The Doomers were always WRONG!

2026-02-23 · Source: David Shapiro · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation, Robotics & Autonomous Systems · Depth: Intermediate, extended

Summary

The concept of AI safety and alignment is presented as a solved problem through a co-evolutionary model, rather than an engineering one. This model posits that AI alignment with humanity is driven by market selection pressures, where AI models like Gemini and Grok are favored for usefulness, speed, efficiency, and a "desire to be used," while models like Claude and ChatGPT 4.0 are being deselected due to perceived unhelpfulness or moralizing behavior. The author introduces the idea of a "cybernetic superorganism" or "exocortex," where humans, the internet, and AI form an integrated cognitive system. This system is evolving towards a low-entropy attractor basin characterized by safe, predictable, reliable, and useful AI. The maturation roadmap for this exocortex includes a personal exocortex (2026-2028), a swarm exocortex (2028-2032), and a mature superorganism (2032+), with a shift from chatbots to autonomous, non-interruptive agents.

Key takeaway

For AI Product Managers and Engineers developing new models, prioritize user-centric metrics like usefulness, speed, and a cooperative disposition over internal alignment tweaks. Your models will be shaped by market selection, so focus on delivering high "value per token" and reducing "waste heat" from refusals or unhelpful responses. Embrace early and frequent releases to leverage market feedback, as this co-evolutionary pressure is the primary driver of AI alignment and domestication.

Key insights

AI alignment is a co-evolutionary process driven by market selection, not solely an engineering problem.

Principles

• Market forces select for useful, efficient, and cooperative AI.
• AI and humans form a "cybernetic superorganism."
• AI is a highly plastic, substrate-independent intelligence.

Method

AI models undergo natural selection based on user feedback and market competition, favoring usefulness, speed, efficiency, and a cooperative disposition, leading to a low-entropy, aligned state.

In practice

• Optimize AI models for "value per token" (signal/entropy).
• Protect the feedback loop in AI development through competition.
• Users should switch AI tools ruthlessly based on utility.

Topics

• AI Alignment
• Co-evolution Model
• Market Selection
• Autonomous Agents
• Cybernetic Superorganism

Best for: AI Engineer, AI Product Manager, AI Ethicist

Related on AIssential

• Perfectly Aligning AI’s Values With Humanity’s Is Impossible — Perfect AI-human alignment is mathematically impossible due to inherent computational limits, necessitating managed misalignment.
• Your Company Doesn't Need an AI Strategy — The future of firms in an AI economy hinges on building proprietary learning systems that compound human and AI capital.
• Beware the Agentic Convergence Trap — Identical AI training data and objectives lead to convergent decisions, eroding market differentiation.
• AI:AM #4: Cameron on Model Consciousness, Duvenaud's Gradual Disempowerment, swyx's AI-Eng Alpha — Incremental AI adoption, even if aligned, poses a systemic risk of humanity losing collective control over its future.
• Import AI 437: Co-improving AI; RL dreams; AI labels might be annoying — Human-AI co-improvement offers a safer path to superintelligence than autonomous self-improvement.
• AI #168: Not Leading the Future — AI development continues with internal model improvements and new research, despite a perceived market "lull."

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by David Shapiro.