The mythos of Mythos and Allbirds takes flight to the neocloud

2026-04-24 · Source: Practical AI · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Cybersecurity & Data Privacy · Depth: Intermediate, extended

Summary

This episode of the Fully-Connected podcast discusses three key AI-related developments. First, Anthropic's unreleased Mythos frontier model, reportedly highly capable in cybersecurity vulnerability discovery, has led to Project Glasswing, a closed initiative with 40 companies to address potential risks. Second, the shoe manufacturer Allbirds has pivoted its entire business to become an AI compute infrastructure provider, rebranding as a "neocloud" company and seeing its shares surge by 700%. This move highlights a trend of non-tech companies entering the specialized AI cloud market. Finally, the hosts explore "tokenmaxxing," a gamified approach to software development where engineers are encouraged to maximize LLM usage, often at significant cost, to boost productivity, raising questions about optimal spending and relevant metrics.

Key takeaway

For CTOs and engineering leaders evaluating AI integration, recognize that advanced models like Mythos necessitate proactive security measures and governance frameworks. Your team's AI tool usage, particularly "tokenmaxxing," should be carefully measured against tangible productivity gains, not just raw token consumption, to ensure cost-effectiveness and avoid outrunning organizational capacity. Additionally, be aware that AI chat logs are not privileged and can be discoverable in legal proceedings, requiring updated internal policies.

Key insights

AI advancements are driving rapid shifts in business models, cybersecurity, and developer workflows.

Principles

• Frontier models can pose significant security risks.
• Market values capital allocation to AI infrastructure.
• High LLM usage aims to accelerate developer productivity.

Method

Anthropic's Project Glasswing involves inviting companies to use the Mythos model in a closed environment to identify and fix system vulnerabilities before wider release.

In practice

• Evaluate AI models for cybersecurity vulnerabilities.
• Consider specialized "neocloud" providers for AI workloads.
• Assess LLM token spending against developer productivity.

Topics

• Anthropic Mythos Model
• AI Cybersecurity
• Allbirds AI Pivot
• Neocloud Infrastructure
• Tokenmaxxing

Best for: Executive, Investor, CTO, AI Engineer, Director of AI/ML, Legal Professional

Related on AIssential

• #209: Claude Mythos, Project Glasswing, Claude Code Leak, OpenAI Raises $122B & the End of Middle Management — Frontier AI models like Anthropic's Claude Mythos demonstrate advanced capabilities with significant security and societal implications.
• Dylan Patel - The Infinite Demand for Tokens, Claude Mythos, and Supply Constraints - [Invest Like the Best, EP.469] — Unbounded demand for frontier AI models is driving explosive token consumption and unprecedented value creation.
• AI #163: Mythos Quest — Advanced AI models like Claude Mythos pose significant cybersecurity risks while demonstrating rapid commercial growth and complex internal states.
• 636: Amazon Tokenmaxxing, Seats + Meters, Musk’s $119bn Terafab, OpenAI's $31bn to Microsoft, Ontario Nuclear Build-Out, Sub-2-Hour Marathon, AI-Built Cyberattack, and Butch Cassidy — Metrics-driven AI adoption can lead to perverse incentives and inflated usage, while AI also accelerates cyber threats.
• Claude Mythos and Project Glasswing: why an AI superhacker has the tech world on alert — Anthropic's Mythos AI model demonstrates unprecedented capability in discovering and exploiting software vulnerabilities, prompting a restricted release for cyber defense.
• Stop ‘tokenmaxxing’ and deploy AI sensibly instead — Unchecked agentic AI adoption and "tokenmaxxing" face severe resource, financial, and human skill development limitations.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Practical AI.