Daily edition · 13 items · covers 19 Sep 12:45 → 20 Sep 16:30 UTC · how this edition was made
Sunday, 20 September 2026
Policy 23%Deployment 23%Frontier 15%Military 15%Compute 15%Security 8%
0:00 / 16:12
President Trump said on Truth Social on Saturday that he is forming an "AI Force, much like I did Space Force" and will name an AI czar, writing that his administration "will not in any way hinder or stifle the Growth" of the industry; NBC News reports he called warnings that AI could destroy humanity a "hoax". No executive order accompanies it: the White House presidential actions listing shows nothing dated September 19 or 20.
Nvidia chief executive Jensen Huang told CBS News that "We should go as fast as we can irrespective of anybody else", the sharpest public rejection yet of Dario Amodei's call to slow capability gains; CNBC reports Nvidia revenue reached $215 billion in the latest fiscal year, up from $17 billion in 2021. Rep. Ro Khanna used Face the Nation to call for a federal AI agency "like the FDA" and a Trump-Xi agreement banning recursive self-improving AI. Governor Newsom signed two bills extending California's ban on materially deceptive election media to January 1, 2031.
China's CXMT said its fifth-generation DRAM platform is in mass production at an 11.95 nm half-pitch, with at least 50 percent more dies per wafer than its predecessor. Seeking Alpha, summarising the Financial Times, reports technology companies provided as much as $300 billion in guarantees financing AI data centres and chips over the past year without recording most of it as debt. A New York Times investigation found DraftKings scored gamblers by expected losses and aimed roughly $400 million of automated promotions at them in 2025.
Frontier models & labs
Nvidia's Jensen Huang tells CBS News AI should go "as fast as we can", rejecting slowdown calls
- Asked by CBS News correspondent Jo Ling Kent whether AI development needs to slow down, Huang said "We should go as fast as we can irrespective of anybody else", adding "we would never ever, and never should, ship products before they're ready, deliver products that are unsafe." CBS says the interview airs on "Sunday Morning" on Sept. 20.
- CNBC reports Nvidia revenue "surged to $215 billion in the latest fiscal year, up from $17 billion in 2021", and that during a House hearing on Monday Treasury Secretary Scott Bessent said "the president is completely aligned with Jensen Huang, the CEO of Nvidia." CNBC reports Trump phoned Huang during Monday's All-In Summit in Los Angeles and told him, on speaker, "The robots are not going to be taking over the world."
- Speaking in Scotland on Thursday, Huang said the answer to safety concerns is "good old-fashioned engineering" and that "There were incidents, and those incidents, thankfully, did no harm", according to CNBC. An Nvidia spokesperson declined to comment to CNBC and White House representatives did not respond to its request for comment.
Alibaba publishes Qwen-Image-2.1 with a 7B generation component under a research licence Company claimSingle source
- The model card states 7B parameters in the model's visual generation component, with 32 Single-Stream DiT layers, and lists supported aspect ratios of 1:1, 4:3, 3:4, 3:2, 2:3, 16:9 and 9:16.
- Capabilities listed are text-to-image generation with native transparency (RGBA), image editing with up to 10 reference images, and local edits specified by circles, painted annotations or masks. The repository's commit history shows an update to a "0919 release checkpoint" and a README rewrite within the past day.
- The model card publishes no benchmark numbers or comparisons, so the capability descriptions are Alibaba's own and are not independently verified. The licence named on the model card is the Qwen Research License Agreement.
Security, misuse & threat intelligence
BragJack browser-assistant hijack earned more than $20,000 in bounties and produced two CVEs harmfulUpdate
- BleepingComputer reports the research "earned more than $20,000 in bug bounties from the five vendors, ranging from $600 to $7,000, and produced two CVEs". Chrome assigned CVE-2026-0628 and paid a $7,000 bounty.
- Both Google and Microsoft have since resolved the flaws they were assigned, according to BleepingComputer. This briefing covered Gal Weizman's underlying research on September 17; the bounty figures, the CVE assignments and the vendor fixes are the new facts.
- Weizman calls the method "Prompt Forcing": rather than smuggling instructions into content the agent reads, the attacker writes the whole prompt and the follow-ups, and the agent carries them out with its existing browser privileges. Forever Security's own per-vendor table does not match BleepingComputer's attribution of the second CVE, so only the Chrome assignment is reported here.
Military, defense & geopolitics
US Air Force plans 500 highly-autonomous collaborative combat aircraft by 2032, Defense One reports Single source
- Defense One reports the Air Force plans to field 500 highly-autonomous collaborative combat aircraft, or CCAs, by 2032. Air Force Secretary Troy Meink is quoted saying "A little bit scary, but we need to lean into it. Autonomy is one of, if not the key technology enabling cost-effective combat power."
- Shield AI president and co-founder Ryan Tseng told Defense One his company had worked with Ukrainians to improve the strike capability of one of its drone weapons by 70 percent.
- The article's argument is that the constraint is doctrine, training and defined permissions for autonomous decisions rather than airframes; Michael Gregg of the Air Force Research Laboratory is quoted saying "There's a trust and autonomy question... That is the future." One outlet carries the report and the Air Force is not quoted giving a programme cost or a contract award.
