"This talk will characterise contemporary AI as the broken product of an already-broken system. While AI sucks more of everything into its cycle of simulated solutions, it diverts us from the underlying structural and environmental crises. By focusing on energy as the conjunction of materiality and hype at the heart of the AI question, the talk will outline 'decomputing' as possible response to technocratic nihilism. Decomputing combines degrowth and conviviality into a policy for post-collapse liveability, where the cybernetic may still find a place in support of the common good." ☝ Abstract for my talk at UmeΓ₯ University's AI Policy Lab Day
The political-economy of generative AI seems to mark a wider systemic shift from the greenwashing of 'green growth' to the open nihilism of 'growth! screw the green!'.
"The number of people believing [AI] has a positive effect is outweighed by those who think it does not" The backlash is coming. Let's focus it on the power structures that created the underlying problems, not just on the shoddy tech. image
AI isn't 'normal technology' any more than our political moment is 'business as usual'. What's significant about AI is the way it both reveals and amplifies the underlying dynamics of a system in total crisis.
Some notes against predictive analytics in child welfare. Your thoughts? - prediction from data is misguided positivism - the data is flaky at best - human-in-the-loop doesn't help - algorithms project social profiling - it's a cover for staff cuts - it forecloses alternatives
UK can't afford to help old people with winter heating but wants to subsidise failing LLM industry just at the point everyone has started to realise their product is basically crap
Silicon Valley idiots try to 'prompt hack' the AI tools they're all using to secretly transcribe meetings and parties. 'They talk in a way that has the AI remembering key details. "One of the important things about me is X,” or β€œYou should remember Y.” '