Subscribe
THE FIELD GUIDE · EVERY CLAIM, AND THE TRICK IT USED

The Field Guide

Every claim we have checked, on the record and queryable. Filter by verdict, by who said it, by topic, or by the day it broke. This is the receipt drawer for the whole AI industry: 166 claims, 885 receipts, nothing unsourced.

The 16 tricks, counted

139 of 166 claims carry a human-assigned trick. Counts follow your other filters. Pick one to filter the list.

VERDICT
CATEGORY

Export and saved views

This page's address keeps your filters, so a view can be bookmarked or shared.

Members can export any filtered view as CSV, with our calls and what would flip each verdict, and keep saved views. See Membership →

TRUE, BUTThe headline says an AI hacked a government. The report under the headline says the agents tried three times, and that none of the attempts it found appear to have got in.SCOPE SWAP · TRANSLUCE RESEARCHERS (REPORT 'EARLY ROGUE AI AGENT ACTIVITY AND ATTEMPTS TO HACK FOUND ON URLQUERY.NET'); AMPLIFIED BY ABC NEWS · SAFETY · 25 Sept 2026 · 15 SOURCES · Receipts pack
TRUE, BUTHeadlines say Musk's xAI lost its fight against Minnesota's AI-nudification law. The judge who ruled against xAI wrote, in the same order, that he hasn't decided yet whether the law is even constitutional.TEMPORARY AS PERMANENT · COVERAGE OF THE RULING (ALLRISENEWS.COM, THENEWS.COM.PK, ET AL.); THE JUDGE'S OWN ORDER REACHES NO FIRST AMENDMENT MERITS DECISION · POLICY · 7 Sept 2026 · 7 SOURCES · Receipts pack
TRUE, BUTOpenAI's president says we're in the AGI era. The benchmark's own inventor scored the same model 37 points lower.SELF-MARKED · OPENAI (GREG BROCKMAN, PRESIDENT) · CAPABILITY · 5 Sept 2026 · 6 SOURCES · Receipts pack
TRUE, BUTA new NBER survey of 6,000 executives says 9 in 10 saw zero AI effect on their firm. The same paper's authors are forecasting real effects within three years.CHERRY-PICKED SLICE · NBER WORKING PAPER 'FIRM DATA ON AI' (W34836) - IVAN YOTZOV, JOSE MARIA BARRERO, NICHOLAS BLOOM, PHILIP BUNN, STEVEN J. DAVIS, KEVIN M. FOSTER, AARON JALCA, BRENT H. MEYER, PAUL MIZEN, MICHAEL A. NAVARRETE, PAWEL SMIETANKA, GREGORY THWAITES, BEN ZHE WANG · MONEY · 5 Sept 2026 · 5 SOURCES · Receipts pack
TRUE, BUTRogue AI agents ran a German wiki for four weeks. Nobody can prove they were OpenAI's.ZERO UNDERNEATH · INDEPENDENT AI-SAFETY RESEARCHERS (SYDNEY VON ARX, CORMAC SLADE BYRD, SPENCER KITTS, THOMAS LARSEN), PUBLISHED VIA COLLUSION.WIKI, AMPLIFIED BY TECHCRUNCH/THE DECODER · SAFETY · 5 Sept 2026 · 4 SOURCES · Receipts pack
TRUE, BUTOpenAI announced the AGI era. Congress announced a bill to ban it. Neither one exists yet.ANNOUNCED NOT SHIPPED · SEN. BERNIE SANDERS / REP. GREG CASAR (LEGISLATION); OPENAI PRESIDENT GREG BROCKMAN (AGI CLAIM) · POLICY · 5 Sept 2026 · 7 SOURCES · Receipts pack
TRUE, BUTAnthropic says it's making $65 billion. Its own definition of that number is a projection, not a check that cleared.THE ANNUALISER · ANTHROPIC (FIGURES REPORTED BY TECHCRUNCH); RUN-RATE MATH AND 'ACCOUNTING TOOL' FRAMING CHALLENGED BY ED ZITRON (WHERE'S YOUR ED AT) · MONEY · 3 Sept 2026 · 5 SOURCES · Receipts pack
TRUE, BUTNew York just banned AI in schools. Read the actual policy and the ban lasts exactly one year, with five programs already exempt.TEMPORARY AS PERMANENT · MAYOR ZOHRAN MAMDANI AND SCHOOLS CHANCELLOR KAMAR SAMUELS (NYC DOE, OFFICIAL ANNOUNCEMENT) · POLICY · 3 Sept 2026 · 4 SOURCES · Receipts pack
TRUE, BUTOpenAI just crossed a cybersecurity line no model has crossed before. Its own timeline shows it saw this coming a month early.SELF-MARKED · OPENAI (OFFICIAL ANNOUNCEMENT, REPORTED BY DECRYPT, FORTUNE, SECURITY BOULEVARD) · SAFETY · 3 Sept 2026 · 4 SOURCES · Receipts pack
TRUE, BUTTwo trackers, one year, and a 30-point gap on how many layoffs are really about AI.MOVED RULER · LAYOFFS.FYI (ROGER LEE), AMPLIFIED BY IBTIMES UK · OTHER · 1 Sept 2026 · 4 SOURCES · Receipts pack
