A Donor

[QACI](https://www.lesswrong.com/posts/4RrLiboiGGKfsanMF/the-qaci-alignment-plan-table-of-contents) is a plan for aligning AI in a way which is robust to superintellgent optimization, under development by [Orthogonal](https://orxl.org/), an agent foundations research organization which is searching for formal goals which cause good things when maximized.

Our current best aligned formal goal is [QACI](https://www.alignmentforum.org/posts/CYtzXadXFtBSBYm3J/a-narrative-explanation-of-the-qaci-alignment-plan "‌") (see also [a story of how it could work](https://www.alignmentforum.org/posts/CYtzXadXFtBSBYm3J/a-narrative-explanation-of-the-qaci-alignment-plan "‌") and [a tentative sketch at formalizing it](https://www.alignmentforum.org/posts/qeRqmdadsdj8Frvyn/a-rough-sketch-of-formal-aligned-ai-using-qaci "‌")), which implements something [like coherent extrapolated volition](https://carado.moe/cev-coherent-enough.html "‌") by extending a ["past user"](https://carado.moe/outer-alignment-past-user.html "‌")'s reflection to be simulated/considered arbitrarily many times, until alignment is solved.




Project	Earliest donation	Latest donation	Total amount
AI Safety Upskilling of Matt Watkins	almost 2 years ago	about 1 year ago	$11,303.12
Alignment Continuity	over 1 year ago	over 1 year ago	$26,295.03

A Donor Loading...

A Donor
Loading...