Q1 2026 Timelines Update
Researchers now predict AI will replace human software engineers by 2028, 1.5 years sooner than expected.
The AI Futures research team, including Daniel Kokotajlo and Eli, has published a quarterly update to their AI timelines forecast, significantly accelerating predictions for when AI will automate software engineering. Their median estimate for achieving 'Automated Coder' (AC)—the point where an AGI company would fire all human software engineers in favor of AI—has moved forward from late 2029 to mid-2028, a shift of about 1.5 years. The forecast for 'Top-Expert-Dominating AI' (TED-AI), an AI that outperforms top human experts in virtually all cognitive tasks, has also moved forward by a similar margin.
Key drivers for the accelerated timeline include analysis of newly evaluated models like Gemini 3, GPT-5.2, and Claude Opus 4.6, which showed faster-than-expected progress in agentic coding capabilities. The team revised their estimate for the doubling time of the METR coding time horizon metric down from 5.5 months to 4-4.5 months. Real-world adoption also played a role: Claude Code reportedly reached over $2.5 billion in annualized revenue just nine months after launch, indicating explosive growth in coding agent utility. Furthermore, private discussions with AI company researchers suggest some insiders believe automated AI R&D is coming even sooner than these public forecasts.
- Forecast for 'Automated Coder' milestone moved from late 2029 to mid-2028, a 1.5-year acceleration.
- Analysis based on new models (GPT-5.2, Claude Opus 4.6) shows METR coding time horizon doubling every ~4 months.
- Claude Code reached $2.5B annualized revenue, signaling rapid real-world adoption of coding agents.
Why It Matters
This forecast suggests the displacement of human software engineers could begin within two years, forcing urgent strategic planning for tech companies and professionals.