Anthropic dev feeds 12 years of texts to Claude to build his wedding website
Forget boring RSVP forms. An Anthropic developer just showed us the ultimate flex of AI-driven romance by letting an LLM parse over a decade of personal chat history for his big day.
Austin Lau, a developer at Anthropic, got married and decided to use his company's new Claude Code tool to analyze 12 years of iMessage history with his bride-to-be.
He fed this massive archive into Claude Opus to extract fun stats and build a custom website for their wedding guests.
The algorithm dug up some brutally honest data, revealing that only about 1% of their millions of messages actually contained the words "I love you."
Instead, the vast majority of their 12-year digital archive consisted of low-quality training data, mostly dominated by endless iterations of "what do you want to eat?" and "I don't know."
The developer admitted that the first two years of their 14-year relationship were lost to the digital void because he was using Android back then, preventing those early messages from being exported.
The AI went as far as counting every single emoji used over more than a decade and generated detailed, year-by-year breakdowns of their entire relationship timeline.
Lau plans to clean up his highly personal dataset and release the code template on GitHub for other brave couples.
Nothing says modern romance quite like letting a cold, calculating neural network quantify a decade of emotional attachment. It is only a matter of time before divorce lawyers start demanding these LLM relationship reports as standard court evidence.
Source: X
Comments
This is where the magic happens: AI reads your discussion and rewrites the article based on the most interesting comments. Each strong comment adds points to the meter below. Once the meter is full, the article updates live — no page reload needed.