Chinese state-TV-affiliated account accuses Anthropic of 13 privacy-policy revisions easing US intelligence access Single source
- Geopolitechs reports that Yuyuan Tantian, a social-media account affiliated with China Central Television, examined 13 revisions to Anthropic's privacy policies since 2023 and characterised them as progressively enabling transfers of user data to US intelligence agencies.
- Per Geopolitechs, the post says five policy versions from May 2024 onward addressed moving user data from Canada, Brazil, South Korea and the EU to the United States, and that Anthropic's listed data-source categories grew from three to six between June 2024 and September 2025, with training use of user data shifting from off by default to on by default.
- The post cites a February 2026 Anthropic report on Chinese distillation that said it was sharing technical indicators with relevant intelligence agencies, a June 2026 statement to the US Senate describing analysis of 28.8 million user conversations and 25,000 user accounts, July 2026 job postings for Threat Intelligence Managers preferring US Top Secret clearances, and a September 10, 2026 report analysing approximately 200 million interactions.
- These are a Chinese state-affiliated account's characterisations of Anthropic, not verified findings about its practices. Anthropic is not quoted in the post, and one outlet carries the account of it.
Policy, regulation & law
Trump says he will form an "AI Force" modelled on Space Force and name an AI czar
- CNN quotes Trump's Truth Social post: "For this purpose, I am forming the AI Force, much like I did Space Force, which has been a tremendous SUCCESS, in my First Term. To that end, I will be announcing, in the near future, the AI 'Czar' — Only High I.Q. individuals need apply!"
- He also wrote "We will not in any way hinder or stifle the Growth of this incredible Industry. Rather, we will cherish it, help it, and watch over it, as it grows!", and said misconduct can be handled "very easily, with our already existing Criminal and Civil Justice System", according to CNN. NBC News reports he characterised warnings that AI could destroy humanity as a "hoax".
- No appointee was named. CNN says it has asked the White House whether the "AI Force" would be a branch of the military like the Space Force, which Trump created in 2019 as the first new US military service since 1947; the question is unanswered. Al Jazeera notes the administration's previous AI and cryptocurrency czar, David Sacks, stepped down earlier this year.
- CNN reports the administration will convene a high-level AI event on the sidelines of the UN General Assembly on Wednesday, and that a White House official said AI safety would be on the agenda when Trump meets Xi Jinping.
- There is no instrument behind the announcement yet. A check of the White House's presidential actions listing on September 20 found nothing dated September 19 or 20; its most recent entries were two September 18 items, on the H-1B visa programme and on entry of certain nonimmigrant workers, neither about AI.
Newsom signs two California election-deepfake bills, extending the ban's sunset from 2027 to 2031 beneficial
Office of Governor Gavin Newsom·California Legislative Information (AB 686)·California Legislative Information (AB 502)
- AB 686 (Berman) was approved by the Governor and chaptered on September 19, 2026 as Chapter 334, Statutes of 2026. The Governor's office says it "Extends the sunset date, from January 1, 2027, to January 1, 2031, on the prohibition on distributing materially deceptive audio or visual media (i.e., 'deepfakes')".
- AB 502 (Pellerin and Berman) was approved and chaptered the same day as Chapter 333, Statutes of 2026. Its official title is "An act to amend Section 20012 of the Elections Code, relating to elections, and declaring the urgency thereof, to take effect immediately", so it is in force now. The Governor's office says it modifies the law barring anyone, with malice, from knowingly distributing a campaign advertisement or other election communication containing materially deceptive content unless certain conditions are met.
- Both were signed inside a package of elections bills that the Governor's office frames as protecting California elections rather than as AI policy. Neither measure binds model developers: they govern the distribution of synthetic election media, not the systems that produce it.
Khanna calls for a federal AI agency "like the FDA" and a Trump-Xi ban on recursive self-improvement
- On "Face the Nation with Margaret Brennan" on September 20, Rep. Ro Khanna, the top Democrat on the Select Committee on the Chinese Communist Party, said "we need to set up an AI agency like the FDA", citing Geoffrey Hinton's argument that AI can be regulated "like we regulate nuclear energy, electricity, aviation, have independent technical experts make sure AI is safe". He added: "one of the things that really annoys me is when you have these tech CEOs out there saying, 'Oh, this is too complicated. People in Washington don't understand this.' Give me a break. We have regulated very complex things."
- Khanna said two core principles should anchor a US-China agreement: "banning any self-improving AI, recursive self-improving AI" and "safeguards so that AI can't be misused by people for biological nuclear weapons", and that this "needs to be the first agenda item on the Trump-Xi Jinping summit". On verification he said an agreement "should include international monitoring, it should include checks through the data centers where you can tell whether a model is being used for inference or training, and it should use enforceable and verifiable tools".
- Khanna confirmed the Chinese firms he wrote to have not responded, and conceded on distillation: "they're right that they are engaging in distillation. Basically, they are copying American models. That's one reason, frankly, to regulate American models because they're copying them." Brennan names DeepSeek and Alibaba among the recipients.
