I'm excited about @collect_intel and @audreyt testing a step toward more productive AI policy discourse: https://t.co/ivLHwWXsmR
For intl diplo, large-scale deliberation yielding common ground is powerful in building cross-regional coalitions with critical mass.
Join the Collective Intelligence Project and @audreyt in piloting alignment assemblies: a series of collective input processes that can start building towards collectively-governed AI.
>> https://t.co/RDRx5I7J71 <<
Just had the scary experience of hearing my own voice on one of my favourite podcasts...All the better to do it with an old friend: Thank you for joining me Igor. AI risk is boring and political! Have a listen����Many thanks also @MLStreetTalk for giving airtime to ML governance
We spoke with Carla Cremer @CarlaZoeC and
Igor Krawczuk @thegermanpole about existential risk, alignment, effective altruism and governance. Filmed in-person at Oxford University! https://t.co/jTZrPVLbjV
long podcast on the same subject #FTXbankruptcy#EffectiveAltruism#DemocratisingRisk also just came out - second half allowed us to expand on the more controversial ideas and some of the criticisms we received https://t.co/ZdoOrSDkTX
If EAs have a next decade, it should be that of the institutional turn. They should offer itself as the testing ground for real innovation in institutional decision-making.
I mustn't rest until I hear @RoryStewartUK and @landemore discuss democratic experimentation on @RestIsPolitics, 🙃so: my arguments for risk reduction via 'epistemic mechanism design' applied to Effective Altruism and their exposure to FTX / @SBF_FTX https://t.co/I2pPwGV6Ac 🧵
some more reform ideas: @GiveDirectly could advance towards deliberation platforms for recipients of their US disaster relief program, @LongResilience
could run scaled citizen assemblies on risks from automation, @80000Hours could manage an open source jobs wiki, ...
This is a really, really stunning piece by @CarlaZoeC about the ideology that drove not only Sam Bankman-Fried, but also many of the most powerful people in AI right now https://t.co/pnACIOnC0Q