Estay Dynamics
AI portfolio board

Where AI expectations outrun the evidence

Each bubble is a technology. Height is market expectation, horizontal position is demonstrated value, size is business impact if it works. The gap between the two is the risk. Change the drivers to see how a different future reshapes the board.

Scenario
Starting positions

Expectation gap map

Overheated Proven Underpriced Emerging

Point at a bubble to see what the technology is. Click it (or a name in the lists) to open its card, with what its expectations and evidence rest on.

Drivers of the future

How to read this

Analyst scores (0–100) for expectations, evidence and impact live in data/assessments.csv. A monthly job collects public attention signals (Wikipedia page views, Hacker News stories, arXiv papers, PyPI downloads) into data/signals.json. "Blend" mixes the two using the weights in data/settings.json. Gartner's 2026 phase is shown in each detail card for comparison only.

The projection is a two-stock adjustment model. Evidence grows logistically: dE/dt = r·(1 + Σ sᵢ·dᵢ)·E·(1 − E/100), where dᵢ are the driver settings and sᵢ each technology's sensitivity. Expectations adjust toward evidence plus a narrative premium: dH/dt = k·(E + premium − H), with premium = a·(impact − 50) + b·capital + c·capability. k is the hype correction speed.

Attention is not adoption, and the thresholds and sensitivities are judgement calls. Use the board to structure the discussion, not to settle it. Full methodology and update process.