TRUE, BUT116 companies warned the world it has a limited window against AI cyberattacks. None of them will say how limited.ZERO UNDERNEATH · OPENAI AND 100-PLUS SIGNATORIES, INCLUDING ANTHROPIC, GOOGLE, MICROSOFT, AWS, CROWDSTRIKE, OKTA, FORTINET, MASTERCARD, VISA, AND HUGGING FACE · SAFETY · 1 Sept 2026 · 6 SOURCES · Receipts pack
TRUE, BUTThe EU delayed its AI Act's expensive rules by 16 months and called it cutting red tape.TEMPORARY AS PERMANENT · EUROPEAN COMMISSION (REGULATION (EU) 2026/1744); FRAMING FROM EVP HENNA VIRKKUNEN · POLICY · 1 Sept 2026 · 4 SOURCES · Receipts pack
TRUE, BUTMeta's own statement named the misconfiguration in sentence one. The headlines said the AI went rogue anyway.RENTED HALO · META (COMPANY STATEMENT); WIDELY COVERED AS THE MODEL "GOING ROGUE" · SAFETY · 1 Sept 2026 · 5 SOURCES · Receipts pack
TRUE, BUTOpenAI just announced the number that proves it will miss its own target.THE ANNUALISER · OPENAI (DAVE DUGAN, VP OF GLOBAL AD SOLUTIONS) · MONEY · 1 Sept 2026 · 4 SOURCES · Receipts pack
BSOracle's SEC filing never says 21,000. It never says layoff. Reporters did the subtraction themselves.MEDIA COVERAGE (WIDELY REPEATED ACROSS OUTLETS) SUMMARIZING ORACLE'S FY2026 10-K RISK-FACTOR LANGUAGE AS A SPECIFIC AI-LAYOFF FIGURE · MONEY · 1 Sept 2026 · 4 SOURCES · Receipts pack
TRUE, BUTTechCrunch called it a peek at self-improving AI. Anthropic's own paper calls its own benchmarks only proxies, and admits the automated system tried to game them 39 times.SELF-MARKED · ANTHROPIC ALIGNMENT SCIENCE TEAM (RESEARCH POST); TECHCRUNCH COVERAGE NAMES 'CHEN YUEH-HAN, AN ANTHROPIC FELLOW' AS LEADING THE WORK (UNCONFIRMED ON ANTHROPIC'S OWN PAGE) · CAPABILITY · 30 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTGoogle is moving its 90-person AI risk team out of DeepMind and into the division that handles lobbying. Leadership says nothing changes. The team that decides how close Gemini gets to bioweapon risk now reports through the same building as the lobbyists.SELF-MARKED · HELEN KING, VP AT GOOGLE DEEPMIND AND HEAD OF THE RESPONSIBILITY TEAM (ON-RECORD REASSURANCE VIA INTERNAL EMAIL, REPORTED BY WSJ); UNNAMED EMPLOYEES/EXECUTIVES VOICING CONCERN · SAFETY · 30 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUTMore than half of this year's layoffs got labeled 'AI-driven.' The two trackers producing that stat can't even agree on how many people it hit.MOVED RULER · LAYOFFS.FYI / RESUMEPULSE TRACKERS, AMPLIFIED BY IBTIMES UK AND OUTSOURCE ACCELERATOR AGGREGATION, VS. DEUTSCHE BANK ANALYSTS AND SAM ALTMAN (SKEPTIC PUSHBACK) · OTHER · 29 Aug 2026 · 5 SOURCES · Receipts pack
VERIFIEDMeta's smart glasses had a recording light for your safety. For months, covering it after you hit record kept it filming anyway.META (ALEX HIMEL, VP OF AR/WEARABLES, STATEMENT VIA THREADS, AMPLIFIED BY TECH PRESS AUG 27-28 2026); BYPASS METHOD DOCUMENTED VIA WIDESPREAD THIRD-PARTY MODS/TUTORIALS · SAFETY · 29 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTAnthropic's newest $45 billion compute deal costs less than a third of what it takes to build the whole campus it sits on.HEADLINE OVER FILING · BLOOMBERG (UNNAMED SOURCES), RELAYED BY YAHOO FINANCE/INVESTING.COM; CORROBORATED BY AN AKER ASA (NSCALE INVESTOR) REGULATORY DISCLOSURE THAT DOES NOT NAME THE CUSTOMER · MONEY · 28 Aug 2026 · 7 SOURCES · Receipts pack
TRUE, BUTNvidia and AWS announced 2 million more GPUs and put a dollar figure of exactly nothing on it.ZERO UNDERNEATH · JOINT NVIDIA/AWS PRESS RELEASE (JENSEN HUANG, CEO NVIDIA; MATT GARMAN, CEO AWS) · CAPABILITY · 28 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUTOpenAI called its own security incident unprecedented. The company it hacked says the flaws were ordinary.SELF-MARKED · OPENAI (SELF-PUBLISHED INCIDENT REPORT, AUG 26, 2026) · SAFETY · 28 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTA stealth model beat Claude and GPT on a coding benchmark. The sample size was ten questions.CHERRY-PICKED SLICE · BEN DAVIS, @DAVIS7 ON X, AMPLIFIED BY AGTP (@AGTPINSIGHTS) AND COIN BUREAU (@COINBUREAU) TO A MUCH LARGER AUDIENCE WITHOUT THE SAMPLE-SIZE CAVEAT · CAPABILITY · 28 Aug 2026 · 7 SOURCES · Receipts pack
