The next 36 months will be WILD

2026-02-26 · Source: David Shapiro · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Data Science & Analytics · Depth: Advanced, extended

Summary

Leading AI figures like Dario Amodei, Jensen Huang, and Sam Altman are converging on a 2027-2028 timeline for Artificial General Intelligence (AGI) and Artificial Super Intelligence (ASI), with recursive self-improvement (RSI) anticipated even earlier, potentially by 2027. This accelerated timeline is supported by super-exponential progress in machine autonomy, exemplified by Claude Opus 4.6's median autonomous success rate of 14.5 hours, with doublings occurring every 90 days. Key pillars for RSI, such as algorithmic research, data generation, and code writing/execution, are largely considered solved or near-solved, while model training and evaluation remain the primary bottlenecks. The "industrial siege" describes an intense, irreversible race among companies and nations, driven by massive investments (e.g., $600 billion in data centers) and a "point of no return" mentality. Current bottlenecks like chips and high-bandwidth memory are being addressed by market forces, but energy demand, projected to reach 500 terawatt-hours for AI, presents a significant challenge, with solutions including microgrids and small modular reactors (SMRs) by 2029-2030. The economic impact, characterized by "Solow's Paradox 2.0" and "ghost jobs," is already manifesting as jobless growth, with entry-level hiring imploding.

Key takeaway

For CTOs and VPs of Engineering assessing future technology roadmaps, the convergence of expert predictions on AGI/ASI by 2027-2028, coupled with super-exponential progress in machine autonomy, necessitates immediate strategic planning. Your organization should accelerate investment in AI integration and automation, particularly in areas like model training and evaluation, to capitalize on the rapid advancements and mitigate risks associated with the "automation cliff" and jobless growth. Prepare for significant economic shifts and potential talent displacement.

Key insights

Expert consensus points to AGI/ASI by 2027-2028, driven by super-exponential AI progress and an irreversible industrial race.

Principles

• AI capabilities follow a super-exponential growth model.
• Economic impact of AI is measured by substitution, not just AGI definition.

Method

Recursive self-improvement (RSI) requires advancements in algorithmic research, data generation/curation, code writing/execution, model training, and model evaluations, with the latter two posing current bottlenecks.

In practice

• Monitor Meter scores for task horizon crossing 24 hours.
• Observe entry-level hiring trends as an indicator of automation's impact.

Topics

• AGI Timelines
• Recursive Self-Improvement
• Machine Autonomy Metrics
• AI Infrastructure
• AI Economic Impact

Best for: Investor, CTO, VP of Engineering/Data, AI Engineer, Data Scientist, Director of AI/ML

Related on AIssential

• Experts give a 50% chance of AGI by 2050, 75% chance by the mid-2060s, and 95% chance by 2090 — Experts predict AGI has a 50% chance by 2050, rising to 95% by 2090.
• From AGI to ASI — The AGI-to-ASI transition is plausible via four pathways, facing significant uncertainties and bottlenecks requiring urgent research.
• AGI in 2025 |Do you think what matters today will still matter in the coming months? TL;DR: No! — AGI's imminent arrival will fundamentally reshape industries, making current business models and workforce dynamics obsolete.
• 🔮 Exponential View #573: Are the AI labs building for an intelligence explosion? — A frontier AI model training its successor by 2028 is a 60% possibility, shifting AI development to an industrial scaling problem.
• The AGI Goalpost Keeps Moving — The lack of a consistent AGI definition creates significant information asymmetry for investors.
• AI Grill Cooks With Light!? — AGI by 2030 requires advancements in continual learning, reasoning, and memory, with agents as a key path.

Open in AIssential →

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