Continual-learning ladder (highest production rung)

The highest rung of the continual-learning ladder (L0 static snapshot to L6 six-month-employee test) with a shipped production system on record, from dated rows citing the vendor's or lab's own description. Derived metric continual_learning_level.

Why it matters. The crux of the timelines argument. If systems learn on the job, deployment feeds invention automatically and the return arrow stops depending on human-data vendors; Narayanan and Kapoor's speed limits assume it stays unsolved.

Proxy types

product, model_release

Unit

rung

Cadence

quarterly

Valve

return arrow

• derived metric continual_learning_level (formula in the semantic layer)

Rows are vendors describing their own systems; the rung is what they claim to ship, not what an independent evaluation found.

2 observations. Hollow points are disputed (see counterevidence). Every point links to its observation.

Highest production rung on record is L3: Trajectory's per-customer LoRA adapters refreshed hourly and A/B-routed behind provenanced endpoints, per Baseten's co-authored post of 27 May 2026; Cursor's Tab model sits at L2 (one online-trained model for all users, 1.5-2 hour checkpoint cycles). L4 (firm knowledge in weights, Engram + Harvey) and L5 (rank-1 LoRA merge, Nested Learning, TTT, self-distillation) exist only as research and do not move the level. L3 is inside the normal band, but every production row is a vendor describing its own system (tier 7), so the status is capped at emerging. Initial seed.

The tracker's prior expectation was consistent with normal; the evaluator reads emerging. The evaluator wins until a reviewed override.

• 2026-05-27 · 3.00 rungas of 2026-05-27(2 obs)
• 2025-09-12 · 2.00 rungas of 2025-09-12

1. 2026-09-10○unmeasured →to ◐emergingconf — → 45 · evaluate

   Highest production rung on record is L3: Trajectory's per-customer LoRA adapters refreshed hourly and A/B-routed behind provenanced endpoints, per Baseten's co-authored post of 27 May 2026; Cursor's Tab model sits at L2 (one online-trained model for all users, 1.5-2 hour checkpoint cycles). L4 (firm knowledge in weights, Engram + Harvey) and L5 (rank-1 LoRA merge, Nested Learning, TTT, self-distillation) exist only as research and do not move the level. L3 is inside the normal band, but every production row is a vendor describing its own system (tier 7), so the status is capped at emerging. Initial seed.

   obs:1b7d08dbobs:dc983a81

45 / 95 — limited or vague evidence

Confidence is independent of status: 90–95 multiple strong independent sources; 70–89 good evidence, some ambiguity; 50–69 mixed or hard to operationalise; below 50 limited or vague.

• Agent-workdays per human workday in frontier research (self-reported) ◐emerging
• Expert-data market run-rate (Mercor) ▲faster than normal
• Crosswalk: Return arrow ⇄ Training input (return arrow)