LESSWRONG
LW

454
Trent Hodgeson
1333220
Message
Dialogue
Subscribe

AE Studio is a team of 160+ programmers, product designers, and data scientists focused on increasing human agency through neglected high-impact approaches. Originally successful in BCI development and consulting, we're now applying our expertise to AI alignment research, believing that the space of plausible alignment solutions is vast and under-explored.

Our alignment work includes prosociality research on self-modeling in neural systems, with attention schema theory in particular, self-other overlap mechanisms, and various neglected technical and policy approaches. We maintain a profitable consulting business that allows us to fund and pursue promising but overlooked research directions without pressure to expedite AGI development.

Learn more about us and our mission here: 
https://ae.studio/ai-alignment

Posts

Sorted by New

Wikitag Contributions

Comments

Sorted by
Newest
No wikitag contributions to display.
I am worried about near-term non-LLM AI developments
Trent Hodgeson1mo6144

To the degree worries of this general shape are legitimate (we think they very much are), seems like it would be wise for the alignment community to more seriously pursue and evaluate tons of neglected approaches that might solve the fundamental underlying alignment problem, rather than investing the vast majority of resources in things like evals and demos of misalignment failure modes in current LLMs, which definitely are nice to have, but almost certainly won't themselves directly yield scalable solutions to robustly aligning AGI/ASI.

Reply1
AE Studio is hiring!
Trent Hodgeson5mo*20

Thanks Lucius, yes, this was tongue-in-cheek and we actually decided to remove it shortly thereafter once we realized it might not come across in the right way. Totally grant the point, and thanks for calling it out.

Reply
20AE Studio is hiring!
5mo
2
81Mistral Large 2 (123B) seems to exhibit alignment faking
Ω
6mo
Ω
4
162Reducing LLM deception at scale with self-other overlap fine-tuning
Ω
6mo
Ω
46
68Alignment can be the ‘clean energy’ of AI
7mo
8
208Making a conservative case for alignment
10mo
67
100Science advances one funeral at a time
10mo
9
91Self-prediction acts as an emergent regularizer
Ω
11mo
Ω
9
77The case for a negative alignment tax
1y
20
226Self-Other Overlap: A Neglected Approach to AI Alignment
Ω
1y
Ω
51
27Video Intro to Guaranteed Safe AI
1y
0
Load More