Lukas Finnveden

Karma: 2,981

Previously “Lanrian” on here. Research analyst at Redwood Research. Views are my own.

Feel free to DM me, email me at [my last name].[my first name]@gmail.com or send something anonymously to https://www.admonymous.co/lukas-finnveden

Notes on cooperating with unaligned AIs

Lukas Finnveden24 Aug 2025 4:19 UTC

45 points

8 comments21 min readLW link

(blog.redwoodresearch.org)

Being honest with AIs

Lukas Finnveden21 Aug 2025 3:57 UTC

63 points

6 comments17 min readLW link

(blog.redwoodresearch.org)

AI-enabled coups: a small group could use AI to seize power

Tom Davidson, Lukas Finnveden and rosehadshar

16 Apr 2025 16:51 UTC

132 points

23 comments7 min readLW link

What’s important in “AI for epistemics”?

Lukas Finnveden24 Aug 2024 1:27 UTC

50 points

2 comments28 min readLW link

(www.forethought.org)

Project ideas: Backup plans & Cooperative AI

Lukas Finnveden8 Jan 2024 17:19 UTC

18 points

0 comments13 min readLW link

(www.forethought.org)

Project ideas: Sentience and rights of digital minds

Lukas Finnveden7 Jan 2024 17:34 UTC

20 points

0 comments20 min readLW link

(www.forethought.org)

Project ideas: Epistemics

Lukas Finnveden5 Jan 2024 23:41 UTC

43 points

4 comments17 min readLW link

(www.forethought.org)

Project ideas: Governance during explosive technological growth

Lukas Finnveden4 Jan 2024 23:51 UTC

20 points

0 comments16 min readLW link

(www.forethought.org)

Non-alignment project ideas for making transformative AI go well

Lukas Finnveden4 Jan 2024 7:23 UTC

44 points

1 comment3 min readLW link

(www.forethought.org)

Memo on some neglected topics

Lukas Finnveden11 Nov 2023 2:01 UTC

28 points

2 comments7 min readLW link

(open.substack.com)

Implications of evidential cooperation in large worlds

Lukas Finnveden23 Aug 2023 0:43 UTC

39 points

4 comments17 min readLW link

(lukasfinnveden.substack.com)

PaLM-2 & GPT-4 in “Extrapolating GPT-N performance”

Lukas Finnveden30 May 2023 18:33 UTC

57 points

6 comments6 min readLW link

Some thoughts on automating alignment research

Lukas Finnveden26 May 2023 1:50 UTC

30 points

4 comments6 min readLW link

Before smart AI, there will be many mediocre or specialized AIs

Lukas Finnveden26 May 2023 1:38 UTC

59 points

14 comments9 min readLW link 1 review

PaLM in “Extrapolating GPT-N performance”

Lukas Finnveden6 Apr 2022 13:05 UTC

85 points

19 comments2 min readLW link

Truthful AI: Developing and governing AI that does not lie

Owain_Evans, owencb and Lukas Finnveden

18 Oct 2021 18:37 UTC

82 points

9 comments10 min readLW link

OpenAI: “Scaling Laws for Transfer”, Hernandez et al.

Lukas Finnveden4 Feb 2021 12:49 UTC

14 points

3 comments1 min readLW link

(arxiv.org)

Prediction can be Outer Aligned at Optimum

Lukas Finnveden10 Jan 2021 18:48 UTC

15 points

12 comments11 min readLW link

Extrapolating GPT-N performance

Lukas Finnveden18 Dec 2020 21:41 UTC

112 points

31 comments22 min readLW link 1 review

Formalising decision theory is hard

Lukas Finnveden23 Aug 2019 3:27 UTC

17 points

19 comments2 min readLW link