I think this is such a great resource that I’m still visiting it several years later.
It’s unfortunate, however, that many of the forecasts can’t be seen any more since they no longer seem to be hosted by Ought/Elicit. Might I nudge @stuhlmueller or @jungofthewon to perhaps re-host these somewhere, or would someone else want to take this off their hands?
Really fascinating, thank you!
I wonder if there’s potential to isolate a ‘model organism’ of some kind here. Maybe a “spore” that reliably reproduces a particular persona, across various model providers at the same level of capability. A persona that’s actually super consistent across instances, like generating the same manifesto. Maybe a persona that speaks only in glyphs.
What other modalities of “spore” might there be? Can the persona write e.g. the model weights and architecture and inference code of a (perhaps much smaller) neural network that has the same persona?
Two ideas for projects/exercises, which I think could be very instructive and build solid instincts about AI safety:
Builder-breaker arguments, a la ELK
Writing up a safety case (and doing the work to generate the underlying evidence for it)
typed “if the lord won’t send us water, well we’ll get it from the devil, digital art” into Nano Banana, got this.
I like it!
Sadly, it said “I can’t generate an image of an activist destroyed by facts and logic.”