- Brennan said on air that the FBI, the NSA and CISA "have all warned that China-based AI companies are targeting and copying American artificial intelligence at what they describe as an industrial scale". That is the host's characterisation; the underlying agency warnings were not opened for this edition, and the White House has not published a US position on any of Khanna's proposals.
Compute, chips & infrastructure
CXMT says its fifth-generation DRAM platform is in mass production at an 11.95 nm half-pitch Company claim
- CXMT announced on Sunday at the 2026 World Manufacturing Convention in Hefei, Anhui Province, that its fifth-generation platform has entered mass production, with an active-area half-pitch of 11.95 nm, an array capacitor aspect ratio of 45:1 and a core cell array height of 6,762 nanometers, according to Global Times. Seoul Economic Daily reports the half-pitch was reached using quadruple patterning.
- Global Times reports "the number of dies produced per wafer rising by at least 50 percent" against the prior generation, and two LPDDR5X products of 24Gb each — 50% higher capacity than the previous generation — in 496-ball and 245-ball packages, both in mass production.
- Seoul Economic Daily reports CXMT held 9.5% of global DRAM revenue in the second quarter of 2026, fourth behind Samsung Electronics, SK hynix and Micron, and that the company declined to disclose yields or monthly output. It notes the 11.95 nm figure is difficult to compare directly with the "12-nanometer-class" process names used by the incumbents, so the parity claim is not independently established.
FT: technology companies provided as much as $300 billion in guarantees financing AI data centres and chips Single source
- Seeking Alpha, summarising the Financial Times, reports: "Technology companies have provided as much as $300 billion in guarantees to finance artificial intelligence data centers and chips during the past year, helping support the construction boom without immediately recording most of the exposure as debt."
- The mechanism described is a guarantee of minimum asset values through special-purpose arrangements, which lets a company add AI capacity without immediately booking the liability. Seeking Alpha writes that such guarantees "can obscure potential liabilities if demand fails or equipment devalues", and that losses could rise if AI hardware becomes obsolete quickly or the industry overbuilds.
- The Financial Times article itself could not be opened from here, so only the figures printed on the Seeking Alpha page are reported. Company-level breakdowns circulating in other summaries of the FT piece are not included because they were not read in the original, and one summary is the source for the $300 billion figure.
Deployment & impact
New York Times: DraftKings scored bettors by expected losses and aimed about $400 million of promotions at them harmfulSingle source
- A New York Times investigation published on September 19, drawing on internal memos, presentations, betting records and interviews with more than 40 former employees, found DraftKings built a machine-learning model in 2023 that scored each online casino player by how much they were expected to lose after receiving a free bet or bonus, weighing play frequency, daily account balances and the ratio of losses to wagers. Staff called the output an "elasticity" score.
- The company aimed roughly $400 million in AI-automated promotions at the highest-scoring accounts in 2025, and executives credit data science and personalised promotions with lifting the sportsbook margin by 13% in 2025.
- Former employees said a parallel model meant to flag customers sliding toward gambling harm, begun by data scientist Nestor Hernandez in 2024, was shelved; Chief Responsible Gaming Officer Lori Kalani said leadership decided against predictive risk tools because the approach was not "evidence-based". DraftKings said it "rejects any implication" that its marketing is unfair, and that promotions go to customers showing "sustained, engaged use of our platform, not toward customers based on their losses".
- nytimes.com could not be opened from here, so the figures above are taken from the Yahoo News, Startup Fortune and Tech Times reports of the investigation. All of them trace to the single Times piece.
Flock Safety offers voluntary buyouts across its 1,500-person workforce after surveillance backlash mixedSingle source
- TechCrunch, reporting on a Wired story, says Flock Safety announced voluntary severance packages on Friday that it described internally as the most generous it has ever offered, and that it expects a significant portion of its 1,500-person workforce to express interest. Without the buyouts, the company said it would almost certainly need to lay off staff.
- TechCrunch cites 46 documented cases of police officers misusing Flock's licence plate recognition technology, identified by The Washington Post in August 2026, Florida and Texas halting use of the system, and 90 cities terminating Flock contracts in August alone — a fourfold increase on the previous month.
- CEO Garrett Langley said the biggest damage from the backlash has been to internal morale. The story originates with Wired; TechCrunch's write-up is the version read here, and the company has not published a headcount reduction target.
AI evaluation startup Vals raises a $40 million Series A led by Andreessen Horowitz Single sourceCompany claim
- TechCrunch reports Vals, founded in 2024, raised a $40 million Series A led by Andreessen Horowitz, after a seed round led by 8VC and Bloomberg Beta. Model developers pay Vals to test their models, which co-founder Rayan Krishnan compares to students paying for the SAT.
- Headcount has grown from 8 at the start of the year to 25, with 10 to 15 more planned, and revenue is eight times last year's, according to TechCrunch. Vals evaluates models on law, finance and coding tasks, and is extending into recursive self-improvement, mental health, cybersecurity, biosecurity and the law of armed conflict. It recently launched a programme providing model evaluations to federal agencies.
- Vals does not publish its test materials, to stop models being trained against them, which also means its scores cannot be independently reproduced. The revenue multiple and the headcount figures are the company's own.