Aligned to Flourish

It’s incredibly easy to be fooled by the capabilities of the current top-performing tech (LLM agents).

For 20 years, the startup mantra has been: “Ideas are worthless.

I’ve been posting more on LinkedIn recently about AI safety, AI security, and the future of AI.

Epistemic status Work-in-progress.

Epistemic status Uncertain; trying to articulate my cruxes.

Epistemic status Exploratory.

This is the talk I gave at the PIBBSS Symposium in September 2025, at the end of a summer working with PIBBSS on automating AI safety research.

There have been multiple recent calls for the automation of AI safety and alignment research.

[Nov 27, 2024] I have some new alignment project ideas I quickly wrote up.

This post is broken down into two parts: Which AI productivity tools am I currently using?

Imagine you’re building a sandcastle on the beach.

I gave a keynote talk on how we should be thinking about accelerating AI alignment (safety) research.

This is a post on a recent paper I thought was cool.

Note I wrote this post in less than a day and didn’t want to spend more time on it, but if you have further insights or clarifications, let me know!

About this series This is the first installment of AI Insights, a series of shorter posts focused on insights relevant to AI Safety.

I wrote this as a bio for EAG Bay Area 2024.

I just published a new post on LessWrong.

Wrote up a short (incomplete) bullet-point list of the projects I’d like to work on in 2023.

In March 22nd, 2022, we released a survey with an accompanying post for the purpose of getting more insight into what tools we could build to augment alignment researchers and accelerate alignment research.

As we are trying to optimize for actually solving the problem, we should not fall into the trap of learning just to learn.

I’m an alignment researcher, but I still think we should be vigilant about how models like GPT-N could potentially be used to make the world a better place.

I’ve been ruminating on an idea ever since I read the section on deception in “The Core of the Alignment Problem is…” from my colleagues in SERI MATS.

Epistemic status Quite confused.

For discussion: Link to LessWrong post.

Top Diplomacy players focus on gigabrain strategies rather than deception.

This post is currently in the Appendix of a much longer post I’m currently editing and waiting for feedback.

TLDR of what I’ve been thinking about lately: Learning is a set of skills.

“If a movement is too demanding of everyone, it will inevitably collapse.”

Note: This is an old blog post I wrote in 2019.

This is a quick post to mention that I’ll be in Berkeley attending the SERI MATS program to work on aligning language models.

I created two AI Alignment playlists on Youtube.

TL;DR: In this project, we collected and cataloged AI alignment research literature and analyzed the resulting dataset in an unbiased way to identify major research directions.

Crossposted from the AI Alignment Forum.

What this Post is About A few months ago, I gained access to a new exciting tool called Elicit from Ought.