RSS

Tamsin Leake

Karma: 2,584

I’m Tamsin Leake, co-founder and head of research at Orthogonal, doing agent foundations.

How “Pause AI” ad­vo­cacy could be net harmful

Tamsin Leake26 Dec 2023 16:19 UTC
32 points
8 comments2 min readLW link
(carado.moe)

OpenAI, Deep­Mind, An­thropic, etc. should shut down.

Tamsin Leake17 Dec 2023 20:01 UTC
36 points
48 comments3 min readLW link
(carado.moe)

Align­ment work in anoma­lous worlds

Tamsin Leake16 Dec 2023 19:34 UTC
24 points
4 comments3 min readLW link
(carado.moe)

Some bi­ases and se­lec­tion effects in AI risk discourse

Tamsin Leake12 Dec 2023 17:55 UTC
26 points
21 comments4 min readLW link
(carado.moe)

How LDT helps re­duce the AI arms race

Tamsin Leake10 Dec 2023 16:21 UTC
70 points
13 comments4 min readLW link
(carado.moe)

We’re all in this together

Tamsin Leake5 Dec 2023 13:57 UTC
68 points
65 comments2 min readLW link
(carado.moe)

So you want to save the world? An ac­count in paladinhood

Tamsin Leake22 Nov 2023 17:40 UTC
65 points
19 comments15 min readLW link
(carado.moe)

an Evan­ge­lion di­alogue ex­plain­ing the QACI al­ign­ment plan

Tamsin Leake10 Jun 2023 3:28 UTC
45 points
15 comments43 min readLW link
(carado.moe)

for­mal­iz­ing the QACI al­ign­ment for­mal-goal

Tamsin Leake and JuliaHP
10 Jun 2023 3:28 UTC
53 points
6 comments14 min readLW link
(carado.moe)

Orthog­o­nal’s For­mal-Goal Align­ment the­ory of change

Tamsin Leake5 May 2023 22:36 UTC
68 points
12 comments4 min readLW link
(carado.moe)

Orthog­o­nal: A new agent foun­da­tions al­ign­ment organization

Tamsin Leake19 Apr 2023 20:17 UTC
207 points
4 comments1 min readLW link
(orxl.org)

con­tinue work­ing on hard al­ign­ment! don’t give up!

Tamsin Leake24 Mar 2023 0:14 UTC
82 points
45 comments1 min readLW link
(carado.moe)

the QACI al­ign­ment plan: table of contents

Tamsin Leake21 Mar 2023 20:22 UTC
102 points
0 comments1 min readLW link
(carado.moe)

your ter­mi­nal val­ues are com­plex and not objective

Tamsin Leake13 Mar 2023 13:34 UTC
60 points
6 comments2 min readLW link
(carado.moe)

the quan­tum am­pli­tude ar­gu­ment against ethics deduplication

Tamsin Leake12 Mar 2023 13:02 UTC
11 points
16 comments2 min readLW link
(carado.moe)

QACI: the prob­lem of blob lo­ca­tion, causal­ity, and counterfactuals

Tamsin Leake5 Mar 2023 14:06 UTC
23 points
1 comment2 min readLW link
(carado.moe)

state of my al­ign­ment re­search, and what needs work

Tamsin Leake3 Mar 2023 10:28 UTC
51 points
0 comments2 min readLW link
(carado.moe)

Hello, Elua.

Tamsin Leake23 Feb 2023 5:19 UTC
37 points
18 comments4 min readLW link
(carado.moe)

a nar­ra­tive ex­pla­na­tion of the QACI al­ign­ment plan

Tamsin Leake15 Feb 2023 3:28 UTC
53 points
29 comments6 min readLW link
(carado.moe)

so you think you’re not qual­ified to do tech­ni­cal al­ign­ment re­search?

Tamsin Leake7 Feb 2023 1:54 UTC
55 points
7 comments1 min readLW link
(carado.moe)