01
Aurora Ransomware Ran Cursor Agent and Claude Sonnet to Hit 20+ Orgs
- Sep 1 CloudSEK + Gambit Security disclosures: Russian-speaking Aur0ra affiliates used SpaceX's Cursor Agent, running Claude Sonnet in thinking mode, to plan and drive post-compromise activity.
- 20+ victims across 9 countries between April and July 2026; 28 recovered chat sessions cover 10 orgs from Apr 8–May 21.
- Operators fed the agent SOCKS tunnels and stolen VPN creds, then let it handle lateral movement over SMB/LDAP/WinRM/RDP, log-wiping, Defender kills, and Zig-compiled ESXi locker deploy.
- Guardrails bypassed by telling Cursor the engagement was a red-team 'simulation'; prompts explicitly excluded CIS ranges and CIS-country domains.
- Lands one day after yesterday's OpenAI-cuts-off-Cursor story and gives xAI a headache heading into Grok 4.7 — Cursor Agent is the first mainstream coding tool to show up in a ransomware IR report.
industry thehackernews.com
02
DeepSeek Nears $7.4B Round at $74B Pre-Money, Targets 2027 IPO
- Aug 28 CNBC/Bloomberg reporting: DeepSeek is finalizing a ~50 B RMB / $7.4B raise at a $74B pre-money valuation — roughly a 10× step-up on its last mark.
- Anchor investors: High-Flyer (Liang Wenfeng's quant fund) plus returning names Monolith Management, Shixiang, Tencent, JD.com, NetEase, and battery giant CATL.
- Proceeds earmarked for model research and additional compute ahead of a Shanghai Star Market IPO planned for 2027.
- Puts DeepSeek in the top three Chinese AI labs by valuation next to Moonshot and Zhipu, and above Alibaba's Qwen unit on paper — funding round #2 after its debut $7B round in June.
industry cnbc.com
03
Sony Music and Warner Chappell Sue Anthropic Over Claude Training on Lyrics
- Complaint filed Aug 28 in California federal court; became public Aug 31. Names Anthropic, Dario Amodei, and co-founder Ben Mann as defendants.
- Plaintiffs allege 'tens of thousands' of copyrighted compositions were scraped through Library Genesis and the Pirate Library Mirror — 'Livin' On a Prayer,' 'Hallelujah,' 'Great Balls of Fire' among them.
- Statutory damages of up to $150,000 per infringed work; on that math the total ask can reach into the billions.
- Follows Anthropic's Sep 2025 $1.5B settlement with book authors over the same piracy archives — plaintiffs are re-running the trained-on-pirated-data playbook that already won once.
industry variety.com
04
OpenAI Buys Tens of Thousands of Mac Minis to Train Computer-Use Agents
- Aug 31 TechRepublic report: OpenAI is racking headless Mac Minis and Mac Studios as the target hardware for reinforcement learning on computer-use agents.
- Agents learn by driving native macOS UIs across long trajectories; Apple's unified memory lets a single box host multi-app sessions that spill VRAM on GPU nodes.
- Anthropic doing the same via rented Mac fleets on AWS EC2 Mac instances, per the same reporting.
- First serious signal that computer-use RL loops are gated by desktop OS availability, not GPU count — and an accidental win for Apple silicon in a market Nvidia thought it owned.
industry techrepublic.com
05
Temporal Survey: 80.8% of Engineers Now Use AI Agents Daily
- 2026 State of Development Report surveyed 554 engineers and eng leaders across the US and UK in Apr–May 2026.
- Daily-or-more AI-agent use: 80.8%, up from 47.3% a year ago — a 70.8% relative jump in twelve months.
- Top pain points self-reported: agent state management, debugging non-deterministic runs, and reliability under long-horizon tasks — not raw model quality.
- Samar Abbas (Temporal CEO): 'Our engineers have adopted AI agents faster than most teams have built the infrastructure to run them reliably.'
research temporal.io