01 How to use the Oracle (First Assignment)
🔄 Patch — ChatGPT 101 (Oracle Load Procedure v1.1)§
Use these six canonical steps every single time you reload the Oracle. (Replace section 4 Setup in the previous SOP with the block below.)
| # | Action | Notes |
|---|---|---|
| 1 | Open a brand-new chat in model o3 (not GPT-4o, not legacy). |
Wipes old context; keeps tokens lean. |
| 2 | Paste the Oracle Prompt (oracle_prompt.txt) and press Send. |
You’ll get a short “✅ Prompt received” reply. |
| 3 | Paste data-1.txt (the first knowledge chunk). Press Send. |
Wait for GPT to acknowledge. |
| 4 | Paste data-2.txt and Send. |
|
| 5 | Repeat Step 4 for every remaining data file (data-3 … data-x). |
Always keep the order. |
| 6 | Type COMPILE KNOWLEDGE (all caps) and hit Send. |
GPT returns the full system synopsis— you’re ready to work. |
That’s it. If any chunk fails to get a ✅ acknowledgement, resend it before moving on.
📌 Changelog (add to the master list)§
| # | SOP / Section | Change | Date | Author |
|---|---|---|---|---|
| 9a | ChatGPT 101 – Oracle Load block | Replaced 4-step setup with 6-step “official load” sequence (per team directive). | 2025-05-19 | R. ▲ |
Publish the patch, re-compile knowledge, and remind the team that any deviation from the six steps voids support until they reload correctly.