Re:How I (almost) replaced two machine learning models with ...
by countWSS on Sunday January 28, 2024 @03:30PM
(#39169515)
Sunk cost fallacy: they invested time, effort and their reputation
on the "smart solution", its going against the consensus.
There is no point optimizing overengineered stuff,
it will be simplified to that closed-form equation on its own,
and the company will gain some technical know-how managing
that system to implement advanced ML analytics.
Re:How I (almost) replaced two machine learning models with ...
by frangolino on Sunday January 28, 2024 @03:18PM
(#39169413)
You'd have made the PM look like an utter fool, so he shut you down. Classic office politics. What's good for (i) the company, (ii) you, (iii) your boss... often are not the same things... sometimes in direct contradiction.
The interests of the involved parties must always be taken into account.
Re:How I (almost) replaced two machine learning models with ...
by icosian on Monday January 29, 2024 @02:53AM
(#39173831)
Perhaps it could have been offered as a "naive" interim solution to be replaced by machine learning, with the benefit that the ML team would have a benchmark to show how good their model was by comparison.
The interests of the involved parties must always be taken into account.