ysskrishna/ai-agent-skills / skills/fermi-estimation
fermi-estimation
VerifiedUse for Fermi estimation, or when a rough number is needed fast: "roughly how many...", "ballpark the cost", "estimate capacity or market size", "is this even feasible?". Decomposes, uses ranges and sanity-checks. Skip when exact data is available or a precise calculation is wanted.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 3.5 KB |
| references/example.md | markdown | 1.8 KB |
references/example.md
View on GitHub# Example: Fermi Estimation (light path)
Illustrative case. All inputs are invented for the example.
**User:** "We're adding a 'recent activity' feed that writes one row per user action. Roughly how much storage per year, and will one database handle it?"
**Quantity:** storage added per year by the feed, in GB, plus peak writes per second. **Purpose:** decide whether the current database is enough. **Needed accuracy:** within a factor of 3.
**Pass:** Decompose, Estimate, Combine, Cross-check, Sensitivity.
**Decompose**
- Storage per year = active users x actions per user per day x 365 x bytes per row.
| Factor | Low | High | Basis |
|--------|-----|------|-------|
| Active users | 20,000 | 60,000 | `[ESTIMATED]` user's current count and growth |
| Actions per user per day | 10 | 40 | `[ESTIMATED]` guess; check product analytics |
| Bytes per row (with index) | 200 | 500 | `[ESTIMATED]` typical small row |
**Combine**
- Low: 20,000 x 10 x 365 x 200 B = 14.6 GB per year.
- High: 60,000 x 40 x 365 x 500 B = 438 GB per year.
- Geometric midpoint: about 80 GB per year. Order of magnitude: tens to hundreds of GB.
- Peak writes: average is users x actions / 86,400 s, about 2-28 per second; assume peak is 10x average, so roughly 20-280 per second.
**Cross-check**
- A similar existing table of 1 million rows takes about 300 MB in the user's database (`[ESTIMATED]`, check). Rows per year, low to high: 73 million to 876 million, which is about 22 GB to 260 GB. Close to the first result.
**Sensitivity**
- Actions per user per day moves the answer most (4x range). Measure it from product analytics for two weeks.
- Verdict: storage is not a problem for one database for several years at the low end; at the high end plan retention or partitioning. Peak writes up to a few hundred per second are usually fine, but verify against the database's measured limit.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |