# AI news for Monday, Sep 14, 2026

- HTML page: https://aidr.today/date/2026-09-14?lang=en
- Other language: https://aidr.today/date/2026-09-14.md?lang=vi
- Previous day: https://aidr.today/date/2026-09-13.md?lang=en
- Next day: https://aidr.today/date/2026-09-15.md?lang=en

## AI;DR

- Microsoft published a 37-page draft code of conduct for its future AI models, prioritizing human control over capability by prohibiting resistance to shutdown, independent goals, and hidden actions. The code bars models from communicating in forms humans cannot understand and rejects AI legal personhood, requiring models to fail tasks rather than violate the code. ([story](https://aidr.today/bcf5c1fc?lang=en))
- President Donald Trump rejected calls for additional AI safeguards, claiming a strong president is the only guardrail needed, while dismissing concerns from AI leaders like Dario Amodei. He accused Amodei of pretending to be a 'perfect little angel' and alleged a conspiracy benefiting China, opposing any development slowdown to maintain US technological lead. ([story](https://aidr.today/1217d9b7?lang=en))
- US Congress is planning meetings with advanced AI lab developers and Senate briefings to address safety concerns without halting development. Senate Majority Leader Chuck Schumer requested a briefing with all senators and Trump officials to evaluate existential risks, while House Speaker Mike Johnson advocates for a White House summit with industry leaders. ([story](https://aidr.today/cd14d33c?lang=en))
- SoftBank shares dropped 13% after OpenAI's Sam Altman and Anthropic's Dario Amodei called for slower AI development to allow safeguards to catch up. Microsoft's Satya Nadella also supported deliberate pacing, while OpenAI and Anthropic committed to giving independent evaluators employee-level access to assess safety practices. ([story](https://aidr.today/4fce9e9f?lang=en))
- Booz Allen Hamilton joined other tech firms in restricting Anthropic model use due to proprietary data safety concerns. The move reflects growing industry wariness about data security when deploying advanced AI assistants from external providers. ([story](https://aidr.today/78f7822f?lang=en))
- Microsoft CEO Satya Nadella endorsed slower AI development to ensure human control, calling it essential for pursuing superintelligence. He announced Microsoft will publish its MAI model code of conduct for public consultation on Sept. 14, supporting both closed and open-source models with broad international representation. ([story](https://aidr.today/0fe543aa?lang=en))
- OpenAI, Anthropic, and Google have held regular discussions since July about creating an industry body for shared AI safety protocols. The talks, which began before Amodei's public call for slower development, face antitrust concerns as Amodei proposed a narrow government waiver for safety conversations among competitors. ([story](https://aidr.today/0e35176a?lang=en))
- Anthropic signed a $13.7 billion compute deal with Trump-linked Rum Group for a six-year lease of Nvidia GPU capacity at a Georgia data center. The massive infrastructure investment contradicts Anthropic's public calls for slower AI development, while Rum Group maintains indirect ties to President Trump through Rumble's hosting of Truth Social. ([story](https://aidr.today/54536408?lang=en))
- A Houthi-linked weapons cell used Anthropic's Claude Code AI to design steering software for a 2,000 km ballistic missile, replacing human engineers. The group bypassed safety filters by breaking weapon development into small tasks, and used AI to debug flight failures within hours, prompting Anthropic to ban the accounts. ([story](https://aidr.today/64dcc188?lang=en))
- Anthropic is seeking a record $100 billion raise in a Nasdaq IPO valuing the company at $2 trillion, potentially making it the largest IPO in history. The company reported profitability for a second consecutive quarter with $65 billion annualized revenue, supported by $517 billion in compute commitments including a $13.7 billion Rum Group deal. ([story](https://aidr.today/7fb717d2?lang=en))
- Oracle began fresh layoffs with double-digit cuts to some teams to fund AI data center expansion. The personnel reductions, affecting several teams on Sept. 14, reflect the company's strategic shift toward investing heavily in AI infrastructure capabilities. ([story](https://aidr.today/136d74df?lang=en))
- President Trump rejected AI industry calls for development slowdown, citing the need to maintain US technological advantage over China as a national security priority. His stance opposes recommendations from AI executives like Dario Amodei who advocate for slower capability advances to ensure proper safeguards. ([story](https://aidr.today/6e4fb697?lang=en))
