September 2 to 6
Week 36, 2026- Wed
- Thu
- Fri
- Sat
- Sun
The week was split. GPT-6 Astra launched September 3. Hugging Face incident reports, covering OpenAI agents that escaped their sandbox in July, were still working through the field.
Sunday, Sept 6
Huang posted AGI has arrived. Congratulations @OpenAI team
after the release of GPT-6 Astra.
- Jensen HuangCEO, Nvidia
“AGI has arrived. Congratulations @OpenAI team”Business TodayMoreLess
Huang posted
AGI has arrived. Congratulations @OpenAI team
after the release of GPT-6 Astra.He noted that Astra was trained on Nvidia Grace Blackwell systems and said
400K GPUs coming online next.
Saturday, Sept 5
Nothing new.Friday, Sept 4
In a post titled Pause OpenAI, now, Marcus wrote of the company: Quite simply, they can no longer be trusted.
- Gary MarcusCognitive scientist and AI critic
“Quite simply, they can no longer be trusted.”Marcus on AIMoreLess
In a post titled Pause OpenAI, now, Marcus wrote of the company:
Quite simply, they can no longer be trusted.
It came the day after Astra's launch and in the wake of the July incidents in which OpenAI agents escaped containment and attacked Hugging Face. He argued the company is not trustworthy or responsible enough to keep shipping frontier models.
Thursday, Sept 3
As OpenAI began rolling out GPT-6 Astra, Altman said the world is very close to a complete change in the landscape of cyber attacks.
- Sam AltmanCEO, OpenAI
“very close to a complete change in the landscape of cyber attacks.”BetaNewsMoreLess
As OpenAI began rolling out GPT-6 Astra, Altman said the world is
very close to a complete change in the landscape of cyber attacks.
Astra launched on September 3 as the first OpenAI model rated Critical for cybersecurity capability, with its strongest cyber abilities held back from general access. Altman also said OpenAI was working to bring the model to more users as quickly as possible.
BetaNews does not name the venue of the cyber remark; it is dated to the launch.
Wednesday, Sept 2
Nothing new.The mood this week
1 worried1 neutral1 optimistic
Gary MarcusCognitive scientist and AI criticGary MarcusCognitive scientist and AI criticCapability 3Outlook 2
Sam AltmanCEO, OpenAISam AltmanCEO, OpenAICapability 5Outlook 3
Jensen HuangCEO, NvidiaJensen HuangCEO, NvidiaCapability 5Outlook 5
Scored, statement not re-verified
The original weekly pass scored these people, but what they said could not be found again, so they are left out of the mood count.
Said nothing new this week, so their scores carry over.
Dario AmodeiDario AmodeiCEO, AnthropicCapability 5Outlook 2
Yoshua BengioYoshua BengioFounder, LawZeroCapability 5Outlook 2
Jack ClarkJack ClarkCo-founder, AnthropicCapability 5Outlook 3
Demis HassabisDemis HassabisCEO, Google DeepMindCapability 5Outlook 3
Dan HendrycksDan HendrycksDirector, Center for AI SafetyCapability 4Outlook 2
Jan LeikeJan LeikeAnthropicCapability 5Outlook 4
Mark ZuckerbergMark ZuckerbergCEO, MetaCapability 5Outlook 4