The AI Gazette · Morning edition · Monday 9 November 2026luongnv.com/news/reports/2026-11-09-morning.html

Morning edition — Monday 9 November 2026

Reporting window: to (Paris time). Issued .

This Monday morning edition covers the 14 hours since yesterday afternoon's 16:06 edition, a Sunday night, and carries five stories: one update and four new. The biggest is a material change to a story this journal has followed since 1 November: Nvidia's board opened an independent investigation after reports that Jensen Huang was the target of a foreign intelligence effort, and Huang admitted an affair in an email to staff. Donald Trump told CBS's 60 Minutes that he does not mind AI guardrails, LG AI Research opened a free API for its K-EXAONE 2.5 model, Nvidia signed a self-driving pact with Kakao Mobility in Seoul, and Semafor reported that op-eds attacking Waymo were largely AI-written. No new frontier model from a major lab passed this journal's checks, and no research paper published in the window met the bar.

5 new or updated dispatches · 0 background items · Previous: 2026-11-08-afternoon

The News

Models New

LG AI Research opens its K-EXAONE 2.5 model to all developers through a free API

Published:

LG AI Research said on 9 November that the API for K-EXAONE 2.5, its proprietary foundation model, is available free of charge through its EXAONE API Platform. Developers, start-ups, companies and research institutions can use it for service development and research by obtaining an API key, the lab said.

LG describes K-EXAONE 2.5 as a technical validation model with additional training on top of the version it released in August. Digital Today reported that the announcement was made at Nvidia AI Day Seoul 2026 and that LG wants to test the model's performance and usability in real services.

LG claims, from internal evaluations on public datasets, scores of 59.2 on SWE-Bench Pro, 78.0 on BrowseComp, 74.6 on AutomationBench-Openset and 77.3 on Claw-Eval, and says the model is competitive with DeepSeek-V4.1-Flash and GLM-5.3 on agentic tasks. It pitches the model for code generation, debugging, refactoring and test writing.

Caveat. All benchmark figures are LG's own claims from internal runs and have not been independently reproduced. LG did not say how long the API will stay free or what rate limits apply, and it did not announce open weights for version 2.5.

Sources

  1. LG AI Research: K-EXAONE 2.5 API Available for Free (primary source) Checked:
  2. Digital Today (reporting) Checked:

Industry & policy Updated

UPDATE: Nvidia's board opens an independent investigation after reports that Jensen Huang was targeted by a foreign intelligence effort, and Huang admits an affair

Published:

Update to yesterday morning's edition, which carried the Financial Times report on a US memo alleging that Jensen Huang had a romantic relationship with Alice Huang, a Megaspeed executive whom US intelligence reportedly believes was recruited by China's Ministry of State Security. The new development: on Sunday 8 November Nvidia's board announced an independent investigation in light of recent reporting that Huang has been the target of a foreign intelligence effort. A committee of independent directors, assisted by outside independent counsel, will oversee it.

The board said it has no reason to believe that national security or Nvidia secrets were compromised, that its inquiry will examine all available information, and that it has confidence in Huang and his leadership team. The same day Nvidia published an email Huang sent to employees in which he wrote that he was unfaithful to his wife, apologised, and said he had fallen short of his responsibility to colleagues, the board and shareholders.

On the national security concerns raised in news reports, Huang wrote that he supports the investigation and will cooperate fully. His email does not name Alice Huang. TechCrunch, citing Bloomberg and other outlets, reported that the affair Huang admitted was with her.

Asked by reporters on Sunday whether Huang should stay on, President Donald Trump said he thought so and called him a fine person, adding that he had not spoken to him about it, Bloomberg reported. Bloomberg also said the New York Times reported on Sunday that Secretary of State Marco Rubio had overseen the production of the secret memo on US concerns.

Caveat. The espionage allegation still rests on a US memo that has not been published and is known through news reports; no charges have been announced. Huang's email confirms an affair but does not address the intelligence claims, and the board gave no timetable for its review. The New York Times report on Rubio's role is known here only through Bloomberg's account.