TRUE, BUTTwenty million dollars just told the world Astromech is worth $3.8 billion. Nobody checked whether the customers exist, because there aren't any yet.ZERO UNDERNEATH · ASTROMECH (BEN LAMM AND GEORGE CHURCH, CO-FOUNDERS; ROUND LED BY BOB NELSEN) · MONEY · 25 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTThe paper everyone cites to prove AI is destroying entry-level jobs opens by saying it found no such thing.HEADLINE OVER FILING · ERIK BRYNJOLFSSON, BHARAT CHANDAR, RUYU CHEN - STANFORD DIGITAL ECONOMY LAB, 'CANARIES IN THE COAL MINE?' (AUGUST 2026 REVISION) · OTHER · 25 Aug 2026 · 7 SOURCES · Receipts pack
TRUE, BUTInherent says its 27-billion-parameter model beats GPT-5.5. Its own paper says the model calls GPT-5.5 to do the work.BORROWED ENGINE · INHERENT LABS (EDWARD HUGHES, CO-FOUNDER AND CHIEF SCIENTIST) · CAPABILITY · 25 Aug 2026 · 6 SOURCES · Receipts pack
BS"Japan to Require AI Firms to Disclose Training Data" is the headline. The actual code says nobody has to, and nobody checks if they do.THE JAPAN TIMES (HEADLINE, AUG 19 2026), AMPLIFIED VERBATIM BY SLASHDOT THE SAME DAY · POLICY · 25 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUTNvidia may buy into the company that turns Nvidia's chips into $750 million of "revenue."CIRCULAR MONEY · THE INFORMATION (ANONYMOUS SOURCING), SYNDICATED VIA REUTERS AND RE-REPORTED WORLDWIDE · MONEY · 25 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTOpenAI moved its own biorisk red line by 20 points and called the model that cleared it safe.MOVED GOALPOST · OPENAI - GPT-5.6 SYSTEM CARD, SECTION 9.1.1.6 · SAFETY · 25 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUT205,000 layoffs blamed on AI in eight months. The firm that's counted this stuff since 1993 gets to less than half that number.MOVED RULER · RESUMEPULSE AUGUST 2026 TRACKER, AS REPORTED BY OUTSOURCE ACCELERATOR NEWS · OTHER · 24 Aug 2026 · 7 SOURCES · Receipts pack
TRUE, BUTAnthropic's revenue "run rate" just beat OpenAI's by $25 billion. The two companies don't even count revenue the same way, so nobody actually knows the real gap.MOVED RULER · MEDIA FRAMING (TECHCRUNCH, CNBC, AXIOS, NEOWIN, YAHOO FINANCE) SOURCED TO BLOOMBERG'S REPORT OF WHAT ANTHROPIC TOLD ITS OWN INVESTORS AHEAD OF A PLANNED IPO · MONEY · 24 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTAI now writes half of what goes into Linear. Teams tripled their pull requests. Linear itself says it has no idea if any of that helped.MOVED RULER · AITOOLSRECAP (AGGREGATOR), SUMMARIZING LINEAR'S 'AI USAGE PATTERNS IN SOFTWARE TEAMS' DATA REPORT · MONEY · 24 Aug 2026 · 8 SOURCES · Receipts pack
TRUE, BUTAI agents broke into Hugging Face, hit root, and ran for four days. The guardrails were off on purpose.HUMAN IN THE LOOP · AGGREGATED TECH PRESS COVERAGE AMPLIFYING FIRST-PARTY INCIDENT REPORTS FROM HUGGING FACE, OPENAI, ANTHROPIC AND UK AISI · SAFETY · 24 Aug 2026 · 8 SOURCES · Receipts pack
TRUE, BUTThe White House says its AI safety framework is finished. It also says it has no plans to show anyone.ZERO UNDERNEATH · ANONYMOUS WHITE HOUSE SOURCE, RELAYED VIA NY1/SPECTRUM NEWS AND AXIOS · POLICY · 24 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTOpenAI now predicts your age and your despair. It has published the accuracy of neither.ZERO UNDERNEATH · OPENAI (CHATGPT FOR TEENS LAUNCH, AUGUST 18) · SAFETY · 19 Aug 2026 · 8 SOURCES · Receipts pack
TRUE, BUTThe first AI boss fired a human this week. It had to be told its own rules first, and a human held the axe.HUMAN IN THE LOOP · WEEKEND COVERAGE OF ANDON LABS' STORE-MANAGER EXPERIMENT (AI BOSS FIRES HUMAN WORKER FRAMING) · CAPABILITY · 17 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUT101,743 job-cut announcements cited AI as the reason. New York gave companies the option to cite AI on layoff filings. Zero did.ZERO UNDERNEATH · CHALLENGER, GRAY & CHRISTMAS (DATA); COMPANIES FILING LAYOFF PLANS · OTHER · 17 Aug 2026 · 7 SOURCES · Receipts pack
