[cadamsdotcom]

Most positive AI thing shared on HN in a while, and here’s why.

Even after handing the agent a validator and iterating the validation, even after the validator get really good at telling the model what to do and what not to do, there’s still an emptiness.

This gives the author strong ground to rightly assert that both now and after new models render 2026 model writing’s emptiness old news, there’ll remain reasons to write, to communicate, to share understanding.

It’s nice to read something not proclaiming doom and not lamenting career death.

No matter how smart our tools become, the human condition still has plenty ahead.