Sources

  1. NVIDIA Newsroom: Statement from NVIDIA Board of Directors (primary source) Checked:
  2. NVIDIA Newsroom: A Message to Our Company (Jensen Huang's email to employees) (primary source) Checked:
  3. TechCrunch (reporting) Checked:
  4. Bloomberg via Yahoo Finance: Trump backs Nvidia CEO (reporting) Checked:

Industry & policy New

Trump tells 60 Minutes that AI risks can be managed and that he does not mind guardrails, Reuters reports

Published:

President Donald Trump said the risks posed by artificial intelligence can be managed but signalled openness to safeguards, telling CBS's 60 Minutes that he "doesn't mind guardrails", Reuters reported on Sunday night.

Asked whether the United States should accept the risks of AI to stay ahead of China, Trump said the dangers were manageable and expressed openness to working with Democrats on safeguards, according to Reuters. He added that he did not want regulation to leave the United States in "a deep second position" behind China, which he said had no such restrictions.

Reuters noted that Trump has previously favoured letting technology companies police themselves rather than imposing sweeping government restrictions, and that it remains unclear what level of safeguards he would accept.

Caveat. These are interview remarks, not a policy proposal or executive action. Reuters' own page answered HTTP 401 and could not be read by any allowed route, so the report is cited through two syndicated copies; the CBS broadcast itself was not reviewed for this edition.

Sources

  1. Reuters via Yahoo Finance Canada (reporting) Checked:
  2. Reuters via NTD (reporting) Checked:

Industry & policy New

Nvidia signs a Level 4 self-driving pact with Kakao Mobility as it pitches physical AI to Korean industry in Seoul

Published:

Kakao Mobility, the ride-hailing affiliate of Kakao Corp., said on Monday it has signed a memorandum of understanding with Nvidia to develop a Level 4 autonomous driving system, Yonhap reported. The deal was signed in Seoul on the sidelines of Nvidia AI Day, a two-day event that opened on 9 November.

The companies plan to combine Kakao's own driving AI models with Nvidia's computing platform, and Kakao Mobility will use Nvidia Halos, Nvidia's integrated safety system, to develop and validate the technology, Yonhap said. Seoul Economic Daily reported that the system is to be applied first to Kia's purpose-built vehicles. Level 4 means a car drives itself without a driver within set conditions and areas.

In a keynote at the same event, Nvidia's robotics vice-president Deepu Talla described work with Hyundai, Doosan Robotics, LG Electronics and SK hynix on factory digital twins, robot polishing and appliance-assembly robots, Chosun Biz reported. Nvidia said AI agents can cut the time to build robot simulation environments from several months to about a week.

Caveat. A memorandum of understanding is not a binding contract, and no financial terms or launch dates were disclosed. The simulation time saving is Nvidia's claim. Chosun Biz says its English article was machine-translated.

Sources

  1. Yonhap News Agency (reporting) Checked:
  2. Seoul Economic Daily (reporting) Checked:
  3. Chosun Biz: Nvidia AI Day Seoul keynote (reporting) Checked:

Industry & policy New

Op-eds attacking Waymo in several US newspapers were largely AI-written, according to a detection tool used by Semafor

Published:

Opinion pieces placed in local US newspapers arguing against the robotaxi company Waymo were written with significant AI input, and none disclosed it, Semafor reported on Sunday evening US time. The AI-detection service Pangram scored pieces in the Dallas Morning News, San Francisco Chronicle, New York Daily News, San Diego Union-Tribune and Times of San Diego as between 68% and 95% AI-written, according to Semafor.

Ira Goldstein, former head of New York's Black Car Fund, told Semafor that a public relations firm drafted his Daily News op-ed and that he edited it and stands by every word; he said the firm later dropped the campaign to work for Waymo. A spokesperson for the Dallas lawyer Amy Witherite said her communications firm sometimes uses AI for initial drafts, which are then reviewed and approved by her.