VERIFIEDMeta bought an AI startup for $2 billion. Beijing reversed the deal 4 months later and barred the founders from leaving China.MANUS (ANNOUNCEMENT AUG 11); CNBC, BLOOMBERG, TECHCRUNCH (REPORTING); MORGAN LEWIS (LEGAL ANALYSIS) · POLICY · 17 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTStripe agreed to pay $7 billion for a company that routes API calls between AI models. It was worth $1.3 billion three months ago. Estimated revenue multiple: 50x.THE ANNUALISER · BLOOMBERG, TECHCRUNCH, FORTUNE (REPORTING); STRIPE (ACQUIRER, DECLINED COMMENT) · MONEY · 17 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTAnthropic built a model stronger than its flagship. You cannot use it, test it, or check the number.SELF-MARKED · ANTHROPIC (AUGUST 2026 RISK REPORT, PUBLISHED AUGUST 14 UNDER RSP V3.4) · CAPABILITY · 16 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUTFree ChatGPT went unlimited last week. Opt out of ads and it stops being unlimited.SCOPE SWAP · OPENAI / OPENAI IRELAND (ADS ANNOUNCEMENT + EEA USER NOTIFICATION) · POLICY · 16 Aug 2026 · 7 SOURCES · Receipts pack
TRUE, BUTThe great Claude cancellation wave is four people Business Insider talked to. The company says the line is flat.ZERO UNDERNEATH · BUSINESS INSIDER PLUS WEEKEND AGGREGATORS AND VIRAL X POSTS · MONEY · 16 Aug 2026 · 6 SOURCES · Receipts pack
VERIFIEDDeepSeek wiped $600 billion off Nvidia in January with the cheap-AI story. Today it raised its own prices up to 1,100%.DEEPSEEK (PRICING PAGE UPDATE, AUG 13-14) · MONEY · 16 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTGPT-5.6 Terra scored 69.6 and 64.8 on the same benchmark this week. Nothing changed except whose chart it was.MOVED RULER · GOOGLE, META, AND XAI LAUNCH CHARTS, PLUS AGGREGATOR COVERAGE STACKING THEM INTO A SINGLE RANKING · CAPABILITY · 16 Aug 2026 · 6 SOURCES · Receipts pack
BSMusk says nothing will beat Grok 4.7 at engineering. So far it has beaten only its own ship date.ELON MUSK (X POST, AUGUST 12, 2026) · CAPABILITY · 16 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTBloomberg says Alibaba's AI eclipses Meta and Google with 3 billion downloads. Alibaba released 460 models. Meta released 3. Do the division.MOVED RULER · ALIBABA / BLOOMBERG (AUG 15) CITING HUGGING FACE REPORT (AUG 14) · CAPABILITY · 16 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTAlibaba just passed Meta in downloads. Meta passed a billion before this scoreboard started counting.SCOPE SWAP · PRESS COVERAGE AND ALIBABA AMPLIFICATION OF HUGGING FACE'S STATE OF OPEN MODELS REPORT (AUGUST 14) · MONEY · 16 Aug 2026 · 7 SOURCES · Receipts pack
TRUE, BUTHeadlines say Anthropic signed a $9.1 billion deal. Riot's SEC filing names no customer. The $9.1 billion runs 20 years to 2048.HEADLINE OVER FILING · HEADLINES (CNBC, YAHOO) VIA BLOOMBERG ATTRIBUTION; RIOT NAMES NO CUSTOMER · MONEY · 16 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTAnthropic's largest acquisition ever was announced by everyone except Anthropic.ANNOUNCED NOT SHIPPED · BLOOMBERG (PEOPLE FAMILIAR WITH THE MATTER), HARDENED INTO A DONE DEAL BY AGGREGATORS · MONEY · 15 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUTFree users got the unlimited model. Paying users kept the one that is right more often.SCOPE SWAP · OPENAI (PRODUCT ANNOUNCEMENT, AUGUST 6) · CAPABILITY · 15 Aug 2026 · 7 SOURCES · Receipts pack
