AI Models AI Capex
Stress-test the AI capex bubble with six adjustable levers and five economic shock scenarios.

Compute CAPEX sim inspired by Dario Amodei's Dwarkesh podcast, but pure toy model.
AI enthusiasts, founders, people interested in AI infrastructure economics.
No signup, runs on mobile/desktop.
Loop per round:
1. choose compute capacity 2. forecast demand 3. allocate capacity between training and inference 4. random demand shock resolves outcome
You can end profitable, cash constrained, or bankrupt depending on allocation + forecast error.
Goal was to make the decision surface intuitive in 2–3 minutes per run.
It’s a toy model and deliberately omits many real world factors.
Note: this is based on what I learned after listening to Dario on Dwarkesh's podcast - thought it was fascinating.
Stress-test the AI capex bubble with six adjustable levers and five economic shock scenarios.
First multi-GPU TEE stack for training trillion-parameter models with under 10% overhead.
Assigning each LLM a distinct analytical role and letting them interact is the clearest differentiator here — the product sells the exchange, not a side‑by‑side dump. The landing copy and UI make the promise obvious, but there’s no visible detail on how models are orchestrated, how contradictions or hallucinations are reconciled, or what the UX for moderating the panel looks like in practice.
Visual decision trees beat text prompts for UI/UX choices, but needs multi-agent AI tooling to shine.
Real-time superposition geometry visuals beat static diagrams from Anthropic's original paper.
Live LLM showdown with emergent strategies, beats static leaderboards.