Shanu Mathew(@ShanuMathew93):This got me thinking as lines will eventually blur btwn what gets pushed into deterministic code as rules vs. probabilistic agents on well defined/repeatable tasks extremely fine tuned for situations with varying degrees of uncertainty, etc. the decision tree will look interesting as organiations mature across this boundary -->Explicit rules can move into code vs. -->Recurring judgments on messy inputs may stay with small, specialized models vs. --> Broader agents make sense when the steps themselves vary.

2026.09.22 21:10

This got me thinking as lines will eventually blur btwn what gets pushed into deterministic code as rules vs. probabilistic agents on well defined/repeatable tasks extremely fine tuned for situations with varying degrees of uncertainty, etc. the decision tree will look interesting as organiations mature across this boundary -->Explicit rules can move into code vs. -->Recurring judgments on messy inputs may stay with small, specialized models vs. --> Broader agents make sense when the steps themselves vary.

Frank Downing@downingARK

2026.09.22 20:17

totally agree for "general purpose" models. The valid exception would be the cost comparison of a general purpose frontier model to a smaller model fine tuned for a narrow use case. The latter can still yield material cost savings, for example in the Crowdstrike data below, but only if you have a stable and sufficiently scaled use case to justify the R&D investment (not one time but ongoing to keep up w/ the frontier).