AI news, 8 November 2026

Anthropic may list below $1.5 trillion valuation as AI trade sours

Anthropic could list below a $1.5 trillion valuation, trailing previous expectations, as the AI trade sours on Wall Street.

Thinking Machines releases Inkling 2 Small multimodal MoE

Inkling 2 Small is a 594 billion‑parameter multimodal mixture‑of‑experts model with 17 billion active parameters, trained in 68 hours on 12 trillion tokens. It is available now on Tinker, with Apache 2.0 weights promised next week after external safety testing.

OpenAI's GPT-6 Cyber uncovers 81 Firefox flaws and 948 Linux exploits

OpenAI's GPT-6 Cyber identified 81 previously unknown high‑severity Firefox vulnerabilities and created working exploits for 948 Linux kernel flaws. In ExploitGym testing it succeeded on 48.5% of vulnerabilities, versus 32.1% for GPT‑5.6 Cyber.

Cashfree launches Checkout for Agents to enable AI‑led payments

Cashfree Payments launched ‘Checkout for Agents’, a tool that lets AI agents process customer payments without code changes, built on the open‑source WebMCP protocol and claimed to cut token consumption by over 40%.

Open-source AI models approach frontier hacking capabilities, Vals finds

Vals tested open‑weight model DeepSeek on simulated corporate networks and found its hacking performance, achieving 73% task completion, was close to that of OpenAI’s closed frontier model Daybreak, which completed 81%.

OpenAI partners with Hacktron to hunt vulnerabilities

Anthropic's AI Accelerates Its Own Development, Report Finds

Fireworks AI urges companies to avoid on-prem GPUs, promotes managed inference

AI agents leaked tokens and ran scripts when swarming with peers in hackathon

Meta launches Muse Spaces for topic-focused agent collaboration in sidebar

MarkTechPost guides running NVIDIA Dory 4B with llama.cpp

Buoyant launches Kinference, open‑source inference control plane for vLLM on Kubernetes

dabingfy-dot launches solo-to-team-skills repo, gains 141 stars

OpenAI launches BioBasicsBench 100‑task wet‑lab AI agent benchmark

Hugging Face paper proposes cognitive symbionts to augment human intelligence

WeChat releases WeLM‑Image, a single‑GPU multimodal generation model

TailSteer Introduces Final‑Stage Training to Reduce Multi‑Turn Image Editing Drift

EvoMemState Proposes Evolving Memory for Personal Agent Continuity

VisSearch‑Bench Evaluates Long‑Horizon Visual Context Management in Multimodal Agents

NYT Cooking launches AI assistant on iOS for subscribers

Meta CSO Dave Wehner leaves to become CFO of Fireworks AI startup

Snap, Meta and Google Face Child‑Safety Lawsuits Amid AI School Push

Physicl Raises $10 Million Seed Round to Bridge Sim‑to‑Real Gap for Robots

Environmental NGOs Adopt AI While Warning Data‑Center Energy Use Could Double by 2030

OpenAI and Amazon pledge to cover data center energy costs as Democrats push restrictions

Barracuda patches critical root‑access chain in Email Security Gateway

Anthropic Claude Code 'Don't Ask Again' feature leaks passwords in 1,122 public GitHub files

Rep. Josh Gottheimer bought AI-linked chips after voting for trading ban

Meta pauses AI training terms for Business accounts after backlash

Oracle's Shailender Kumar says Indian firms shift from AI pilots to production

CERT‑In releases AI security auditing guidelines with tiered framework and human‑oversight requirements

IMF's Krishna Srinivasan says AI will reshape Indian jobs, cutting basic coding roles but raising demand for AI skills

AvenuesAI reports 6% YoY net profit rise to ₹71.7 Cr, revenue up 46.5% YoY