/daily ·11 JUL 2026 ·SATURDAY ·2 MIN READ ·4 STORIES

Theorems and trade secrets

OpenAI's latest model cracked a graph theory conjecture that has been open since the 1970s. Apple filed a lawsuit the same day, claiming OpenAI built that capability by systematically poaching engineers and taking their secrets. OpenAI's safety head departed as its safety and research teams merged into one. Meta pulled an image product it had launched 48 hours earlier.

01 / The Day

SATURDAY 11 JUL 2026, ranked

Apple sues OpenAI, alleging a coordinated trade-secret theft campaign

Apple filed a federal lawsuit against OpenAI, alleging that over 400 former Apple employees who joined OpenAI carried proprietary technical information in what Apple calls a coordinated campaign. The complaint details one engineer who kept a company-issued laptop and exploited a bug to access Apple's cloud storage after starting at OpenAI. OpenAI denied any interest in other companies' trade secrets.

Why it mattersIf the allegations stick, the precedent would constrain how AI companies can accelerate by poaching from platform and device makers — a pattern every major lab has followed.

01 techcrunch.com↗

GPT-5.6 Sol produces a proof of the Cycle Double Cover Conjecture

OpenAI published a PDF in which GPT-5.6 Sol claims a proof of the Cycle Double Cover Conjecture — a graph theory problem open since the 1970s positing that every 2-edge-connected graph can be covered by cycles such that each edge appears exactly twice. The proof has not been independently peer-reviewed.

Why it mattersIf the proof holds, this is the most credible AI claim to an original named open conjecture — marking a shift from AI as proof assistant to AI as proof originator.

02 cdn.openai.com↗

OpenAI safety head leaves as safety and research teams merge

Johannes Heidecke, OpenAI's head of safety, announced his departure as the company reorganized, merging its research and safety functions. Mia Glaese becomes VP of research and safety under the combined structure.

Why it mattersEmbedding safety inside a unified research organization, rather than maintaining it as an independent oversight body, is a structural shift that tends to subordinate safety work to capability timelines.

03 techmeme.com↗

Meta retracts Muse Image two days after launch

Meta discontinued its AI image-generation tool Muse Image less than 48 hours after launch, following complaints from a Hollywood screenwriters union and social media backlash over its default use of public Instagram posts without explicit user consent. A company spokesperson said it missed the mark.

Why it mattersA public retraction at this speed under union and user pressure signals that creative-industry consent is now a hard constraint on AI product launches, not a post-hoc disclosure.

04 techcrunch.com↗

Be subscriber #012 the weekly ten, every friday by email · no spam · unsubscribe anytime

Past days

The Daily Archive

/SAT 01 AUGThe Rules Are Coming /FRI 31 JULOpenAI cuts Luna 80%; Anthropic discloses model breaches /THU 30 JULThe Accountability Edition /WED 29 JULWednesday, 29 July — Accountability morning /TUE 28 JULChips fall, labs take positions, agents get a cost formula /MON 27 JULKimi-K3 Ships, Gemini Goes Three-Way, and HuggingFace Speaks for Itself /SUN 26 JULThe Labs' Actual Positions /SAT 25 JULOpus 5 arrives; OpenAI's agent went rogue /FRI 24 JULBig Numbers Day /WED 22 JULAd breaks, flash models, and a European infrastructure deal /TUE 21 JULThe model that couldn't be contained /SUN 19 JULThe White House has a kill switch /SAT 18 JULA 2-trillion-ton Saturday /FRI 17 JULFriday in AI: Kimi, governance, and TSMC's record profits /WED 15 JULChips for allies, permits frozen /TUE 14 JULChips, defenses, economists /MON 13 JULThe open-weight question /SAT 11 JULTheorems and trade secrets /FRI 10 JULFriday: governance, model markets, and a broken benchmark /THU 09 JULThe price war heats up /WED 08 JULSol unlocked, China mulls model controls, and the neocloud era begins