Edouard Harris

Karma: 542

Co-founder @ Gladstone AI.

Contact: edouard@gladstone.ai

Website: eharr.is

Defining capability and alignment in gradient descent

Edouard Harris5 Nov 2020 14:36 UTC

22 points

6 comments10 min readLW link

AI takeoff story: a continuation of progress by other means

Edouard Harris27 Sep 2021 15:55 UTC

76 points

13 comments10 min readLW link

AI Tracker: monitoring current and near-future risks from superscale models

Edouard Harris and Jeremie Harris

23 Nov 2021 19:16 UTC

64 points

13 comments3 min readLW link

(aitracker.org)

Instrumental convergence in single-agent systems

Edouard Harris and simonsdsuo

12 Oct 2022 12:24 UTC

31 points

4 comments8 min readLW link

(www.gladstone.ai)

Misalignment-by-default in multi-agent systems

Edouard Harris and simonsdsuo

13 Oct 2022 15:38 UTC

19 points

8 comments20 min readLW link

(www.gladstone.ai)

Instrumental convergence: scale and physical interactions

Edouard Harris and simonsdsuo

14 Oct 2022 15:50 UTC

15 points

0 comments17 min readLW link

(www.gladstone.ai)

POWERplay: An open-source toolchain to study AI power-seeking

Edouard Harris24 Oct 2022 20:03 UTC

27 points

0 comments1 min readLW link

(github.com)

Inverse scaling can become U-shaped

Edouard Harris8 Nov 2022 19:04 UTC

27 points

15 comments1 min readLW link

(arxiv.org)