And so far the demonstrated benefits are... what, exactly? Despite being absolutely flooded by attempted uses of these tools, I can't think of a single time that an LLM has made my life unequivocally easier or better. When weighed against the above list, I'm happy being unilaterally against it all.

Show Post

- AI use in warfare (all signs point to the girls' school in Iran being partly the fault of Claude, arguably the least evil model) - if these tools end up being as capable as they're promising, the economy as we know it will not exist. what's a job?

Show Post

- diving headfirst into a world where it gets harder and harder for even educated and critical people to be sure of what's "real" - no more entry level jobs, which has knock on effects for the entire workforce - cognitive atrophy (this has been demonstrated experimentally)

Show Post

- unpredictable negative externalities (things that have already happened: jailbreaking, malware, sexual bullying) - environmental damage (recarbonizing the grid, water use, rising energy costs) - this is prolly a bubble that we don't know how it's going to pop

Show Post

Laundry list of hazards/dangers of LLM tools, that will generally still be problems regardless of who owns the models: - internet of slop - unpredictable/unexplainable/uncontrollable algorithms - IP theft - AI induced psychosis

Show Post