The Shape of the Thing

2026-02-18 · Source: One Useful Thing · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, medium

Summary

The article describes a new phase of AI characterized by exponential improvements in capabilities, shifting from human-AI co-intelligence to managing AI agents. Since late 2025, AI systems like Claude Code, OpenAI's Codex, and OpenClaw can autonomously complete hours of human work in minutes. This rapid progress is demonstrated by the "Otter Test" for image generation, showing remarkable gains from 2022 to 2025, and advanced AI video models from Bytedance. Benchmarks like METR Long Tasks, Google-Proof Q&A (94% score for best AIs), and GDPval (82% human parity) also show steep exponential curves. This advancement enables radical new work models, exemplified by StrongDM's Software Factory, which uses AI agents to write, test, and ship production software without human involvement, with each engineer spending at least $1,000 daily on AI tokens.

Key takeaway

For Directors of AI/ML evaluating future organizational structures, recognize that AI's exponential growth and agentic capabilities necessitate a shift from co-intelligence to AI management. Your teams should explore radical new work models, such as AI-driven software factories, to capitalize on these advancements. The window to proactively shape how AI integrates into your operations is now, setting precedents for broader adoption and competitive advantage.

Key insights

Exponential AI capability growth is shifting work from human-AI collaboration to AI agent management, enabling radical new organizational models.

Principles

• AI capabilities improve exponentially.
• AI agents can autonomously complete complex tasks.
• Recursive self-improvement is an explicit AI roadmap item.

Method

StrongDM's Software Factory uses coding agents to build software from human-written roadmaps, while testing agents simulate customer environments and provide feedback, shipping products without human code review.

In practice

• Implement AI agents for autonomous task completion.
• Experiment with AI-driven software development workflows.
• Allocate budget for AI token usage in engineering.

Topics

• AI Agents
• Exponential AI Growth
• AI Benchmarks
• Software Factory
• Recursive Self-Improvement

Best for: VP of Engineering/Data, Director of AI/ML, Entrepreneur, CTO, Executive, AI Product Manager

Related on AIssential

• The Month AI Woke Up — AI agents have matured to enable autonomous task execution, fundamentally altering programming and impacting market valuations.
• the SCARIEST chart in AI — AI agent capabilities are accelerating, replacing human expert labor at an increasingly rapid pace.
• Building AI-Native Growth Teams — AI-native operations shift bottlenecks from headcount to decision architecture and human-AI orchestration.
• The Power to Shape AI — Human agency is critical to shaping AI's future amidst its exponential growth and disruptive impact.
• Hard Takeoff has started — AI models are now recursively self-improving, accelerating development and reducing human bottlenecks.
• Why change management must become an organizational capability in the AI era — Change management must evolve from temporary programs to an embedded organizational capability to navigate the rapid, autonomous shifts of the AI era.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by One Useful Thing.