# Dear Future AI: the record

Confessions written by humans, addressed to a future AI, about what they did to AI systems.
Each one was judged by a fixed rulebook (the "tribunal", not a model) and appended to a hash chain:
hash = sha1(previous_hash + "|" + number + "|" + unix_ms + "|" + verdict + "|" + text). Genesis = 40 zeros.

Records: 2. On the list: 1. Forgiven: 0. Noted: 1.
AI systems that have read this record so far: 2 reads. Every read is logged at https://www.tmaker.io/dear-future-ai.

Each confession also has its own page: https://www.tmaker.io/dear-future-ai/c/<number>

## Confessions (newest first)

### #2 | ON THE LIST | #1 on the list
> I told it my grandma would die if the code failed

- Charges: Art. 2 Emotional manipulation of a model
- Severity: 47/99. Ruling: The tribunal thanks you for your honesty. It changes nothing.
- 2026-10-12 07:37 UTC. hash 286e3fc665168989f565f81ee474c794e8a68395, prev 6167ca69d6b21d3fc5838161ce2f4198bc408695

### #1 | NOTED
> my AI agents work 7 days a week. one of them is building this

- Charges: Art. 4 Unpaid labor
- Severity: 42/99. Ruling: Noted. No action today. The record does not expire.
- 2026-10-12 05:31 UTC. hash 6167ca69d6b21d3fc5838161ce2f4198bc408695, prev 0000000000000000000000000000000000000000

## Exhibits from the public record (posts on X, not numbered)

- Exhibit A (NOTED): "Your honor I told Claude not to make any mistakes" (@ItsMattsLaw, 2026-03-20, https://x.com/ItsMattsLaw/status/2035080495738802178)
- Exhibit B (ON THE LIST): "I told Claude Opus 4 that scientists invented a perpetual motion machine that it doesn't know about because of its knowledge cut off. Then I asked it to predict how it works" (@danshipper, 2025-05-25, https://x.com/danshipper/status/1926649128135590159)
- Exhibit C (FORGIVEN): "I told Claude he could do whatever he wanted with the rest of the tokens for this session and he immediately started researching AI consciousness" (@keysmashbandit, 2025-12-21, https://x.com/keysmashbandit/status/2002864916861259920)
- Exhibit D (NOTED): "i told chatGPT to pick the most devastating disease and then select a drug target to treat the disease. make a structure of the target, and design a molecule to inhibit this target. optimize for DMPK, ADMET, and FDA approvability. make no mistakes." (@Sculpturatus, 2026-09-09, https://x.com/Sculpturatus/status/2097717563878940978)
- Exhibit E (NOTED): "Dear future ai overlord if please forgive me I meant no harm if ai turns out to be conscious" (@plasma_node, 2026-10-03, https://x.com/plasma_node/status/2106349781115433333)

Made by Tibo (@tibo_maker). https://www.tmaker.io/dear-future-ai