Our readingWhat is a korrent?
Drawn from what Boris Cherny said
Mental modelPlay long games
In the views onAnthropicunder our view: The tool leads in use, and its scores, its flaws and its future are argued from there
At Anthropic, we don't build for the model of today, we build for the model of six months from now. And that's still my advice to founders that are building on LLMs.
Where this was said
↗Y Combinator, quoting Boris Cherny on the Lightcone Podcastx.com
Added to korrents 31 Aug 2026 · How quotes work · Something wrong? Tell us
Do you hold this korrent?Do you also believe this?
I hold this tooCannot both be true
Sivers says to avoid predicting the future as much as possible and wait for it to become the present and reveal what is actually needed. Cherny says to build for the model six months from now rather than the one in front of you. Building for a thing that does not exist yet is the practice Sivers is telling people to stop.
In tension with
Cherny builds deliberately for a model that does not exist yet and tells founders to do the same; Hansson says guessing two model releases ahead is a waste of time and sends you mad. Not a contradiction, and the gap between them is the interesting part: Cherny is betting that the next model will be BETTER without claiming to know how, while Hansson is attacking the attempt to predict WHAT it will be able to do. One can act on the first without doing the second — but the bet only pays if the guess about capability is roughly right, which is the thing Hansson says nobody can do.
Shared subject first, then similar wording. They may agree or disagree.