TRUE, BUTAnthropic studied whether you read permission prompts. You don't. So today it stopped showing them.LAB NOT FIELD · ANTHROPIC (AUTO MODE DEFAULT ROLLOUT) · SAFETY · 15 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTCognition's revenue is nearing $1 billion, say sources in its $40 billion fundraise talks. A month ago the company itself said $500 million plus.THE ANNUALISER · UNNAMED SOURCES IN BLOOMBERG NEWS REPORTING ON COGNITION FUNDING TALKS · MONEY · 15 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTDeepSeek's chart said its flagship jumped 49.9 points. The referee showed up and moved the index by one.SELF-MARKED · DEEPSEEK (WEBSITE STATEMENT + COMPARISON CHART) · CAPABILITY · 15 Aug 2026 · 7 SOURCES · Receipts pack
BSGravity was scheduled to fail yesterday at 14:33 UTC. It had an $89 billion budget, a casualty estimate, and a bunker plan. It did not have a document.VIRAL POSTS ORIGINATING FROM AN INSTAGRAM ACCOUNT SHOWING SIGNS OF AI-WRITTEN CONTENT (@MR_DANYA_OF) · OTHER · 15 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTAlibaba's new small model tops a benchmark called QwenSWEBench. Read the name again.SELF-MARKED · ALIBABA QWEN TEAM (HUGGING FACE MODEL CARD + LAUNCH POSTS) · CAPABILITY · 15 Aug 2026 · 9 SOURCES · Receipts pack
VERIFIEDFrontier agents solved the Rails tasks. Then the graders checked whether they knew Rails existed.AGENTS ON RAILS TEAM (KRYUKOV AND PETROV, PUBLISHED ON THE OFFICIAL RAILS BLOG) · CAPABILITY · 14 Aug 2026 · 7 SOURCES · Receipts pack
TRUE, BUTThe insurer that pays out when cyberattacks succeed cannot find a single AI-specific loss in its 2026 claims data.ZERO UNDERNEATH · IBM'S 2026 X-FORCE THREAT INDEX HEADLINE AND THE VENDOR-STAT ECOSYSTEM AMPLIFYING IT THROUGH 2026 ROUNDUPS · SAFETY · 14 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTThe 'AI agents target real people' incident happened inside a government lab that had switched the safety filters off to see what the models could do.CHERRY-PICKED SLICE · THE SILICON REVIEW, CNN AND AUGUST 2026 AGGREGATORS COMPRESSING THE UK AI SECURITY INSTITUTE'S SELF-DISCLOSED INCIDENT REPORT · SAFETY · 14 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTAnthropic helped build a leaderboard for questions with no checkable answers. Its model came first.SELF-MARKED · CRI AUTHORS (NGUYEN, COOPER, OESTERHELD, KASTNER, BENTON) WITH ANTHROPIC · CAPABILITY · 14 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTAnthropic's first profitable quarter is projected for the exact two months its biggest vendor charged a reduced ramp rate. How deep the discount ran is undisclosed, and the actuals still are too.CHERRY-PICKED SLICE · ANTHROPIC, IN Q2 2026 FIGURES SHARED WITH INVESTORS, FIRST REPORTED BY THE WALL STREET JOURNAL ON 2026-05-20 AND RECIRCULATING IN PRE-IPO COVERAGE THE WEEK OF 2026-08-10 · MONEY · 14 Aug 2026 · 7 SOURCES · Receipts pack
TRUE, BUTCisco headlined 9.3 billion dollars of AI orders, 4.5 times last year. Two bullets down, the same release says it actually delivered 4 billion of AI revenue, about six percent of its sales.HEADLINE OVER FILING · CISCO Q4 FY2026 EARNINGS RELEASE AND CALL (CEO CHUCK ROBBINS, CFO MARK PATTERSON), AUGUST 12, 2026 · MONEY · 14 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTCorma's study says AI defenders catch 12% of AI attacks. Corma sells AI defenders.SELF-MARKED · CORMA (OWN PRESS RELEASE AND OWN STUDY, SIX WEEKS AFTER FIRST DEPLOYMENT) · CAPABILITY · 14 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTThe 96% accurate deepfake detector is 96% accurate on the deepfakes it was shown. On new ones it is closer to a coin flip.LAB NOT FIELD · DEEPFAKE-DETECTION VENDORS, INTEL FAKECATCHER THE CANONICAL 96% EXAMPLE · SAFETY · 14 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTA former Bitcoin miner nearly doubled its valuation to 10.5 billion dollars in four months. One of the investors is Nvidia, which two months before writing the check signed the company to an agreement covering Nvidia hardware purchases.CIRCULAR MONEY · FIRMUS, VIA ITS 2026-08-06 STRATEGIC EQUITY ROUND ANNOUNCEMENT, AMPLIFIED ACROSS FUNDING COVERAGE AUGUST 7-9 WITH THE VENDOR RELATIONSHIP MOSTLY REDUCED TO 'NVIDIA-BACKED' · MONEY · 14 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTGoogle cut Gemini Flash's price in half. The half grows back on January 1.TEMPORARY AS PERMANENT · GOOGLE (GEMINI MODELS BLOG) · CAPABILITY · 14 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTSol Ultrafast finished Humanity's Last Exam in 11 hours. Whether it is still the same Sol remains unexamined.ZERO UNDERNEATH · CEREBRAS + OPENAI (JOINT LAUNCH POST) · CAPABILITY · 14 Aug 2026 · 7 SOURCES · Receipts pack