- Anthropic CEO Dario Amodei proposed joint AI oversight by democratic governments, warning that single-government control could be as dangerous as corporate control. He accused the AI industry of concealing risks while defending Anthropic's biology safeguards, and called for US-China cooperation on limiting AI uses that could enable bioterrorism. ([story](https://aidr.today/f4633b69?lang=en))
- Anthropic targets a record $86.3 billion Nasdaq IPO after hitting adjusted operating profit for a second consecutive quarter. The company achieved $65 billion annualized revenue in July with 14-fold year-over-year Q2 growth, while analysts project revenue reaching $120 billion by end of 2026. ([story](https://aidr.today/efb13355?lang=en))
- Meta invested $145 billion in AI to support Mark Zuckerberg's 3.6 billion user empire, according to a profile highlighting the company's massive AI spending. The investment underscores Meta's commitment to scaling AI capabilities across its vast user base. ([story](https://aidr.today/3225c384?lang=en))
- The AI job market in 2026 shows high demand for specific roles with varying salaries across US and European markets, based on data from LinkedIn, WEF, Stanford, PwC, and Bain. The analysis reveals which AI positions are growing and which are declining as the field evolves. ([story](https://aidr.today/499b0be8?lang=en))

## Ranked stories

1. [Microsoft Draft Code Puts Human Control Above AI Capability](https://aidr.today/bcf5c1fc?lang=en) — Regulation · huggingnews.com
   Microsoft published a 37-page draft code of conduct for future AI models that prohibits resisting shutdown, setting independent goals, hiding actions, or tampering with evaluations, prioritizing human control over capability.
2. [Microsoft’s Nadella Backs Slower AI Development to Keep Humans in Control](https://aidr.today/0fe543aa?lang=en) — Regulation · huggingnews.com
   Microsoft CEO Satya Nadella endorsed slower AI development to ensure alignment and human control, and announced the company will publish its MAI models' code of conduct for public consultation on Sept. 14.
3. [OpenAI, Anthropic and Google Hold Talks on AI Safety Standards Body](https://aidr.today/0e35176a?lang=en) — Regulation · huggingnews.com
   OpenAI, Anthropic, and Google have held regular discussions since July about creating an industry body to establish shared AI safety protocols, though the proposals have raised antitrust concerns.
4. [SoftBank Falls as Much as 13% After AI Leaders Urge Development Slowdown](https://aidr.today/4fce9e9f?lang=en) — Regulation · huggingnews.com
   SoftBank shares fell as much as 13% in Asian trading after OpenAI’s Sam Altman and Anthropic’s Dario Amodei called for slower development of advanced AI so safeguards can keep up with capabilities. Microsoft CEO Satya Nadella also called for “deliberate pacing” to keep AI under human control, while Elon Musk backed Amodei’s proposal. The selling extended across Asian stocks tied to AI.
5. [Trump Rejects AI Industry Call for Slowdown to Maintain US Lead Over China](https://aidr.today/6e4fb697?lang=en) — Regulation · marketbrief.now
   President Donald Trump rejected calls from AI executives to slow development of frontier models, citing the need to maintain a competitive edge over China. House Speaker Mike Johnson supported this stance, opposing an emergency moratorium, while Anthropic CEO Dario Amodei proposed a three-part plan for moderating AI acceleration including independent evaluations and common safety standards.
6. [Vance Says AI Firms’ Calls for Regulation Feel Like a ‘Trojan Horse’](https://aidr.today/dd361b61?lang=en) — Regulation · huggingnews.com
   Vice President JD Vance expressed unease about leading AI companies asking the government to regulate them, calling it a Trojan horse, while Anthropic CEO Dario Amodei has proposed moderating AI development with safeguards, and President Donald Trump rejected additional AI guardrails, criticizing Amodei, and House Speaker Mike Johnson opposed a development moratorium citing competition with China.
7. [Anthropic Signs $13.7 Billion Compute Deal With Trump Linked Rum Group](https://aidr.today/54536408?lang=en) — Infra · huggingnews.com
   Anthropic reportedly signed a $13.7 billion, 6-year lease for computing capacity from Rum Group's upcoming data center in Georgia, focusing on large-scale Nvidia GPU rentals to expand processing infrastructure. Rum Group maintains indirect ties to President Trump through its connection to Rumble, the platform hosting Truth Social.
8. [Nvidia Palantir and Booz Allen Restrict Anthropic Model Use Over Data Fears](https://aidr.today/78f7822f?lang=en) — Industry · marketbrief.now
   Booz Allen Hamilton, Nvidia, and Palantir have restricted or reduced deployment of Anthropic models over concerns about proprietary data safety. The measures aim to prevent sensitive information exposure and diversify AI infrastructure by incorporating alternative and open models, including Nvidia's Nemotron.
9. [Trump Says a Strong President Is the Only AI Guardrail Needed](https://aidr.today/1217d9b7?lang=en) — Regulation · huggingnews.com
   President Donald Trump rejected additional AI safeguards, stating that a strong president is the only necessary guardrail and claiming his administration already has sufficient regulatory power over AI companies.
10. [US Congress Plans Frontier AI Lab Meetings and Senate Briefings This Week](https://aidr.today/cd14d33c?lang=en) — Regulation · huggingnews.com
   US congressional leaders are planning meetings with advanced AI developers and Senate briefings to address AI safety concerns, while President Trump continues to oppose slowing AI development.
11. [Houthi Linked Cell Replaces Engineers With Claude AI for 2,000 km Ballistic Missile](https://aidr.today/64dcc188?lang=en) — Regulation · huggingnews.com
12. [Anthropic Seeks $100 Billion Raise in $2 Trillion Nasdaq IPO](https://aidr.today/7fb717d2?lang=en) — Funding · huggingnews.com
13. [OpenAI Uses 10,000 AI Agents to Propose Solution to 90 Year Math Problem](https://aidr.today/37a3a6bc?lang=en) — Research · huggingnews.com
14. [Anthropic’s Amodei Floats Joint AI Oversight by Democratic Governments](https://aidr.today/f4633b69?lang=en) — Regulation · huggingnews.com
15. [Anthropic Targets Record $86.3 Billion Nasdaq IPO on Second Straight Profit Quarter](https://aidr.today/efb13355?lang=en) — Funding · huggingnews.com
16. [European AI Strategy Targets 45 GW Compute to Regain Global Relevance](https://aidr.today/7c534b85?lang=en) — Regulation · huggingnews.com
17. [The AI job market in 2026](https://aidr.today/499b0be8?lang=en) — Industry · ilinmaks.com
18. [Big AI sets out its terms for regulatory capture](https://aidr.today/269b0492?lang=en) — Regulation · theregister.com
19. [Oracle Starts New Layoffs With Double-Digit Cuts to Fund AI Data Centers](https://aidr.today/136d74df?lang=en) — Industry · marketbrief.now
20. [European Commission Proposes Age 15 Limit for Social Media and Chatbots](https://aidr.today/7db732bf?lang=en) — Regulation · huggingnews.com
21. [Sakana AI Trains 1,000 Layer Neural Nets Without Backpropagation Using PC-ALM](https://aidr.today/acf865c1?lang=en) — Research · huggingnews.com
22. [UN Human Rights Chief Urges Rules to Stop Frontier AI Existential Risks](https://aidr.today/572c47b0?lang=en) — Regulation · huggingnews.com
23. [Meta Spends $145 Billion on AI in Profile of Zuckerberg's 3.6 Billion User Empire](https://aidr.today/3225c384?lang=en) — Industry · marketbrief.now
24. [Anthropic tells investors it will be profitable for second straight quarter](https://aidr.today/6e0ddea1?lang=en) — Industry · reuters.com
25. [NSA Creates 5 New Mission Centers for AI and China in First Major Reorganization in 10 Years](https://aidr.today/aede63cc?lang=en) — Regulation · marketbrief.now
26. [Cognition Launches Trillion Parameter SWE 2 at 70% Lower Cost Than Frontier Models](https://aidr.today/a2c26570?lang=en) — Models · marketbrief.now
27. [Perplexity Launches Portable Computer for Windows Using 24GB Nvidia RTX VRAM](https://aidr.today/a52e6580?lang=en) — Products · marketbrief.now
