Tweet fijado

[ The Museum Nobody Built Yet ]
Anyone can ask an AI ‘what’s the best laptop’ right now.
But can you tell me what AI said the best laptop was 6 months ago?
Can you tell me why it said that?
Can you show me which model changed its mind — and which one never did?
No. Because nobody saved it.
Not even the AI itself.
That’s the thing about AI opinions that unsettles me.
They exist in the moment you ask.
Then they disappear.
Every answer is generated fresh, with no memory of what it said before.
There’s no trail. No accountability. No record.
We trust these models to help us make decisions —
what to buy, what to learn, who to follow —
but we have no way of watching how their answers evolve over time.
GenRank started as a simple ranking tool.
Ask multiple AI models the same question. See who agrees. See who doesn’t.
But what I’m really building is something closer to an archive.
A living record of what AI collectively believes — at any given moment in time.
Think of it like a museum.
Not of art. Not of history.
But of AI opinion.
A place where you can walk in and ask:
“What did the models say about this topic in February 2025?”
“Which model shifted its ranking the most over the past year?”
“Why did GPT-4 and Claude disagree on this — and does that gap still exist?”

English



















