AI will handle the entire software development process end-to-end within 6-12 months (by late 2026) — humans relegated to reviewer/editor role.

AI will handle the entire software development process end-to-end within 6-12 months (by late 2026) — humans relegated to reviewer/editor role. | Claude Code + Devin-class agent maturity

Key catalyst: Claude Code + Devin-class agent maturity

Watch events: Agentic SWE benchmarks (SWE-bench, Terminal-bench); CS hiring data; Big Tech engineer headcount trajectories.

Claude Code + Cursor + Cognition Devin already handling substantial portions of SDLC; autonomous PR-merging agents in production at major firms 2026.

{
  "nia": false,
  "qty": "end-to-end SWE",
  "mode": "FORECAST",
  "role": "Cited-CEO",
  "context": "Amodei early-2026 forecast; goes beyond code assistants to full SDLC autonomy. Structurally disrupts engineering labor economics.",
  "to_year": 2027,
  "conv_cues": "will be; CEO; specific timeframe",
  "direction": "HAPPEN",
  "from_year": 2026,
  "timeframe": "6-12 months (by late 2026)",
  "conv_level": "HIGH",
  "milestones": [
    {
      "kind": "prereq",
      "label": "Recursive self-improvement is already happening now (no longer three years out)",
      "status": "hit",
      "weight": 0.5,
      "ordinal": -9,
      "source_id": "238_009",
      "expected_date": "2026-04-29",
      "observed_date": "2026-04-29"
    },
    {
      "kind": "quartile_checkpoint",
      "label": "Q1 window check-in (25%)",
      "status": "overdue",
      "weight": 0.05,
      "ordinal": -8,
      "source_id": null,
      "expected_date": "2026-05-04",
      "observed_date": null,
      "miss_emitted_at": "2026-05-12T22:09:45.491809+00:00",
      "miss_emitted_by": "metadata_milestone_sweep"
    },
    {
      "kind": "llm_pre_event",
      "label": "SWE-Bench Verified hits 90%+ score",
      "notes": "SWE-Bench Verified is the canonical 'AI handling software dev' benchmark. 90% indicates near-human performance on real GitHub issues.",
      "source": "Anthropic blog, OpenAI evals page, Papers With Code SWE-Bench page, METR evaluations",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -7,
      "source_id": null,
      "confidence": 0.75,
      "expected_date": "2026-07-01",
      "research_origin": "training",
      "expected_date_range": {
        "to": "2026-09-30",
        "from": "2026-04-01"
      },
      "measurement_criterion": "Top model score on SWE-Bench Verified reaches ≥90% on the public leaderboard (current top is in the 70-80% range as of late 2025)"
    },
    {
      "kind": "quartile_checkpoint",
      "label": "Q2 window check-in (50%)",
      "status": "pending",
      "weight": 0.05,
      "ordinal": -6,
      "source_id": null,
      "expected_date": "2026-09-04",
      "observed_date": null
    },
    {
      "kind": "llm_pre_event",
      "label": "AI agent demonstrates ≥1 workday autonomous task completion",
      "notes": "METR's 'task length' evaluation is the canonical agent-autonomy metric.",
      "source": "Anthropic blog, Cognition (Devin) blog, METR.org evaluations, Apollo Research reports",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -5,
      "source_id": null,
      "confidence": 0.65,
      "expected_date": "2026-09-15",
      "research_origin": "training",
      "expected_date_range": {
        "to": "2026-12-31",
        "from": "2026-06-01"
      },
      "measurement_criterion": "Public demo or third-party evaluation (METR, Apollo Research) shows an AI agent (Devin-class, Claude Code-class) completing tasks requiring ≥8 hours of senior-engineer work without human intervention"
    },
    {
      "kind": "llm_pre_event",
      "label": "Major tech company discloses majority of new code is AI-generated",
      "notes": "Sundar Pichai stated 25%+ at Google in Oct 2024. Trajectory implies 50%+ within 2 years.",
      "source": "Earnings call transcripts, GitHub Octoverse report, Stack Overflow Developer Survey, Microsoft/Google blog posts",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -4,
      "source_id": null,
      "confidence": 0.55,
      "expected_date": "2026-11-29",
      "research_origin": "training",
      "expected_date_range": {
        "to": "2027-04-30",
        "from": "2026-07-01"
      },
      "measurement_criterion": "Public statement (earnings call, blog, internal memo leak) from top-5 tech company (Microsoft, Google, Meta, Amazon, Apple) that >50% of new code is AI-generated. Or developer survey shows similar at industry level."
    },
    {
      "kind": "quartile_checkpoint",
      "label": "Q3 window check-in (75%)
... (truncated)