gavin leech @gl@schelling.pt
context maximiser @Arb
Joined Dec 2022
"Morality is complicated, AIs try to be simple"
he can't keep getting away with it (writing the best critical work on AI forecasting in the world and putting it up on tumblr and only tumblr)
https://nostalgebraist.tumblr.com/post/705192637617127424/gpt-4-prediction-it-wont-be-very-useful
'Posts labeled "בבל" are experimental, unreliable, poetic, prophetic, metaphoric, contradictory, false, confused, incoherent, unclear, inchoate, incontinent, insane, repetitive, low-effort, pointless, cringe, fringe, binge, silly, facile, babbling, ...'
on the psychology of shame
https://forum.effectivealtruism.org/posts/zh6pZTZisqWq7AgnD/on-being-compromised
I can always tell when someone spends Too Much Time designing their blog. It is extremely endearing
"in comparison to less fluid research areas, adaptability is more important than experience. Before the Transformer, NLP relied on insights from linguistics... now... irrelevant. The Transformer was five years ago, so nobody has >5 years of experience."
blogs I found in 2023
Icon my graphic designer picked for Alignment. Nice but unintended ironies:
the two systems begin with the same orientation then there's a drastic orientation shift; they converge *with equal gravity*, & then the values handshake completes, merging them into one effective agent
The Tbilisi Wedding Palace / Palace of Rituals. A complete dream.
And meta-bullshit:
https://forum.effectivealtruism.org/posts/xomFCNXwNBeXtLq53/bad-omens-in-current-community-building
https://forum.effectivealtruism.org/posts/GqGvCSjyTAxRcpa5G/research-deprioritizing-external-communication
https://forum.effectivealtruism.org/posts/av7MiEhi983SjoXTe/some-core-assumptions-of-effective-altruism-according-to-me
https://forum.effectivealtruism.org/posts/AjxqsDmhGiW9g8ju6/effective-altruism-in-the-garden-of-ends
https://forum.effectivealtruism.org/posts/ZkhABk4rRMqsNmwvf/changing-the-world-through-slack-and-hobbies
EA on nuclear war and expertise
https://forum.effectivealtruism.org/posts/bCB88GKeXTaxozr6y/ea-on-nuclear-war-and-expertise
A visit to the idea-machine fair
https://forum.effectivealtruism.org/posts/Kv2sTdBF2FczLEktn/a-visit-to-the-idea-machine-fair
Five [planet-sized] steps for quantifying speculative interventions
Useful Vices for Wicked Problems
https://forum.effectivealtruism.org/posts/yPsG9fr3coczCm7Ku/useful-vices-for-wicked-problems
Does putting kids in school now put money in their pockets later?
Critical review of givewell's 2022 cost effectiveness model
context maximiser @Arb
Joined Dec 2022
