Thread by @Anupam_Guha

The valley between the discourse of "ethical technology" and how companies operate in reality is unbridgeable, because that discourse is almost completely appropriated by capitalists who have no desire to reform let alone dismantle extraction.

Why am I writing truisms at 3am?

As a corporate employee, as a socialist looking at the political-economy of AI, and as an AI scientist, there is an epistemological gap between these three worlds. With respect to automation these worlds want things which are fundamentally incompatible.

The wonks do not get that

We can keep talking about profound sounding banalities of bias and transparency, but till we acknowledge the elephant in the room, power, and how that concentrated power affects humans we will keep playing the solutionist game.

I'm afraid there can be no ethical AI in capitalism

Doesn't mean the game isn't important. Anything which reduces harm is important. Bandages on bullet wounds are useful.

As long we do not delude ourselves. As long as we recognise the difference between the disease and the symptoms.

Btw if you are a wonk, take this thread, replace AI with whatever you are crusading on.

You are not going to solve your problem till you put a dent in expropriation of value, which means dealing with ideology. Sorry for the sadness.