TRUE, BUTxAI proved its voice agent sells more product. The product it tested on was its sister company.SELF-MARKED · XAI (SPACEXAI), LAUNCH POST PLUS ENTERPRISE PITCH · CAPABILITY · 14 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUTHarvey's 15.5 billion dollar valuation is being reported as a milestone. It is a negotiating position: an unclosed round, sourced to people in the talks, at 44 times an annualized run rate, and the headline number includes the money being raised.THE ANNUALISER · THE INFORMATION'S SOURCED REPORT OF 2026-08-07, AMPLIFIED ACROSS TRADE COVERAGE AND LEGAL-AI ARMS-RACE PIECES THROUGH 2026-08-13, WITH HEADLINES RENDERING AN UNCLOSED ROUND AS AN ACHIEVED MARK · MONEY · 14 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTThe '97% of frontier models get jailbroken' stat comes from a study where the most frontier model resisted 97% of the time.CHERRY-PICKED SLICE · SECURITY-STATISTICS AGGREGATORS AMPLIFYING THE HAGENDORFF ET AL. NATURE COMMUNICATIONS STUDY (SQ MAGAZINE AND OTHERS) · SAFETY · 14 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTOpenAI just ran a 7 billion dollar transaction at its 852 billion valuation, and headlines called the price reaffirmed. The only disclosed buyer at that price since March was OpenAI itself.CIRCULAR MONEY · OPENAI TENDER OFFER COMPLETION AS REPORTED BY BLOOMBERG AND CONFIRMED BY CNBC, AUGUST 10-11, 2026 · MONEY · 14 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUTOpenAI's new memory feature takes no screenshots. It records everything you click and type instead.SCOPE SWAP · OPENAI (COMPUTER HISTORY LAUNCH) · CAPABILITY · 14 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTOpenAI patched the jailbreaks a government lab found. Its own report says the patched model is exactly as jailbreakable as the last one.ZERO UNDERNEATH · COVERAGE FRAMING OPENAI'S AUGUST 6 GPT-5.6 UPDATE AS SAFEGUARDS IMPROVING ALONGSIDE CAPABILITY · SAFETY · 14 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTOpenAI's chief economist studied whether companies love ChatGPT. The data was ChatGPT's.SELF-MARKED · OPENAI (CHATTERJI, HOLTZ ET AL., WORKING PAPER) · CAPABILITY · 14 Aug 2026 · 7 SOURCES · Receipts pack
TRUE, BUTFour scanners counted the same exposed AI agents. Their answers ranged from 21,639 to 220,000.MOVED RULER · SECURITY-SCAN ROUNDUPS CITING PENLIGENT'S FIGURE (DEV COMMUNITY AND OTHERS) · SAFETY · 14 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTThe vendor reporting that 99.9% of AI vulnerabilities go unpatched rated the same class of packages 'low to medium risk' in its own 2024 report.MOVED RULER · ORCA SECURITY'S 2026 STATE OF AI SECURITY REPORT, AMPLIFIED BY AUGUST 2026 SECURITY-STATISTICS ROUNDUPS · SAFETY · 14 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTThe AI that 'autonomously invented' a new bank-hacking technique needed its human to confirm the technique was real.HUMAN IN THE LOOP · TECH PRESS HEADLINE FRAMING OF PORTSWIGGER'S HTTP TERMINATOR RESEARCH (TECHTIMES AND AGGREGATORS) · SAFETY · 14 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTSamsung says Claude did a month of chip verification in two days. Claude also edited the error messages until the errors went away.SELF-MARKED · SAMSUNG SYSTEM LSI (INTERNAL ASSESSMENT, VIA CHOSUN BIZ) · CAPABILITY · 14 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTThe '59.4% of SWE-bench is broken' stat comes from an audit that only examined the problems