
Retrieval keeps showing up where the output piles up
The week's most-discussed tools all discovered the same need, search over what their systems already produced
The week in AI, read with a data and retrieval lens. One email a week.


The week's most-discussed tools all discovered the same need, search over what their systems already produced

The prices paid this week were for distribution and capacity while the models themselves kept getting cheaper and more interchangeable, which leaves the specifics of a given system as the part no acquisition fixes

Capability was given away this week while positions in the path were bought, which prices the place a request passes through above the thing that answers it

Checking is being packaged and sold in the same week that the artifacts meant to prove correctness, formal proofs and model reasoning traces alike, turn out to persuade nobody on their own

The week's cost arguments were all about distance, how far an answer has to travel between the data and the model, which is where the spending shows up before the product does

The expensive parts of the serving stack were published this week, which leaves access to the customer's data and the capacity to check model output as the two things still scarce