This is a linkpost for https://www.aria.org.uk/wp-content/uploads/2024/01/ARIA-Safeguarded-AI-Programme-Thesis-V1.pdf

The programme thesis of Davidad's agenda to develop provably safe AI has just been published. You can find extra details by downloading this doc. For context, Davidad is a Programme Director at ARIA who will grant somewhere between £10M and £50M over the next 3 years to pursue his research agenda.

It is the most comprehensive public document detailing his agenda to date.

Here's the most self-sufficient graph explaining it at a high-level although you'll have to dive into the details and read it several times to start grasping the many dimensions of it.

I'm personally very excited by Davidad's moonshot that I currently see as the most credible alternative to scaled transformers, which I consider to be too flawed to be a credible safe path, mostly because:

Ambitious LLM interpretability seems very unlikely to work [...]

---

First published:
February 1st, 2024

Source:
https://www.lesswrong.com/posts/qweXJ6v9heSn4wvdk/davidad-s-provably-safe-ai-architecture-aria-s-programme

Linkpost URL:
https://www.aria.org.uk/wp-content/uploads/2024/01/ARIA-Safeguarded-AI-Programme-Thesis-V1.pdf

---

Narrated by TYPE III AUDIO.

[Linkpost] “Davidad’s Provably Safe AI Architecture - ARIA’s Programme Thesis” by simeon_c

LessWrong (30+ Karma)