amadan
paulca / paulca.com Public
Ledger: sharpen the method section's opening line
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
1 file changed,
+13
−4
ai.html+13 −4
diff --git a/ai.html b/ai.html| index a629bf8..af56bdc 100644 |
| --- a/ai.html |
| +++ b/ai.html |
| @@ -65,15 +65,24 @@ estimate of the energy and carbon the AI work cost.</p> |
| <td>~15 Wh</td> |
| <td>~6 g</td> |
| </tr> |
| +<tr> |
| + <td>July 21, 2026</td> |
| + <td>Quoting this page’s method section — “> Anthropic doesn't |
| + publish per-request energy figures” — “Anthropic, in their |
| + cowardice, doesn't”.</td> |
| + <td>Sharpened the editorial line below. About 5 model invocations.</td> |
| + <td>~5 Wh</td> |
| + <td>~2 g</td> |
| +</tr> |
| </table> |
| -<p><b>Running total:</b> ~69 Wh, ~28 g CO₂e — roughly the energy it takes |
| -to boil two-thirds of a litre of water, or the emissions of driving a petrol car |
| -about 170 metres.</p> |
| +<p><b>Running total:</b> ~74 Wh, ~30 g CO₂e — roughly the energy it takes |
| +to boil three-quarters of a litre of water, or the emissions of driving a petrol car |
| +about 180 metres.</p> |
| <h2>Method</h2> |
| -<p>Anthropic doesn't publish per-request energy figures, so these are |
| +<p>Anthropic, in their cowardice, doesn't publish per-request energy figures, so these are |
| order-of-magnitude estimates, not measurements. The arithmetic: |
| <i>(model invocations in the session) × 1 Wh per invocation × |
| 400 g CO₂e per kWh</i> (a rough global-average grid intensity). Published |