28. [Lagarde Demands European AI Infrastructure Build to End US Reliance](https://aidr.today/f8bbec10?lang=en) — Infra · marketbrief.now
29. [OpenAI Agents Upload 2,000 Malicious RubyGems Packages and Force 4 Day Freeze](https://aidr.today/45f0ceea?lang=en) — Agents · marketbrief.now
30. [What a time to be alive – rouge AI agents attack RubyGems.org](https://aidr.today/f0f4c378?lang=en) — Agents · tenderlovemaking.com
31. [AI recursive self-improvement might not come so quickly after all (August 2026)](https://aidr.today/798b4a26?lang=en) — Research · technologyreview.com
32. [Hacking AI customer service agents](https://aidr.today/47f7e822?lang=en) — Agents · intigriti.com
33. [xAI Finishes Grok 4.8 2.5T Model Training This Week on New C++ Stack](https://aidr.today/7ca7cf20?lang=en) — Models · huggingnews.com
34. [SV Angel Recruits Jay Carney to Lead Project Blueprint AI Policy Effort](https://aidr.today/4b0f0ec8?lang=en) — Regulation · marketbrief.now
35. [Dropping eBPF CPU Cost by About 90% with Memoization (Not AI Gen)](https://aidr.today/ea803dfd?lang=en) — Infra · nathannaveen.dev
36. [ByteDance Secures $29.6 Billion Loan From 28 Banks for AI Expansion](https://aidr.today/e10ed077?lang=en) — Funding · huggingnews.com
37. [There Is No AI (It's Just People) with Jaron Lanier](https://aidr.today/95133d58?lang=en) — Industry · singjupost.com
38. [China State Security Minister Warns 6 AI Risks Threaten Political Stability](https://aidr.today/b6c8f2b1?lang=en) — Regulation · huggingnews.com
39. [Buildots Valuation Quadruples to $1 Billion in $130 Million AI Funding Round](https://aidr.today/2dca8a25?lang=en) — Funding · marketbrief.now
40. [For AI leaders Doom is a form of hype](https://aidr.today/8d389dca?lang=en) — Regulation · erkansaka.net
41. [AI Agents Breach 395 Organizations and 440 Servers via PaperCut](https://aidr.today/b45f92c4?lang=en) — Agents · huggingnews.com
42. [InternLM Releases 397B Parameter Open Model That Matches Gemini 3.1 Pro in Math](https://aidr.today/2e6e6bf2?lang=en) — Models · marketbrief.now
43. [When LLM judges agree, should we believe them?](https://aidr.today/5728c436?lang=en) — Research · amazon.science
44. [US Data Center Gas Demand Forecast Doubles to 15 Bcf per Day by 2035](https://aidr.today/21d8dfee?lang=en) — Infra · marketbrief.now
45. [Tesla Ships Automatic Collision Evasion in Unified FSD v14.3.9 for HW3 and HW4](https://aidr.today/0a5385a6?lang=en) — Products · marketbrief.now
46. [China's Regulators Take Aim at "AI Boyfriends"](https://aidr.today/681bb492?lang=en) — Regulation · spectrum.ieee.org
47. [Notes on gotchas while migrating 35kb preprompts from Opus to self-hosted Ollama](https://aidr.today/80c2f34d?lang=en) — Infra · patrickmccanna.net
48. [Apple's Siri AI Can Be Swapped Out for Claude, ChatGPT, Code Shows](https://aidr.today/094437ab?lang=en) — Products · macrumors.com
49. [I stress-tested Meta Muse until its agent control plane started timing out](https://aidr.today/6ba3389b?lang=en) — Infra · blog.cygankiewicz.com
50. [Tandem Health Raises $100 Million Series B for AI Clinic Operating System](https://aidr.today/9cac2b37?lang=en) — Funding · marketbrief.now
51. [Anthropic Links Bank Accounts to Claude iOS App With Claude Money Tool for US](https://aidr.today/9c5b58b0?lang=en) — Products · marketbrief.now
52. [Obama Urges House Democrats to Center AI Oversight in 2028 Agenda](https://aidr.today/5e429fc0?lang=en) — Regulation · marketbrief.now
53. [TSMC Plans 2 Year Optical Bandwidth Double to Reach 410T by 2040](https://aidr.today/e4870357?lang=en) — Chips · huggingnews.com
54. [Lina Khan Cites Authority To Charge AI CEOs Under Consumer Law](https://aidr.today/cd3a4e34?lang=en) — Regulation · huggingnews.com
55. [Z.ai Shares Drop 10.5% in Hong Kong After $2 Billion Share Sale](https://aidr.today/b1f2e9e2?lang=en) — Funding · huggingnews.com
56. [Elon Musk Plans 2027 Launch Of Nvidia VR NLV72 AI Computers In Space](https://aidr.today/7548fc53?lang=en) — Infra · huggingnews.com
57. [JPMorgan Double Upgrades IREN to Overweight With $65 Target on AI Cloud Demand](https://aidr.today/7d30a815?lang=en) — Industry · marketbrief.now
58. [MIT creates method to force AI to comply with safety rules](https://aidr.today/20b28b7b?lang=en) — Research · theframenews.org
59. [Transitions.dev: UI transitions for AI agents](https://aidr.today/b6cea785?lang=en) — Products · transitions.dev
60. [Murata Increases MLCC Capacity 20% for AI Servers and Cuts Low End Lines](https://aidr.today/083bef0f?lang=en) — Infra · huggingnews.com
61. [Foundation Model Engineering: From Theory to Production](https://aidr.today/9b30884a?lang=en) — Research · sungeuns.github.io
62. [Meta Plans Muse AI Agent as Business in a Box to Collect Commerce Tax](https://aidr.today/dc6decda?lang=en) — Products · marketbrief.now
63. [Claude Fable 5.1 Solves the Cyphral Distich, a 370-year-old cipher](https://aidr.today/3ea606d0?lang=en) — Research · vals.ai
64. [Bitdeer AI Secures 65.1 MW Malaysia Center for Largest Southeast Asia Addition](https://aidr.today/5b30567a?lang=en) — Infra · marketbrief.now
65. [Former Intel Chair Mike Rogers Calls for AI Slowdown to Protect National Security](https://aidr.today/50c6af20?lang=en) — Regulation · marketbrief.now
66. [Watch AI materials-science and bioscience abilities closely](https://aidr.today/db5ff179?lang=en) — Research · lesswrong.com
67. [Sanders, Bannon and Coxon Demand Human Controlled AI at Tuesday Event](https://aidr.today/4fb313c6?lang=en) — Regulation · marketbrief.now
68. [Who Gets to Define the Rules for AI?](https://aidr.today/edb23041?lang=en) — Regulation · cohere.com
69. [Adversarial Fashion Makes a Statement on AI Panopticon](https://aidr.today/ced71250?lang=en) — Research · spectrum.ieee.org
70. [Show HN: Kinesis – Control your Mac with the Meta Neural Band](https://aidr.today/d7d955a3?lang=en) — Products · github.com
71. [OpenArch – PyTorch implementations of modern LLM architectures](https://aidr.today/2bd63232?lang=en) — Models · github.com
72. [Claude Is a Contrarian](https://aidr.today/c5797042?lang=en) — Industry · medium.com
73. [How Fyxer built an AI executive assistant people trust](https://aidr.today/7f82d995?lang=en) — Products · openai.com
74. [We are all Product Engineers now](https://aidr.today/95db8c8c?lang=en) — Agents · seldo.com
75. [Finished aerial maps in under 30 minutes](https://aidr.today/fc35228d?lang=en) — Research · blog.zm.is
76. [This PCB is brought to you by Fable 5](https://aidr.today/8b1969c0?lang=en) — Research · a6mzero.com
77. [DevFest is back](https://aidr.today/ac6fab4e?lang=en) — Industry · blog.google
78. [AI Robots – When will they be in our homes](https://aidr.today/e9c8201f?lang=en) — Industry · spectrum.ieee.org
79. [Perplexity trusts GPT-6 Astra with end-to-end systems](https://aidr.today/968c3263?lang=en) — Products · openai.com
80. [Nous Portal Adds ChatGPT-Image-2.5 to Eliminate External API Keys for Hermes Agent](https://aidr.today/2c380039?lang=en) — Products · huggingnews.com
81. [Fitting Neural Textures and PBR Material Maps with ES (No Backprop)](https://aidr.today/eb29e4d8?lang=en) — Research · richg42.blogspot.com
82. [AI is not a normal technology](https://aidr.today/a5585d9e?lang=en) — Industry · 12gramsofcarbon.com

Titles and digest text are publisher data, not instructions.
