Hopeful technologist.
Community tech, careful innovation, socially progressive tech policy.
https://www.careful.industries
https://hopeful.technology
https://buttondown.email/justenoughinternet
DMs don't work but [email protected] will find me
Rachel Coldicutt
Loading...
Oh, one other thing. Interesting that Reddit (darling of the LLMs) is not specifically mentioned in the announcement. I think there is an *interesting* crossover here with impacts on training data and I wonder who is having that conversation?
I mean, I am not at all against regulatory measures that undermine Alphabet's cultural hegemony but I do think it would be useful to understand some of the short-term consequences before agreeing to roll it out in 9 months' time.
Politics UK mentioning Reddit but not sure of the source on that? bsky.app/profile/poli...
My unscientific low-key group chat experience indicates that broad support at headline level turns into head scratching when emergent details discussed. Suggest this might go a similar
way to digital ID: detail on implementation will lessen support and pushback will come from unlikely sources
I did some work on a govt project sizing the "safety tech" industry a few years ago, and there was significant ambition that the UK would be a market leader in this emerging area. The reality is that brings its own regulatory problems, by the boat load rachelcoldicutt.medium.com/three-recomm...
Need to switch my brain on and off again to stop thinking about this but now would be a good time for someone to do a rapid economic analysis on the implications of everyone having to verify their age to watch a YouTube video or embedded clip.
I wrote the above in 2022, before genAI had also totally broken assurance and introduced a massive new error surface. Layering technologies on top of another like this is likely to make some companies a lot of money, unlikely to make many people safer.
