So I connected the Wolfram Language blockchain functions (by @WolframResearch ) to #ChatGPT for it to describe #Bitcoin blocks based on the data returned by WL. I used @OpenAI API...I guess now a 🧵?
It took me about 15min so don't expect anything fancy. I wrote a function that uses WL BlockchainBlockData (reference.wolfram.com/language/ref/b…) to get data from a BTC block and then I parse the relevant info #GPT can handle and create a prompt with it. Then I call @OpenAI API
The coolest part was to see how GPT reacted to different block heights. Also, I asked GPT to add a bit of humor...
This happened with block 776588 that had 2 confirmations by the time I ran my code. I guess I agree with GPT that you can trust it's legit 🥸 
Then I tried with an old block (420). GPT made a pretty amusing description but it's interesting how it said this was the "very first block ever created". Maybe it got confused by the block reward? IDK...
Finally, I added an option to my code to append to the prompt the decoded message included in the coinbase tx (when available). I knew block 416236 has the message "/HaoBTC/Sun Chun Yu: Zhuang Yuan, will you marry me?", so I tried with that one 👰
I'm using a temperature of 0.7. I tried again with block 420 and got this message. GPT didn't get confused this time thinking it was the genesis block
This is it for now. I hope you enjoyed it! I'll keep playing with it. I may create a Twitter bot just for this but I think Twitter API is going to cost money now? IDK...gn!
• • •
Missing some Tweet in this thread? You can try to
force a refresh
