Reading up on Harvey

As AI models commoditize measurable tasks, lasting value lies in 'untrainable' work requiring private data, trust, organizational change, and domain-specific authority.

Anthropic open-sources a suite of Claude plugins covering 60+ legal workflows, deployable as desktop agents, CLI skills, or managed API agents.

Claude Managed Agents adds dreaming for self-improvement through memory review, outcomes for quality grading, and multiagent orchestration for parallel task delegation.

As AI model release cadence accelerates to every 42 days, companies like Harvey, Glean, and ElevenLabs are expanding from single workflow tools into broad platforms, reversing the SaaS-era unbundling strategy.

Sequoia argues the next $1T company will sell work as an AI-native service, not a tool, capturing the larger services budget by automating intelligence-heavy tasks.​

The AI infrastructure bubble is peaking in 2026, and value will shift to application-layer companies that leverage cheap, abundant compute to transform industries.