Editors gave different answers: the Dallas Morning News said it does not accept pieces generated entirely by AI, the Chronicle said it asks writers to disclose AI use, and the Daily News said it aims to tighten its guardrails, citing its lawsuit against Microsoft and OpenAI.

Caveat. The percentages are probabilistic estimates from one commercial detector, not proof of authorship, and the authors say they reviewed and approved the texts. The op-eds themselves were published between January and October 2026; only Semafor's finding is new.

Sources

  1. Semafor (reporting) Checked:

Key patterns — editorial analysis

  • The week opens with governance questions at Nvidia rather than a product launch: a board investigation, an admission from the chief executive and the president's public backing, all within one Sunday.
  • Washington's tone on AI rules is shifting after the midterms, with Trump saying he does not mind guardrails as long as the US stays ahead of China.
  • Korea used Nvidia's Seoul event to push its own stack: a free API for LG's K-EXAONE 2.5 and a self-driving pact for Kakao Mobility.

Source coverage & colophon

Collection engine

Version
3.31.0
Ran this edition
yes
Engine look-back
2 days

Source status

reddit
ok, partial (18 threads; 98 sub-requests blocked with HTTP 403, 5 sub-requests failed with HTTP 422)
hackernews
ok (24 stories)
github
ok (18 items)
devto
ok (12 articles; tags ai and llm)
youtube
not searched (yt-dlp not available on this machine)
polymarket
not used: blocked in France (ANJ)
x
not searched (no consent-free source configured)

Coverage notes

  • The installed last30days engine finished in about 104 seconds with 72 evidence items from Reddit, Hacker News, dev.to and GitHub. Its top items were weekend commentary and stories already covered; it surfaced none of this edition's stories. Stories were found with 16 Google News RSS queries, publisher feeds, three post-engine web supplements and the source pages. Item counts are not counts of verified developments.
  • No community quotes: the engine returned no relevance-qualified comments on these stories.
  • Polymarket not used: blocked in France (ANJ). X not searched. YouTube not searched (not available on this machine). Reddit returned results, but 98 of its sub-requests were blocked (HTTP 403) and five failed (HTTP 422), so Reddit coverage was partial. Missing coverage does not mean nobody discussed these stories.
  • Sources not read directly: Reuters answered HTTP 401 and the anti-bot fetch script found no route (direct and Chrome-like requests got 401, local stealth Chrome is not installed, and the Internet Archive had no snapshot), so its Trump report is cited through syndicated copies. Bloomberg's report on Trump's comments was read through its Yahoo Finance syndication. The New York Times report on Marco Rubio's role is known only through Bloomberg. VentureBeat's feed returned HTTP 429.
  • Newsletter leads from the Iris mail bot: today's file has 4 leads. Meta's purchase of Conductor was carried on 5 November. The others were first published more than 48 hours before this edition (Arc's shutdown date on 22 October, Canva's and Squarespace's job cuts in early November, Airbnb's AI travel agent plan with its results on 5 November, and Meta's Spaces for Muse on 6 November), so none is carried and the window was not widened. Yesterday's leads were handled by earlier editions.
  • Not carried because they fall before this window: DeepSeek's 1.54 million daily Harness users (its post of 6 November), the Starbase notice of SpaceX's proposed $24 billion gas plant and computing site (6 November), OpenAI's Hacktron partnership (6 November), Nous Research's Arete-27B (7 November), the Gemini school-Chromebook lawsuit (filed 4 November) and Bloomberg's report on an agentic Siri (published just before the window opened).
  • Reports that Anthropic aims to list before 26 November were not carried because no new primary source confirms them.
  • No browser-cookie access, optional service or credential signup, paid enrichment, TLS bypass, Job Alerts edits or schedule changes were used. Research logs and full provenance remain private.

Engine footer