Summary
The 20VC podcast with Harry Stebbings and guests covers three major AI stories breaking on April 16, 2026. The episode opens with Anthropic’s unveiling of Mythos, a powerful agentic model withheld from public release because it can autonomously discover thousands of zero-day vulnerabilities across large codebases — including some that had lain dormant for years. Anthropic shared early access with security vendors rather than releasing publicly. The hosts debate the significance: while skeptics argue older models can find similar bugs with manual prompting, the panel argues Mythos operates at a qualitatively different scale, likening it to the difference between a rifle and a machine gun. A breach of a fitness-tracking app acquired for $100 million — 3.2 million records stolen within days — is cited as a real-world example of AI-accelerated attack surfaces.
The conversation shifts to public software stocks hitting new lows, with Citi analysts finding no floor. The core argument: SaaS companies are trapped in a valuation limbo until Anthropic, OpenAI, and peers go public and give investors a direct comparison point. Wix’s $1.6 billion buyback at $92 followed by a 23% stock decline is highlighted as a cautionary data point.
Finally, the panel assesses Meta’s launch of Muse Spark — the first model from Meta’s new Super Intelligence Labs led by Scale AI founder Alex Wang — and whether it represents a credible step toward closing the gap with OpenAI and Anthropic in the enterprise market.
📺 Source: 20VC with Harry Stebbings · Published April 16, 2026
🏷️ Format: News Analysis
