What if a model learns to fine-tune its own successor?
A lab demonstrating a model that fine-tunes its own successor triggers a capability-overhang repricing and a safety-pause debate — compute names (NVDA/AVGO/memory) catch a bid on accelerated capex while displacement fears cap consumer/credit. Rhymes with the May-2023 Nvidia blowout that ignited the AI-capex wave; the recursive-improvement angle is the novel risk premium layered on top. Forward angle: a credible safety-pause push is the two-way risk — it could freeze the very capex the chip bid is pricing, so the long-semis read is conditional on regulators not intervening.
Every number ships with its receipt — the odds, the range, the precedents, and a public grade at Reality Check. The statistical machinery that produces it is proprietary.
The butterfly cascade
How this trigger trickles across markets, left → right — the root shock, its first‑order moves, then the ripple effects. Drag any node; tap a market for its real price history.
Resolution timeline — how this probability is moving
Our model's odds (electric blue) over time vs the market's (Polymarket, amber), from the past toward the 1–3 years horizon. Each dot is a real macro event that nudged the probability — green pushed it up, red pushed it down. Tap a dot for the source. Loading the probability audit trail…
What it would mean
If this plays out, it is a risk-off shock. A lab demonstrates a model that fine-tunes its own successor, triggering safety-pause debate and a capability-overhang repricing. The trigger decomposes into signed root‑shocks — AI capex ▲ · Job displacement ▲ — which propagate through our causal graph to the markets below.