# Greg Osuri

I build things for people that build things

1,689 predictions from 240 talks and podcasts; 41 essays; 1,754 markdown twins.

## Retrieval

1. Start with the recent index or the narrowest recommended theme index below.
2. Fetch the selected prediction twin for its quote, date, source and timestamp.
Avoid broad indexes as an initial fetch; they remain available for cross-theme searches.
Index twins contain titles and IDs only; leaf twins contain full content.

## URL rules

Canonical prediction: https://www.gregosuri.com/predictions/{theme}/{id}/
`id` is the date-first `slug_id` printed in the title index; copy it exactly.
Append `index.md` to a page URL for its markdown twin. Append `card.png` to a prediction, essay or theme URL for its social card.

## Inventory and estimated fetch budgets

Estimates are UTF-8 bytes / 4, rounded up; this heuristic is not a tokenizer count.
Budgets below describe individual fetched files. This inventory does not include its own size.

- https://www.gregosuri.com/index.md — ~380 tokens
- https://www.gregosuri.com/about/index.md — ~572 tokens
- https://www.gregosuri.com/citing/index.md — ~961 tokens
- https://www.gregosuri.com/page2/index.md — ~381 tokens
- https://www.gregosuri.com/page3/index.md — ~381 tokens
- https://www.gregosuri.com/page4/index.md — ~381 tokens
- https://www.gregosuri.com/page5/index.md — ~381 tokens
- https://www.gregosuri.com/page6/index.md — ~381 tokens
- https://www.gregosuri.com/page7/index.md — ~381 tokens
- https://www.gregosuri.com/page8/index.md — ~381 tokens
- https://www.gregosuri.com/page9/index.md — ~381 tokens
- https://www.gregosuri.com/posts/index.md — ~2,250 tokens

## Recommended title indexes (up to ~5,000 tokens each)

- https://www.gregosuri.com/predictions/recent/index.md — ~2,509 tokens
- https://www.gregosuri.com/predictions/ai-agents/index.md — ~1,142 tokens
- https://www.gregosuri.com/predictions/energy-ai/index.md — ~4,193 tokens
- https://www.gregosuri.com/predictions/local-ai/index.md — ~1,585 tokens
- https://www.gregosuri.com/predictions/open-source-ai/index.md — ~1,711 tokens

## Broad indexes (high cost)

Use these only when a cross-theme or full-theme search requires them.
- https://www.gregosuri.com/predictions/index.md — ~52,463 tokens
- https://www.gregosuri.com/predictions/cloud-decentralization/index.md — ~14,165 tokens
- https://www.gregosuri.com/predictions/crypto-depin/index.md — ~7,054 tokens
- https://www.gregosuri.com/predictions/decentralized-ai/index.md — ~5,316 tokens
- https://www.gregosuri.com/predictions/gpu-economics/index.md — ~7,526 tokens
- https://www.gregosuri.com/predictions/local-compute/index.md — ~5,277 tokens
- https://www.gregosuri.com/predictions/other/index.md — ~5,222 tokens

- Prediction leaf twins: 1,689 files; ~233–550 tokens each.
- Essay twins: 41 files; ~131–7,907 tokens each.

Citation contract and navigation guide: https://www.gregosuri.com/citing/

Transcripts and the video catalog are not published.