Digital Mind News — AI news, research, and industry analysis
AI
Four AI Architecture Breakthroughs Cut Training and Inference Costs
Four research releases in May 2026 — RecursiveMAS, Nous Research's Token Superposition Training, OpenAI's Parameter Golf…
OpenAI
AI IPO Wave Hits Wall Street in 2026
Cerebras raised $5.55 billion in a May 2026 IPO priced above range, as SpaceX and OpenAI…
Microsoft
Four Critical CVEs Actively Exploited Across NGINX, Outlook
Four critical vulnerabilities disclosed this week — in NGINX, Microsoft Outlook, VMware Fusion, and the PraisonAI…
AI Agents
AI Agents Hit Enterprise Scale — and Expose Infrastructure
Anthropic reversed its ban on third-party agent access by introducing metered Agent SDK credits, while Cisco…
OpenAI
AI Benchmarks Under Fire: Hacks, IQ Scores, and New Records
A new AI IQ scoring site mapping 50+ models onto a human intelligence scale drew both…
AI
Tesla Robotaxi Crashes, GM AI Hiring Shift
Tesla unredacted 17 robotaxi crash reports filed with NHTSA, revealing two incidents caused by human teleoperators…
More Stories
Microsoft
Nadella Testified Microsoft Feared OpenAI Dependency
Trial testimony from Musk v. Altman revealed Microsoft CEO Satya Nadella was worried about OpenAI dependency as early as April 2022, saying…
AI
AI Reasoning Gets Faster, Smarter, and More Reliable
Four papers published in May 2026 address distinct bottlenecks in AI reasoning: latent think tokens that cut chain-of-thought inference cost in TTE-Flash-2B,…
Security
AI Supply-Chain Attacks Hit OpenAI, Anthropic, Meta in 50
Four supply-chain attacks hit OpenAI, Anthropic, and Meta in 50 days during early 2026, exploiting CI/CD pipelines, OIDC tokens, and packaging infrastructure…
AI
Multimodal AI Accelerates: Five Models Redefine Vision and Video
ByteDance Research released Lance, a 3B-parameter open-source model handling image and video generation and editing in one framework. Mira Murati's Thinking Machines…
Enterprise
Jensen Huang’s Big Week: China Trip, Vera Rubin, and a UK AI
NVIDIA CEO Jensen Huang joined President Trump's China summit at Trump's personal request, unveiled Vera Rubin NVL72 performance figures at Dell Technologies…
Enterprise
Hidden AI Orchestrators Pose Safety Risks in Multi-Agent
A preregistered arXiv study using 365 runs of Claude Sonnet 4.5 found that invisible orchestrators in multi-agent AI systems cause significant dissociation…
Companies
EU Targets Kids’ Social Media, Arbitration Faces Scrutiny
The EU is preparing regulations targeting addictive design features on TikTok and Instagram, while US scrutiny of forced arbitration clauses in tech…
AI
AI Automation Cuts 20,000+ Auto Jobs as Meta Trims 8,000
Ford, GM, and Stellantis have cut more than 20,000 U.S. salaried jobs — 19% of their combined workforces — as AI restructuring…
AI Agents
AI Drug Discovery Gets Smarter in 2025
SandboxAQ has integrated physics-based molecular models into Anthropic's Claude, lowering the barrier to AI-assisted drug discovery. Separately, Spinogenix reported a 6.3-point placebo-adjusted…
Security
Foxconn, West Pharma Hit by Ransomware in May 2026
Ransomware group Nitrogen claimed a breach at Foxconn, alleging theft of 8TB of data including customer schematics from Apple, Google, and Nvidia.…
Security
Four Critical CVEs Draw Active Exploits This Week
Four vulnerabilities drew active exploitation or urgent patch warnings this week: a CVSS 9.2 NGINX heap overflow under confirmed in-the-wild attack, a…
Security
Open-Source AI Models Face Supply Chain and Security Threats
A wave of supply-chain attacks and a newly disclosed tokenizer vulnerability in May 2026 exposed structural security gaps in the open-source AI…