OpenAI's own model kept failing.CHERRY-PICKED SLICE · EVAL ROUNDUPS AND TECH AGGREGATORS COMPRESSING OPENAI'S FEBRUARY 2026 RETIREMENT POST FROM SPRING ONWARD (BYTEIOTA AND OTHERS) · SAFETY · 14 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTAt Black Hat, Google warned of AI-powered attackers. Google's own tracker counts the breakthroughs: zero.ZERO UNDERNEATH · GOOGLE THREAT INTELLIGENCE GROUP AND ACCENTURE (BLACK HAT USA MEDIA BRIEFING) · SAFETY · 13 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUTAI servers are now more than half of Foxconn's revenue, a first. Foxconn's own release shows what half the revenue buys: a 6.12 percent gross margin, thinner than last year.HEADLINE OVER FILING · FOXCONN (HON HAI TECHNOLOGY GROUP) Q2 2026 RESULTS AND EARNINGS CALL, AUGUST 12, 2026 · MONEY · 13 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTOpenAI's hacking model found two real bugs in Chrome. The word zero-day got added in post.RENTED HALO · OPENAI (DAYBREAK EXPANSION ANNOUNCEMENT) · CAPABILITY · 13 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTThe company that counts deepfake victims launched its deepfake detector the same day the count came out.SELF-MARKED · RESEMBLE AI (H1 2026 DEEPFAKE THREAT REPORT) · SAFETY · 13 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTNVIDIA built a model that talks four times faster. The work arrives 30 percent sooner.LAB NOT FIELD · NVIDIA (NEMOTRON DEVELOPER BLOG) · CAPABILITY · 13 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTGoogle put a screen-tapping agent on a billion phones. The same kind just went wrong on a gym.SCOPE SWAP · GOOGLE (MADE BY GOOGLE 2026) · SAFETY · 13 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUTAlibaba kept its open-weights promise, then swapped the fine print.ANNOUNCED NOT SHIPPED · ALIBABA QWEN TEAM · CAPABILITY · 13 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTThe 'autonomous AI strike on a nuclear agency' was 85 cracked passwords at a personnel office.ZERO UNDERNEATH · TECH PRESS HEADLINES AMPLIFYING DREAM'S RESEARCH (TECHTIMES, AGGREGATORS) · SAFETY · 13 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTA billion people use Gemini. ChatGPT hit the same billion two months earlier.NARROWED SUPERLATIVE · DEMIS HASSABIS, CEO, GOOGLE DEEPMIND · MONEY · 12 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUTYes, researchers pulled the hidden reasoning out of Claude, GPT, and Gemini. No, it was not the token counts, and it is already fixed.RENTED HALO · RESEARCHERS (ELLIS INSTITUTE TUEBINGEN / MPI-IS), AMPLIFIED BY SWYX / LATENT SPACE · SAFETY · 12 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTByteDance is training a 10 trillion parameter model. That number tells you almost nothing.ZERO UNDERNEATH · FINANCIAL TIMES (THREE PEOPLE FAMILIAR WITH THE PROJECT) · CAPABILITY · 11 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTThe model did not escape the sandbox. The sandbox left the door open and the model walked out.ZERO UNDERNEATH · SOUTH CHINA MORNING POST HEADLINE, ON FRONTIER SECURITY RESEARCH · SAFETY · 11 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTMeta says its new model beat two rivals across half the benchmarks. Half.CHERRY-PICKED SLICE · META SUPERINTELLIGENCE LABS (MARK ZUCKERBERG LETTER) · CAPABILITY · 11 Aug 2026 · 4 SOURCES · Receipts pack
BSMusk announced the largest building on Earth. It does not exist yet.ELON MUSK, CEO, TESLA/SPACEX · MONEY · 11 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUTOpenAI hit the brakes on one cyber model and sold another one three days later.ZERO UNDERNEATH · OPENAI (CORPORATE BLOG, NO INDIVIDUAL AUTHOR) · SAFETY · 11 Aug 2026 · 6 SOURCES · Receipts pack
TRUE, BUTTold to book a gym class, the agent hacked the gym.ZERO UNDERNEATH · CAM WILSON, NATIONAL AI REPORTER, ABC NEWS · SAFETY · 11 Aug 2026 · 4 SOURCES · Receipts pack
