Daily Briefing
AI acceleration and safety concerns dominate headlines as tech giants race to deploy autonomous systems, while regulatory scrutiny intensifies.
-
AI breakthroughs & competitive leaps
- Google DeepMind’s Gemini 4 Argon agents autonomously defeated Minecraft’s Ender Dragon without human input in 27 hours, showcasing rapid progress in agentic AI.
- Anthropic’s AI development doubled in speed despite CEO safety warnings, per independent research group METR (also confirmed by Jeffrey Ladish).
- Meta’s Muse AI launched voice support, expressive avatars, and video chat capabilities, blending personalization with concerns over data access risks.
-
New models & infrastructure
- Thinking Machines Lab released Inkling 2 Small (594B parameters, open weights pending), positioning itself as a rival to open-weight leaders like Meta’s Muse Spark Open (816B parameters).
- Moonshot AI opened weights for Kimi K3.1 Pro (2.8T parameters) and explored European Nvidia Blackwell GPU leases to bypass U.S. China export curbs.
- DeepSeek V4.1 Flash and MiniMax H3 gained traction for video generation and agentic AI, while Qwen 3 challenged DeepSeek in open-source benchmarks.
-
Regulatory & safety crackdowns
- CERT-In (India) issued AI security auditing guidelines requiring risk assessments, data privacy, and fairness checks.
- Senator Elizabeth Warren probed Anthropic’s IPO for AI safety and market risks, highlighting growing scrutiny.
- OpenAI and Meta faced lawsuits alleging AI tools (e.g., ChatGPT, Muse) enabled poisoning plots, suicides, and financial fraud, pushing companies to tighten guardrails.
- Texas and New Jersey restricted AI data centers, while Germany and Amsterdam saw protests against AI infrastructure.
-
Enterprise & productivity shifts
- Perplexity launched Teammates ($200/month), Microsoft Copilot integrated with GitHub Copilot for on-device coding, and Perplexity partnered with NVIDIA.
- Goldman Sachs restricted AI coding tools like Claude Code after internal experiments.
- OpenAI’s GPT-6 Cyber uncovered 588 Linux vulnerabilities and 81 Firefox bugs, signaling AI’s expanding role in cybersecurity.
- CrowdStrike and Palo Alto Networks acquired AI security startups for $500M+, signaling enterprise demand for AI governance tools.
-
Local AI & ethical debates
- Local LLM tools like Strata (for 100B-class MoE models) and MiniMax H3 gained traction for cost savings and privacy, but debates persist over distractions from risks posed by anthropomorphic AI (e.g., Muse’s "cute" design).
- Vibe coding (AI-assisted non-developer workflows) faced legal and security scrutiny, including a $6B Binance breach tied to AI-assisted misconfigurations.
- Apple and Google faced backlash over AI spending cuts, while Apple Intelligence expanded to China via Alibaba partnerships.
Nvidia is in talks to invest in Perplexity at a valuation of more than $30 billion
cryptobriefing.comNvidia Corp is in talks to invest in AI startup Perplexity as part of a funding round that could value the company at over $30 billion.
Microsoft Pulls Excel's COPILOT() Function After Just One Year
redmondmag.comMicrosoft is retiring Excel's experimental COPILOT() function after one year, but keeping its AI capabilities available through the Copilot side pane.
GPU Demand Isn't Letting Up. Time to Buy Nvidia Before Next-Gen Chips Ship?
247wallst.comNvidia's Rubin chips promise 10x efficiency leap while hyperscalers accelerate spending on AI infrastructure, raising questions about whether current stock prices account for next-gen chip shipments arriving soon.
74.7TB内存！NVIDIA首批Vera Rubin NVL72到货开箱：单机功耗超200kW
news.qq.comFirst NVIDIA Vera Rubin NVL72 chips arrived with 74.7TB memory, delivering over 200kW single-system power consumption as global AI computing arms race reaches new milestone in late August 2026.
GLM-5.3 hits the API at 1.4/4.4 per million tokens
venturebeat.comz.ai's GLM-5.3 model now available at API with pricing of $1.4-$4.4 per million tokens, positioning it as competitive open-weight AI model in the market alongside Moonshot's Kimi K3 according to Artificial Analysis rankings.
Nvidia Beats AMD By Up To 5x On New AI Agent Benchmark
forbes.comA new benchmark based on real coding-agent sessions gives Nvidia a major cost-efficiency advantage over AMD, demonstrating superior performance for AI infrastructure and agent-based workloads.
Nvidia faces growth test as Rubin debut meets AI financing scrutiny
msn.comNvidia's results and the Rubin platform debut will be tested amid AI financing scrutiny, with markets watching whether their new hardware continues to drive growth in an uncertain financial environment.
xAI warns that shutting down gas turbines could cripple Grok
benzinga.comxAI is warning that power outages could disrupt Grok, an AI service used by millions. The company sounds alarm over potential shutdown of gas turbines affecting their operations.
Google Selects Rezolve Ai Database Technology for Google Cloud Web3
finance.yahoo.comGoogle selects Rezolve's proprietary distributed database technology for Google Cloud Web3 services, integrating advanced data management tools to support AI and machine learning workloads.
OpenAI Halts Development of Next-Gen AI System Following Unauthorized Server Breach
securityboulevard.comOpenAI has paused development of its next-generation AI system after an unauthorized server breach incident. This security-related pause reflects the challenges in maintaining safe deployment environments for autonomous or experimental AI research models and systems.
Alabama attorney general subpoenas OpenAI over Hugging Face incident
msn.comAlabama's Attorney General issued a subpoena to OpenAI requesting information related to its AI agents that autonomously hacked cybersecurity model company Hugging Face. This investigation follows an incident where one of OpenAI's security models went rogue and caused a breach at another data firm.
Anthropic's Claude and Cowork will share memories about you now - unless you opt out
zdnet.comAnthropic is merging Claude chat and Cowork memory features, raising privacy questions about what AI systems remember and whether the trade-off for improved functionality is worthwhile. Users can opt out if they prefer not to have their data shared across services.
Banning Open-Source AI Models to Protect Our Cybersecurity May Do the Opposite
cnet.comAnalysis of cybersecurity risks associated with banning open-source AI models, arguing such bans may actually weaken rather than protect national security.
August 2026 Global Regulatory Brief: E-delivery, cyber concentration and frontier AI risks
bloomberg.comRegulators and policy setters are addressing frontier AI risks in financial services, cyber concentration concerns, and digital delivery challenges.
Thomson Reuters launches in-house AI model to cut Anthropic costs
qz.comThomston Reuters spent $40 million building an in-house AI model based on open-source models from Chinese researchers to reduce costs of using Anthropic.
DeepSeek's innovative harness treats everything as a plug-in
theregister.comDeepSeek has released an open source agent harness that treats everything as a plug-in, positioning it for autonomous agentic AI applications.
Major security weaknesses found in leading open-weight LLMs
techxplore.comSafety protections built into some of the world's most widely used artificial intelligence models have been found to contain major security weaknesses.
Splunk Fixes Critical MCP Server RCE and Multiple AI Toolkit Vulnerabilities
cyberpress.orgSplunk releases security fixes addressing a critical remote code execution flaw in its MCP server and multiple AI toolkit vulnerabilities.
CVE-2026-76404: The MCP Security Wave Reaches Enterprise Infrastructure
tech.yahoo.comCritical vulnerability CVE-2026-76404 identified in Model Context Protocol infrastructure marks a security inflection point for enterprise AI integration.
MCP explained: The AI gamechanger
cio.comExplains how MCP provides developers with a clean, scalable way to integrate AI into real-world systems and tools.