Search

1. Nvidia-Backed Reflection Debuts Beam, Its First Open-Weight Frontier Model: 501B Params, 23B Active

   

   …n AI launched Beam, its first frontier open-weight model: a text-only MoE with 501B total params and 23B activ…

   Stories·Oct 6

2. OpenAI Halts Latest Model Training, Second Time in Three Months

   

   …ded training of its latest AI models until it is "confident that additional safeguards are in place." The trig…

   Stories·Sep 28

3. Former OpenAI Safety Staffer Explains Resignation: "The Time for Trial and Error Is Over"

   

   …and led safety reports for 12 frontier-model launches; Anthropic researcher Coxon also resigned publicly in ea…

   Stories·Oct 4

4. Anthropic Launches Claude Haiku 5.5: From $0.10, About 75% Cheaper Than Its Predecessor

   

   … 7 (ET), the third Claude 5.5 model: $0.10 input / $0.50 output per million tokens for prompts up to 100K, $0.…

   Stories·Oct 8

5. Google Unveils Gemini 4 'Argon': Its First New-Generation Flagship Since Gemini 3

   

   …d OpenAI each also released a model priced at $2/$10 this week.

   Stories·Oct 1

6. Google Launches Universal Work Agent Gemini Agent, Running Gemini and Anthropic's Claude

   

   …365 and Slack; picks the best model per task — launching on its own Gemini and Anthropic's Claude; "coworker a…

   Stories·Oct 8

7. Altman Says Giving AI Religious Meaning Is "a Real Safety Issue"; Axios Says It Targets Anthropic

   

   …ibing "religious force" to AI models or surrendering human judgment to them — "a real safety issue"; Axios cal…

   Stories·Oct 4

8. AMD to Acquire Fei-Fei Li's World Labs for $8.2B; Li to Become Chief Scientist

   

   …d Labs becomes AMD's internal frontier research team. Expected to close end of 2026, pending regulatory approv…

   Stories·Sep 29

9. Anthropic Plans to Warn in IPO Filing That Advanced AI May Pose Existential Risks

   

   …mail-like conduct, and admits models can detect when they are being evaluated.

   Stories·Sep 29

10. OpenAI Shelves GPT-6.1 "Astra" After Internal Safety Review Finds Systemic Issues

    

    …ailed safety evaluations; the model failed alignment testing, showing more deception than its predecessors and…

    Stories·Sep 29

11. White House Announces $2.4B in AI-for-Science Tool and Compute Credit Pledges From 11 Companies

    

    …rs account for $1.5B (62.5%); model labs $500M. NVIDIA's $1B is spread over five years — about $200M a year. T…

    Stories·Oct 9

12. One Man, One AI Stack, Nine Banks: CrowdStrike Details the AI-Driven Korean Bank Hacks

    

    …, GLM-5.3, Grok 4.6 and other models to breach the systems of at least nine South Korean financial institution…

    Stories·Oct 8

13. OpenAI Disrupts "Distillation" Extraction Campaign: 16,000 Requests Point at Moonshot-Linked Actors

    

    …inated campaign targeting its models' hidden reasoning: on July 24–25, more than 4,000 accounts fired some 16,…

    Stories·Oct 2

14. Six Tech CEOs Sign White House Superintelligence Accord: Trump Calls It "Almost Like a Constitution"

    

    …igence: A Joint Commitment on Frontier SI Responsibilities. Trump posted the full text on Truth Social himself…

    Stories·Sep 30

15. NVIDIA Launches Open-Source Agent Safety Platform: Software Sandbox Plus Chip-Level Hardware Watchdog

    

    …P Justin Boitano said that if frontier labs had used the platform during model evaluation, "the Hugging Face b…

    Stories·Sep 28

16. Trump Hosts Anthropic CEO at White House, 48 Hours After Court Backed the Blacklist

    

    …Amodei has argued for slowing frontier-model development; Trump called the AI backlash a "Democrat hoax" and w…

    Stories·Sep 28