Tuesday 28 July 2026
Top stories today:
Anthropic’s Amodei rejects open AI model ban, calls for testing
Kimi K3 weights released by China’s Moonshot AI days after debut
Microsoft launches MAI-Cyber-1-Flash, Perception security system
NVIDIA to invest $5B in Ilya Sutskever’s Safe Superintelligence
Elon’s X launches X Money for U.S. Premium subscribers
0. Data & Calendar
All values as of 6 AM ET / 3 AM PT, other than S&P500 and NASDAQ close (4 PM ET / 1 PM PT).
All times are ET.
Listen to our AI-generated podcast summarizing today’s newsletter (beware of hallucinations):
1. Anthropic’s Amodei rejects open AI model ban, calls for testing
Policy interventions to slow China’s AI development, Dario Amodei urged the U.S. to pursue.
He also urged tighter chip export controls on China and a crackdown on industrial-scale AI distillation.
Relatedly, Sam Altman is meeting with U.S. officials in Washington, DC, this week to preview OpenAI’s upcoming AI models.
South Korea’s Kospi slid to its lowest since Apr. as AI spending concerns and China’s chip advances hit semiconductor stocks
China warned it would take “all necessary measures” if the U.S. sanctions Chinese AI companies.
AI agents are only as powerful as the context they can access. A new IDC white paper explains why conversational context is the defining variable for AI agent success.
52% of developers report their organizations have adopted AI agents, and 66% have moved them into production.
Download the report to learn why Slack is positioned to become the coordination layer for multi-agent work.
*Sponsored.
3. Kimi K3 model weights released by China’s Moonshot AI days after debut
Kimi K3 has become the highest-ranked open-weights AI model in the Artificial Analysis Intelligence Index:
Under the Kimi K3 License, the model weights are available for download through Hugging Face.
Moonshot’s daily sales have surged 6x since K3’s debut earlier this month.
Fill key roles faster, without waiting weeks or months for the pipeline to behave.
AI-assisted matching and structured vetting help you find people who fit the role and the team.
Cut time and cost with a process built for scale, not endless back-and-forth.
*Sponsored.
5. Microsoft launches MAI-Cyber-1-Flash cybersecurity model, Perception security system
MAI-Cyber-1-Flash, running with GPT-5.4, scored 96% on the CyberGym AI cybersecurity benchmark:
Microsoft didn’t provide the model to independent testers before its release.
The model is initially limited to Microsoft’s DASH customers.
Aug. 3: Project Perception enters public preview, with three sets of agents to defend against AI-driven attacks.
Get the chance to peek inside founders and leaders’ brains and see how they think about going from zero to 1 and beyond.
Join thousands of weekly readers at Google, OpenAI, Stripe, TikTok, Sequoia, and more.
Check all the tools and more here, and outperform the competition.
*Sponsored.
7. NVIDIA said to invest $5B in Ilya Sutskever’s Safe Superintelligence
NVIDIA’s next-generation Vera Rubin platform will be deployed at Safe Superintelligence, the companies said, without disclosing financial terms.
