomi AI's large-scale model "MiMo" project lead, Luo Fuli. Apparently from DeepSeek, but she's incredibly young. In China, it feels like there are quite a few women making waves in the AI field.”

The claim is supported. Reporting from the South China Morning Post, Wikipedia, and Baidu Baike confirms that Luo Fuli leads Xiaomi's MiMo large model team and previously worked at DeepSeek as a key developer of the DeepSeek-V2 model. She was born in 1995, making her around 30 at the time of the post, and is widely described in Chinese media as a young prodigy in the AI field.

Supported

Karpathy Barred From Anthropic's Top Model Over Citizenship

Claim checked

“JUST IN: Andrej Karpathy, a top AI scientist at Anthropic, is reportedly barred from accessing the company’s most advanced AI model because he is not a U.S. citizen.”

The claim is supported. Reporting from XDA Developers and a statement from Anthropic confirm that the U.S. government issued an export-control directive requiring Anthropic to suspend access to its Fable 5 and Mythos 5 models by any foreign national, including foreign-national Anthropic employees. Andrej Karpathy, who joined Anthropic's pretraining team in May 2026, is...

Misleading

Musk's trillion-dollar rise runs on private stakes, not just government cash

Claim checked

“Elon Musk is only a trillionaire because of government contracts and loans. He's not an entrepreneur — he's a nepo baby mooching off American taxpayers.”

Elon Musk crossed $1 trillion in net worth on June 12, 2026, after the SpaceX initial public offering, with his fortune tied primarily to his ownership stakes in SpaceX and Tesla, according to Forbes and Bloomberg tracking. Reporting does show Musk's companies have received roughly $38 billion in U.S. government contracts, loans, subsidies, and tax credits since 2003, with...

Mostly true

Anthropic Revenue Nears $50 Billion

Claim checked

“Anthropic’s annualized sales have climbed to nearly $50 billion, giving it new leverage over business customers. Some of those same customers now fear they are paying Anthropic to build products that compete with them. Full story:”

The claim that Anthropic's annualized sales have climbed to nearly $50 billion is mostly true. Multiple reports confirm Anthropic's annualized revenue reached $47 billion in May 2026, which is indeed "nearly $50 billion." The claim about customers fearing they are paying Anthropic to build competing products is also supported by The Information's reporting, though the full...

Mostly true

AI Agents Struggle on New Professional Work Benchmark

Claim checked

“A new benchmark called Agents’ Last Exam (ALE) is testing whether AI agents are truly ready for real digital labor-market work. The benchmark includes more than 1,500 expert-sourced tasks. The tasks span 55 occupations. Models tested include Fable 5, GPT-5.5, Composer 2.5, and other frontier agent systems. The benchmark was created by researchers who previously worked on major evals like MMLU, MATH, CyberGym, and ExploitGym. Current agents can solve some real professional tasks. But on ALE’s...”

The claim is mostly true. The Agents' Last Exam (ALE) is a real benchmark designed to test AI agents on professional workflows, and it does include 1,500+ tasks across 55 occupations. Models like GPT-5.5, Fable 5, and Composer 2.5 are indeed tested on the leaderboard. However, the claim that every tested frontier agent scored 0% on the hardest tier is slightly misleading....

Unverified

Claude Mythos 5 Launch Rumors Swirl, But No Official Date

Claim checked

“Anthropic is preparing to release Claude Mythos 5 tomorrow The first 10T+ parameter AI model is being prepared, marking a new era in artificial intelligence Welcome to the Mythos era”

The claim that Anthropic is preparing to release Claude Mythos 5 tomorrow as the first 10T+ parameter AI model is unverified. While Claude Mythos is a real model developed by Anthropic, the specific details in this viral post—a "Mythos 5" product name, a June 10 release date, and a confirmed 10-trillion-parameter count—lack any official backing from the company. According to...

Supported

Siri AI Blocked in EU Over DMA

Claim checked

“Wait Siri AI not supported in EU? What the fuck? You really won't make it if you're in EU FUCKING HELL”

A viral post claims Siri AI is not supported in the EU, expressing frustration that Apple's AI features are unavailable to European users. Apple's own June 2026 press release announcing the new Siri AI confirms that Siri AI will not be available initially in the EU in iOS and iPadOS, while Mac, Apple Watch, and Apple Vision Pro users in the EU will be able to access it. The...