Geoffrey Irving announced his departure from the UK AI Security Institute to return to the Bay Area and establish a new nonprofit dedicated to AI alignment research. He joined AISI as Research Director in April 2024 after leaving DeepMind in December 2023. During his two-year tenure he contributed to work on scalable alignment and model transparency.
A bittersweet announcement! For family reasons, I will be leaving AISI soon to move back to the Bay Area. I will be starting a new nonprofit alignment research org (more to come).
I will miss this place! Here are some reflections about my time at AISI. 🧵❤️
Users congratulated Geoffrey Irving on leaving AISI to launch a new nonprofit AI alignment organization, praising his service and expressing excitement for his future contributions.
Geoffrey left DeepMind to serve in government. For 2 years, he has done work of national and global consequence - at genuine personal cost. It is the kind of choice more of the best technical minds should be making. Thank you, Geoffrey - Britain has been fortunate to have you!
A bittersweet announcement! For family reasons, I will be leaving AISI soon to move back to the Bay Area. I will be starting a new nonprofit alignment research org (more to come).
I will miss this place! Here are some reflections about my time at AISI. 🧵❤️
A bittersweet announcement! For family reasons, I will be leaving AISI soon to move back to the Bay Area. I will be starting a new nonprofit alignment research org (more to come).
I will miss this place! Here are some reflections about my time at AISI. 🧵❤️
The worry about sacrificing future career advancement is super fake, FWIW. People do worry about that, but like, I’m not that worried that I came out of this with a worse reputation and neither should others. :)
Full reflections in blog form if you don’t like tweet threads!
https://naml.us/post/reflections-on-aisi
A bittersweet announcement! For family reasons, I will be leaving AISI soon to move back to the Bay Area. I will be starting a new nonprofit alignment research org (more to come).
I will miss this place! Here are some reflections about my time at AISI. 🧵❤️
I left DeepMind to join the UK AI Security Institute as a Research Director in Dec 2023 and started in April 2024 (time for family travel, wandering around London, and interval arithmetic). The past 2 years have cemented AISI as my favourite job to date!
https://github.com/girving/ray-render
Full reflections in blog form if you don’t like tweet threads!
https://naml.us/post/reflections-on-aisi
You're one of my favourite people, Xander! I will be sad to get fewer hours with you!
@geoffreyirving Thanks for your exceptional work -- you have shaped the world.
A bittersweet announcement! For family reasons, I will be leaving AISI soon to move back to the Bay Area. I will be starting a new nonprofit alignment research org (more to come).
I will miss this place! Here are some reflections about my time at AISI. 🧵❤️
@geoffreyirving ❤️
@geoffreyirving Thanks for your exceptional work -- you have shaped the world.
AISI is a hugely impactful place, especially for Research Director-level folk! We have a deep bench of excellent, driven talent; the ability to drive research on catastrophic risks or large-scale societal impacts that translates unusually directly into policy and action; …
I have mixed feelings! It is right for family reasons and I’m excited for my next phase, but I love AISI and will miss it! My reasons to join stand: safety progress is slow, coordination is under-resourced, and governments have a key role in coordination.
https://www.aisi.gov.uk/blog/why-i-joined-aisi---geoffrey-irving
Thank you, Ian! It took Ian a couple tries to convince me to join, and I've very thankful you persevered. A shame in hindsight I didn't get to the right answer a few months earlier in summer 2023! ❤️
Geoffrey left DeepMind to serve in government. For 2 years, he has done work of national and global consequence - at genuine personal cost. It is the kind of choice more of the best technical minds should be making. Thank you, Geoffrey - Britain has been fortunate to have you!
Likewise, Joseph! It was very fun to help you design and build out the Model Transparency Team, and it’s a neat example of how AISI mixes research (what techniques will work) with the risk mitigation story (how will the decision process handle uncertain scientific evidence). ❤️
Geoffrey is an amazing leader and colleague. It is bittersweet to see him leave AISI, but I’m excited about his new research initiative!
A bittersweet announcement! For family reasons, I will be leaving AISI soon to move back to the Bay Area. I will be starting a new nonprofit alignment research org (more to come).
I will miss this place! Here are some reflections about my time at AISI. 🧵❤️
Geoffrey's work on AI alignment has been an inspiration, all the best for your new endeavors in the bay area!
A bittersweet announcement! For family reasons, I will be leaving AISI soon to move back to the Bay Area. I will be starting a new nonprofit alignment research org (more to come).
I will miss this place! Here are some reflections about my time at AISI. 🧵❤️
A bunch more reflections, but awkward to chop into 280-character form. Please see my personal post for the rest!
https://naml.us/post/reflections-on-aisi
…and lots of access to very senior government decision makers who want to understand AI and influence its course. Please reach out if interested! I will continue on in an advisory capacity at AISI and am eager to pitch interested folk. :)
@geoffreyirving excited about your new org, look forward to running into you more around the bay
A bittersweet announcement! For family reasons, I will be leaving AISI soon to move back to the Bay Area. I will be starting a new nonprofit alignment research org (more to come).
I will miss this place! Here are some reflections about my time at AISI. 🧵❤️
@geoffreyirving @boazbaraktcs Welcome back to SF and thanks for all your service!
A bittersweet announcement! For family reasons, I will be leaving AISI soon to move back to the Bay Area. I will be starting a new nonprofit alignment research org (more to come).
I will miss this place! Here are some reflections about my time at AISI. 🧵❤️
a big loss 😭
A bittersweet announcement! For family reasons, I will be leaving AISI soon to move back to the Bay Area. I will be starting a new nonprofit alignment research org (more to come).
I will miss this place! Here are some reflections about my time at AISI. 🧵❤️
@geoffreyirving Grateful that AISI caused our paths to cross :) I’ll miss occasionally seeing you around, tho not nearly as much as Benjamin will miss you!! Good luck with the new org 🥲
@geoffreyirving sad to see you leave AISI, Geoffrey! Really enjoyed reading about your and Yoshua Bengio’s important work in @scmallaby’s Infinity Machine.
Hope your next venture will continue this important work.
Geoffrey Irving announced his departure from the UK AI Security Institute to return to the Bay Area and establish a new nonprofit dedicated to AI alignment research. He joined AISI as Research Director in April 2024 after leaving DeepMind in December 2023. During his two-year tenure he contributed to work on scalable alignment and model transparency.
A bittersweet announcement! For family reasons, I will be leaving AISI soon to move back to the Bay Area. I will be starting a new nonprofit alignment research org (more to come).
I will miss this place! Here are some reflections about my time at AISI. 🧵❤️
