September 14 to 20
Week 38, 2026- Mon
- Tue
- Wed
- Thu
- Fri
- Sat
- Sun
The week leaned optimistic. Trump called AI doom "a hoax" at the All-In summit on September 14 with Huang onstage. Four subscribers filed a Sherman Act class action against Anthropic, OpenAI, SpaceXAI and Google, naming Amodei, Altman, Musk and Hassabis and calling the September 12 pacing agreement a cartel. The labs did not comment.
Sunday, Sept 20
Ahead of a Sunday TV appearance, Bengio wrote that the commercial race is pushing frontier capabilities sans garanties suffisantes
, without sufficient safeguards.
- Yoshua BengioFounder, LawZero
Ahead of a Sunday TV appearance, Bengio wrote that the commercial race is pushing frontier capabilities
sans garanties suffisantes
, without sufficient safeguards.DiggMoreLessHe tied new Canadian and German government support for LawZero to the cyberattack on Hugging Face, and called for technical solutions and for citizens and policymakers to engage fully.
The post is in French and was read through Digg's capture; it is dated to the Sunday broadcast it announced, so it may have been posted a few days earlier.
Saturday, Sept 19
Nothing new.Friday, Sept 18
Andrew Ng and Databricks' Ali Ghodsi both said extinction fears are overblown, and Huang told CBS there is a 0% chance 2030 ends the world.
- Andrew NgFounder, DeepLearning.AI
“I don't see any step up in the risk of human extinction from AI compared to a few months ago.”The Machine Learning TimesMoreLess
In The Batch, Ng wrote
I don't see any step up in the risk of human extinction from AI compared to a few months ago.
He said
The loudest voices stoking fears about AI dangers have made tremendous headway in the past two weeks
and described the problems asa sign of the engineering work ahead, rather than insurmountable barriers or the sky falling.
Not on the roster; added as a guest. First published in The Batch, issue 371, on September 18; read here through the republication.
- Ali GhodsiCEO, Databricks
“right now the existential risk is close to zero. So why freak everybody out?”a16zMoreLess
On the a16z podcast, Ghodsi said
right now the existential risk is close to zero. So why freak everybody out?
He argued that doom talk
can cause a lot of mental health issues
and separated speculative superintelligence risk from the nearer problem of AI-powered cyberattacks.Not on the roster; added as a guest. Quotes are from FourWeekMBA's transcription of the episode.
- Jensen HuangCEO, Nvidia
“There is 0% chance that's going to be the end of the world”The Daily SignalMoreLess
Huang told CBS News
There is 0% chance that's going to be the end of the world
, about 2030, and called scaring peopleirresponsible.
He said
we would never ever, and never should, ship products before they're ready
and urged applying existing laws before any new ones.The Daily Signal describes a Friday interview with Jo Ling Kent; it is dated to Friday, September 18.
Thursday, Sept 17
Hinton told CNN that a kill switch will not work in the long run because AI will be able to persuade the people in charge of the switch not to pull the switch.
- Geoffrey HintonAI pioneer, Nobel laureate
“will be able to persuade the people in charge of the switch not to pull the switch.”StartupHub.aiMoreLess
Hinton told CNN that a kill switch will not work in the long run because AI
will be able to persuade the people in charge of the switch not to pull the switch.
The same week he told the BBC that a 10% estimate of catastrophe is
not unreasonable.
Dated by the article; the CNN interview date is not given.
Wednesday, Sept 16
Hinton told a closed Senate briefing that Congress has maybe a year, and Kevin Roose launched Machine Gods with NPR.
- Geoffrey HintonAI pioneer, Nobel laureate
“Maybe a year, but not much more than a year”The Next WebMoreLess
At a closed Senate briefing, Hinton said Congress has
Maybe a year, but not much more than a year
to put guardrails on AI.He said
AI has now reached the point where AI is designing better AI
and that itis going to get out of control unless we do something.
The Next Web says he spoke on Wednesday without giving the date; September 16 is the Wednesday before publication.
- Kevin RooseMachine Gods, NPR
“AI is the biggest story of our lifetimes, and most of the coverage is either cheerleading or doomsaying.”NPR (via WUNC)MoreLess
Announcing Machine Gods with NPR, Roose said
AI is the biggest story of our lifetimes, and most of the coverage is either cheerleading or doomsaying.
The twice-weekly show with Casey Newton is aimed, he said, at people
who want to know what is actually happening with AI.
Tuesday, Sept 15
Zuckerberg wrote that any lab not focused on alignment will fall behind, and Jack Clark told NPR that slowing AI is a collective action problem.
- Mark ZuckerbergCEO, Meta
“Any lab that doesn't focus on alignment will fall behind”BlockspaceMoreLess
Zuckerberg posted that
Any lab that doesn't focus on alignment will fall behind
, framing alignment as a competitive edge rather than a reason to slow.He acknowledged
a lot of debate about slowing progress on capabilities until alignment catches up
and wrote that trust and alignment are becoming the most important capabilities for agents and models. - Jack ClarkCo-founder, Anthropic
“We're saying we've seen warning shots, and that gives us a window in which we can act.”NPR (via KPBS)MoreLess
Clark told NPR's Morning Edition
We're saying we've seen warning shots, and that gives us a window in which we can act.
He said
This year, we've seen these things occur in the wild
and thatthere's a collective action problem here
, which is why the labs want to slow together rather than alone.
Monday, Sept 14
On stage at the All-In Summit, with Trump on speakerphone calling AI fears a hoax, Huang said We're going to make sure that everybody wins in the AI race of America.
- Jensen HuangCEO, Nvidia
“We're going to make sure that everybody wins in the AI race of America.”AxiosMoreLess
On stage at the All-In Summit, with Trump on speakerphone calling AI fears a hoax, Huang said
We're going to make sure that everybody wins in the AI race of America.
Huang called predictions that AI could end the world not grounded in science. He also said AI companies should pause or pace their work
if they feel that their company is out of control.
The mood this week
More optimistic than last week
3 worried1 neutral4 optimistic
7 worried2 neutral2 optimistic
Geoffrey HintonAI pioneer, Nobel laureateGeoffrey HintonAI pioneer, Nobel laureateCapability 5Outlook 2
Yoshua BengioFounder, LawZeroYoshua BengioFounder, LawZeroCapability 5Outlook 2
Jack ClarkCo-founder, AnthropicJack ClarkCo-founder, AnthropicCapability 5Outlook 3
Kevin RooseMachine Gods, NPRKevin RooseMachine Gods, NPRCapability 5Outlook 2
Jensen HuangJensen HuangCEO, NvidiaCapability 5Outlook 5
Ali GhodsiAli GhodsiCEO, DatabricksCapability 4Outlook 5
Andrew NgAndrew NgFounder, DeepLearning.AICapability 3Outlook 5
Mark ZuckerbergMark ZuckerbergCEO, MetaCapability 5Outlook 4
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.
Karen HaoKaren HaoJournalist, author of Empire of AICapability 3Outlook 1
Shakeel HashimShakeel HashimEditor, TransformerCapability 5Outlook 3
Demis HassabisDemis HassabisCEO, Google DeepMindCapability 5Outlook 3
Dan HendrycksDan HendrycksDirector, Center for AI SafetyCapability 4Outlook 2
Yann LeCunYann LeCunAMI LabsCapability 2Outlook 3
Jan LeikeJan LeikeAnthropicCapability 5Outlook 4
Elon MuskElon MuskSpaceXAICapability 5Outlook 4