Issue #2

THURSDAY 6 AUGUST 2026 · 3 CLAIMS CHECKED · 1 SURVIVED THE RECEIPTS · ISSUE 2 OF 17

BUSINESS LISA SU, CHAIR AND CEO, AMD · CLAIMED 2026-08-0401/3

AMD posted the most honest number in AI. It still cost them 9 percent in an afternoon.

01THE CLAIM

"We delivered an excellent quarter, with record revenue and profitability as Data Center revenue more than doubled year-over-year." [SOURCE ↗]

VERIFIED3 SOURCES · LIVE 2026-08-28

AMD posted the most honest number in AI. It still cost them 9 percent in an afternoon.

02THE CHECK

THE CLAIM. data-center revenue more than doubled. THE CHECK: 6.7 billion against 3.2 billion, SEC-bound, independently confirmed. Lisa Su did not stretch a single digit.

THE TWIST. the stock fell 9 percent anyway. Gross margin flat at 56 percent. Capex nearly tripled.

TRANSLATION INTO MONEY. demand is real, but even the winners are buying growth at full retail. Revenue tells you the gold rush exists. Margin tells you who keeps the gold.

03SAY THIS IN THE MEETING · 📸 SCREENSHOT IT

"When any vendor shows you doubled revenue, ask: 'And what happened to gross margin?' That one question separates demand from profit."

04YOUR MOVE ⚡ WHAT IGNORING THIS COSTS

This is what honest hype looks like: a true claim that still cannot carry the stock. Revenue tells you demand exists. Margins tell you who actually profits. Always check both before you believe an AI boom story.

RECEIPTS (3) · CONFIDENCE MEDIUM

every URL below answered a live HTTP check before publish · sweep 2026-08-28

▲ ir.amd.com ⧉ · "We delivered an excellent quarter, with record revenue and profitability as Data Center revenue more than doubled year-over-year."
▲ finance.biggo.com ⧉ · "AMD Posts Record $11.5B Revenue as Data Center Sales More Than Double, Lifting 2027 EPS Target Above $20"
● finance.yahoo.com ⧉ · "When revenue jumps 13% from one quarter to the next and gross margin doesn't budge, the new revenue is arriving at margins no better than the company's existing average."

POLICY GEOFFREY HINTON, PROFESSOR EMERITUS, UNIVERSITY OF TORONTO; NOBEL LAUREATE · CLAIMED 2026-08-0502/3

A Nobel laureate just predicted the end of your job. Ask him one question and it falls apart.

01THE CLAIM

"If AI can do routine intellectual labor, any job that consists mainly of routine intellectual labor is going to be done by AI." [SOURCE ↗]

CONTESTED4 SOURCES · LIVE 2026-08-28

1phase number used in the summary's first section

2phase number used in the summary's second section

3phase number used in the summary's third section

A Nobel laureate just predicted the end of your job. Ask him one question and it falls apart.

02THE CHECK

PHASE 1. THE PROPHECY. A Nobel laureate tells the Ai4 keynote that any job made of routine intellectual labor is going to be done by AI. No model. No timeline. No data. Just the medal.

PHASE 2. THE PUSHBACK, SAME STAGE. Andrew Ng: AI automates tasks, not jobs, and writing code is a small part of what engineers do. Fei-Fei Li: bring evidence, not fear.

PHASE 3. THE DATA. Yale Budget Lab, 33 months after ChatGPT: no discernible erosion of demand for cognitive labor. The collapse keeps not starting.

THE HARD TRUTH FOR YOU. your career decisions deserve better inputs than a laureate's hunch.

03SAY THIS IN THE MEETING · 📸 SCREENSHOT IT

"Next time someone quotes Hinton at your planning meeting, be the person who asks: 'What dataset is that prediction built on?' Then watch the room."

04YOUR MOVE ⚡ WHAT IGNORING THIS COSTS

A Nobel Prize is not a methodology. When someone predicts your professional extinction, ask for the model, not the medal. The best data says the collapse has not started. Plan on evidence, not keynotes.

RECEIPTS (4) · CONFIDENCE HIGH

every URL below answered a live HTTP check before publish · sweep 2026-08-28

▲ datacenterknowledge.com ⧉ · "If AI can do routine intellectual labor, any job that consists mainly of routine intellectual labor is going to be done by AI."
▲ forbes.com ⧉ · "What are those people going to do? Anything you could retrain them to do, AI will be able to do."
▼ forbes.com ⧉ · "In software engineering, AI is taking over a lot of the writing of code. But it turns out only a small part of what software engineers do is writing code."
▼ budgetlab.yale.edu ⧉ · "the broader labor market has not experienced a discernible disruption since ChatGPT's release 33 months ago, undercutting fears that AI automation is currently eroding the demand for cognitive labor"

CAPABILITY ALIBABA QWEN TEAM, OFFICIAL LAUNCH ANNOUNCEMENT · CLAIMED 2026-08-0303/3

The new best open model beat GPT and Claude in every test. Try finding one you can verify.

01THE CLAIM

"Meet Qwen3.8-Max, our most capable model to date. Next week, the open weights of Qwen3.8-Max will be released" [SOURCE ↗]

TRUE, BUT3 SOURCES · LIVE 2026-08-28

ALIBABA QWEN TEAM TRACK RECORD2 CLAIMS · 40/100 BS RATE →

The new best open model beat GPT and Claude in every test. Try finding one you can verify.

02THE CHECK

WHAT SURVIVES CHECKING. Qwen3.8-Max is genuinely frontier-class. Number two open model on the independent Vals Index. Real achievement.

WHAT DOES NOT. the launch table where it beats GPT-5.6 and Claude Fable 5 traces entirely to Alibaba's own unpublished internal evals. Not on the public leaderboard. No methodology. And the open weights promised for next week had not shipped.

THE PATTERN TO REMEMBER. strong model, homemade crown. This is the oldest play in AI marketing and it works because nobody checks.

03SAY THIS IN THE MEETING · 📸 SCREENSHOT IT

"Ask any model vendor one question: 'Which of these benchmark numbers appears on a leaderboard you don't run?' Then count the seconds of silence."

04YOUR MOVE ⚡ WHAT IGNORING THIS COSTS

The oldest trick in AI marketing: grade your own exam, publish the grade, skip the exam paper. Respect the model, ignore the crown until the number shows up on a leaderboard someone else runs.

RECEIPTS (3) · CONFIDENCE MEDIUM

every URL below answered a live HTTP check before publish · sweep 2026-08-28

▲ latent.space ⧉ · "Vals Index: #2 among open models, 66.1 score (matched Claude Opus 4.7)"
▼ yottalabs.ai ⧉ · "every stronger claim than that currently traces back to Alibaba's own press materials"
● glbgpt.com ⧉ · "the table mixes official model reports, system cards, public leaderboards, and Qwen's in-house evaluations"

THAT IS THE RECORD FOR ISSUE #2. NEXT VERDICT DROPS 9PM AEST.