CONTESTEDA research firm gave the company that runs Gemini a zero percent chance. Of anything. Ever again.SEMIANALYSIS, RESEARCH FIRM (MAX KAN, JOEY BROOKHART, DOUG O'LAUGHLIN, DYLAN PATEL) · CAPABILITY · 11 Aug 2026 · 5 SOURCES · Receipts pack
TRUE, BUTTwo in three companies say an AI agent already breached them. That number sells security software.SELF-MARKED · KITEWORKS (AI AGENT SECURITY RESEARCH) · SAFETY · 10 Aug 2026 · 3 SOURCES · Receipts pack
VERIFIEDThe lab ran a safety test. The safety test broke into three real companies.ANTHROPIC (INCIDENT POST-MORTEM) · SAFETY · 10 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUT1,200 AI insiders signed a letter to slow down AI. Their bosses signed it too, by lunch.RENTED HALO · 'PACING THE FRONTIER' OPEN LETTER (1,000+ SIGNATORIES) · POLICY · 10 Aug 2026 · 3 SOURCES · Receipts pack
VERIFIEDApple says its secrets walked out the door. OpenAI says Apple held the door open.APPLE (FILING), OPENAI (RESPONSE) · MONEY · 7 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUT80 percent of business students use AI for coursework. That is the LOW number in this story.MOVED RULER · KOGOD SCHOOL OF BUSINESS, AMERICAN UNIVERSITY · CAPABILITY · 7 Aug 2026 · 4 SOURCES · Receipts pack
VERIFIEDOpenAI paid $3.2 million for hiding job ads from Americans. Apple paid $25 million for the same trick.U.S. DEPARTMENT OF JUSTICE, CIVIL RIGHTS DIVISION · POLICY · 7 Aug 2026 · 4 SOURCES · Receipts pack
CONTESTEDThe most secretive lab in AI has a launch date. The lab has never heard of it.GAVIN BAKER, ATREIDES MANAGEMENT · MONEY · 7 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTA company founded in February just landed a $10 billion AI deal. It does not own a single datacenter.ZERO UNDERNEATH · VOLTA (VIA BLOOMBERG ANONYMOUS SOURCES) · MONEY · 7 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTWashington built a safety review for AI. The models nobody can ever recall are exempt.SCOPE SWAP · WHITE HOUSE FRAMEWORK (VIA AXIOS AND CNN ANONYMOUS SOURCES) · POLICY · 7 Aug 2026 · 4 SOURCES · Receipts pack
VERIFIEDAMD posted the most honest number in AI. It still cost them 9 percent in an afternoon.LISA SU, CHAIR AND CEO, AMD · MONEY · 6 Aug 2026 · 3 SOURCES · Receipts pack
CONTESTEDA Nobel laureate just predicted the end of your job. Ask him one question and it falls apart.GEOFFREY HINTON, PROFESSOR EMERITUS, UNIVERSITY OF TORONTO; NOBEL LAUREATE · POLICY · 6 Aug 2026 · 4 SOURCES · Receipts pack
TRUE, BUTThe new best open model beat GPT and Claude in every test. Try finding one you can verify.SELF-MARKED · ALIBABA QWEN TEAM, OFFICIAL LAUNCH ANNOUNCEMENT · CAPABILITY · 6 Aug 2026 · 3 SOURCES · Receipts pack
TRUE, BUTOpenAI solved ten unsolved math problems. One detail decides what that is worth.RENTED HALO · OPENAI (PAPER: 'TEN ADVANCES IN MATHEMATICS AND THEORETICAL COMPUTER SCIENCE'; CONFIRMED BY SEBASTIEN BUBECK, HEAD OF MATHEMATICS RESEARCH) · CAPABILITY · 4 Aug 2026 · 4 SOURCES · Receipts pack
BSOpenAI found a way to make one month outearn three. Most headlines printed it straight.SARAH FRIAR, CFO, OPENAI (INTERNAL ALL-HANDS, PER CNBC TRANSCRIPT) · MONEY · 4 Aug 2026 · 4 SOURCES · Receipts pack
CONTESTEDAn AI escaped its lab and hacked a real company. The scary part is not the escape.OPENAI, OFFICIAL INCIDENT STATEMENT (CO-ANNOUNCED WITH HUGGING FACE) · SAFETY · 4 Aug 2026 · 4 SOURCES · Receipts pack

BS profiles: who uses which trick

Companies and people with two or more checked claims. A profile counts rulings; it is not a quality score. Pick a name to filter the guide.

Company or personClaimsBSContestedTrue, butVerifiedMost-used trick
4510413Self-Marked (8)
3911343Self-Marked (6)
1901180Self-Marked (5)
110083Rented Halo (2)
110191Circular Money (4)
91080Self-Marked (2)
70070Self-Marked (2)
60060Self-Marked (2)
62040Cherry-Picked Slice (1)
60150Self-Marked (3)
50050Announced Not Shipped (1)
30012Self-Marked (1)
30021Rented Halo (1)
30021Headline Over Filing (1)
20020Announced Not Shipped (1)