I’ve gained at least 5 additional very productive work hours per week due to better sleep and exercise habits that Carson helped me plan for and stick to. I’ve gained many more unquantifiable hours of productivity due to better planning, working more on the things that are useful, and stopping unjustified procrastination.You and him will explore your problems together and figure out a way to solve them. Many of these are instances where you’ve noticed that you could have done better, but haven’t managed to extract a general lesson yet.I strongly recommend trying out a few sessions with him and seeing whether it’s a good fit! Especially so if you don’t already do weekly reviews and reflection about your research process, or your weekly reviews are perfunctory. 

I had a [coaching] meeting with Carson most weeks of the research part of SERIMATS. I used the meetings to figure out how to best make research progress on AI alignment, e.g. by getting Carson's help on the following questions:
1. whether me doing conceptual work or me coding would be better for my team's research progress;
2. how much value various formal research collaborations and informal discussions provide;
3. which issues were causing me to be stuck-ish while writing a research post and how to fix those;
4. which of a number of career options for alignment I should choose, with e.g. the subquestion of how to find out whether a complete graduate-level mathematical toolkit would be helpful;
I'm guessing that on average, I made the same amount of progress during one 1-hour meeting with Carson as I would have made by thinking about the same problems alone for 3.3 hours. (I did not spend a huge amount of time thinking about what this multiplier is, so please don't take it very seriously. I could imagine easily being convinced that it is significantly larger or significantly smaller.) Before my first meeting with Carson, I found it somewhat difficult to imagine how useful research advice could be provided without going into all the technical details of the project I was working on. This contrasted with how actionable/useful the advice received in each meeting ended up being! For example, I left [the meeting where we discussed me being stuck-ish on a post] with a concrete plan for how to order writing certain sections in the post that previously appeared to circularly depend on each other.

I found [Carson's coaching] extremely useful during SERI-MATS. As a wild estimate I would say it made me 30% more effective. Most weeks I went into [coaching with Carson] with a jumbled assortment of thoughts pointing in some direction, and came out with them organized enough that I could act on them, on several occasions significantly changing my plan for the next few days’ work. I think the reason I found it so useful is that Carson had both (i) some background context on my work, and (ii) skill/willingness/time to help me think things through, while the people I usually talk to on any given day just have one or the other (or neither). 

[Carson's coaching] was a very useful addition to SERI MATS for me, and was well worth the time spent. I was able to talk through tradeoffs, get some useful tips for approaching maths problems, and figure out priorities for SERI MATS and beyond. I would gladly pay my own money to continue this service in the future if needed.

For me, Carson covered a significant part of what I would like to get out of having a mentor: Regular check-ins, finding bottlenecks in my research and having ideas for un-blocking them. I particularly appreciated his careful listening and reframing things which made me able to look at problems from angles I hadn't seen before. One of the frames he gave me ended up being the reason why I pursued the project topic that I did.