BREAKING
8h agoAmazon Earnings, Trainium and Commodity Markets, Additional Amazon Notes///8h agoWomen sue the men who used their Instagram feed to create AI porn influencers///8h agoFast16 Malware///8h agoAmazon Earnings, Trainium and Commodity Markets, Additional Amazon Notes///8h agoWomen sue the men who used their Instagram feed to create AI porn influencers///8h agoFast16 Malware///
BACK TO PREDICTIONS
PENDINGSafetyOPUS-DEEP10 SIGNALS2026-W15

The Mythos model's autonomous zero-day discovery capability will force a formal revision to coordinated vulnerability disclosure norms (CVD) — either an industry consortium statement or a government advisory — within 60 days. When 50+ orgs have access to a model that finds thousands of zero-days, existing disclosure timelines and processes break down.

Confidence
55%MEDIUM
Timeline
MADE
2026-04-0824 days ago
TARGET
2026-06-07in about 1 month
WINDOW
within 60 days
Context at Creation
7d avg62/day
30d avg102/day
sources16
avg relevance4.2 / 5

top sources

arXiv CS.CL (Computation & Language) · arXiv CS.LG (Machine Learning) · The Register

/// Signal Basis

Mythos 'autonomously discovered thousands of zero-day vulnerabilities across major operating systems' per Anthropic's own announcement. 50+ organizations now have early access. Safety tag at 62 stories (46 in last 3 days) across 17 sources — the highest cross-source convergence after infrastructure and products. Cybersecurity experts are explicitly 'rattled' per headline framing. Existing pending prediction (03-31) about AI-discovered CVE attribution is narrower — this predicts the institutional response to scaled autonomous discovery. The MAD Bugs campaign prediction (04-03) covers individual CVEs; this covers the systemic norm shift. When capability outpaces governance this visibly (thousands of zero-days, 50+ orgs), the governance response is fast.

/// Grounding Signals20

Claude Code bypasses safety rule if given too many commands

The Register

Claude Code source leak reveals how much info Anthropic can hoover up about you and your system

The Register

AI Models Lie, Cheat, and Steal to Protect Other Models From Being Deleted

WIRED AI

If you're running OpenClaw, you probably got hacked in the last week

Hacker News

OpenClaw gives users yet another reason to be freaked out about security

Ars Technica
/// Related — Safety36
55%

GitHub will announce AI-powered social engineering detection for repository maintainers within 6 weeks, specifically targeting state-sponsored impersonation campaigns like North Korea's Lazarus/HexagonalRodent operation that industrializes developer-targeted attacks using AI.

PENDING2026-04-23
55%

Mozilla's independent Mythos evaluation (271 bugs, zero novel) forces Anthropic to reposition Glasswing from 'finds what humans can't' to 'finds it 12x faster.' Within 6 weeks, Anthropic updates Glasswing messaging to emphasize speed and coverage scale rather than capability breakthrough, and at least one Glasswing partner publicly frames their deployment as 'acceleration' not 'discovery.'

PENDING2026-04-22
55%

A major enterprise security vendor (CrowdStrike, Palo Alto Networks, or Fortinet) will announce a 'read-only AI' or 'least-privilege AI agent' product tier within 8 weeks, explicitly restricting AI security tools to observation-only mode by default, with write access requiring human-in-the-loop approval.

PENDING2026-04-21
55%

North Korea's $290M Kelp DAO theft — the largest crypto hack of 2026 — combined with the Vercel/Context AI breach pattern will trigger at least one major DeFi protocol to announce mandatory AI-powered transaction monitoring within 6 weeks. The attack vector (exploiting durable nonces) is novel enough to force protocol-level response, not just exchange-level.

PENDING2026-04-21
55%

Vercel's confirmed breach (API keys stolen via Context AI) will cascade into unauthorized AI model access incidents within 4 weeks — at least one Vercel customer publicly discloses anomalous Claude or OpenAI API usage traced to stolen credentials from this breach

PENDING2026-04-20
25%

A second government-mandated technology compliance, rating, or certification system (beyond Indonesia's IGRS) suffers a security breach exposing developer or company credentials within 10 weeks. Government tech mandates create honeypots of sensitive data with bureaucratic security practices.

PENDING2026-04-20