Pigeon LLM

• 

  Flowchart of our over-engineered solution

• 

Inspiration

Remember when pigeons were the original DMs? They took their sweet time, got distracted mid-flight, and still somehow did their job. We thought: what if we brought that same inefficiency to the 21st century?

What it does

Pigeon LLM is a messaging system that refuses to be instant. It uses GPT-4 to analyze conversations and intentionally delays your messages based on how urgent (or not) they are. If it’s important? We’ll sit on it. If it’s a meme? Probably going out immediately. Bonus: every message comes with a wildly unhelpful status update

How we built it

We used Twilio to create a modern messaging system. Then we looped in GPT-4 to judge how serious the convo is (spoiler: GPT thinks everything is serious). If the model decides the message is “too important,” it triggers a delay, adds it to a MongoDB queue, and schedules it like it’s mailing a handwritten letter in 1860.

Challenges we ran into

Turning “delayed texting” into a legitimate project. Convincing ourselves this wasn’t just procrastination disguised as innovation.

Accomplishments that we're proud of

We recreated the pain of ancient communication systems - with AI! Also proud of how often we said “carrier pigeon” with a straight face in tech meetings.

What we learned

That life today is absurdly convenient and maybe we don't deserve it. Also, that GPT-4 has no chill when it comes to judging what’s important.

What's next for Pigeon LLM

Turning this into a chatbot so everyone can experience the joy of having their messages delayed with dignity. Coming soon: AI pigeons with commitment issues.

Built With

• grok
• openai
• python
• rabbitmq
• twilio