Lorxus has applied to the SERI ML Alignment Theory Scholars Program. They spent most of 2023-04-08 and 04-09 writing their answers to the application questions.
The market resolves YES if Lorxus gets an interview with SERI or is hired without interview. It resolves NO if they are rejected out of hand. The market may thus resolve early, and without warning; the close date is expected to be well after the point by which they will have heard any rejection or acceptance.
A few relevant facts:
-SERI MATS has said that it will announce its decisions "by mid-late May".
-Lorxus has recently (~1.5 y) completed their math PhD. They have published a single preprint on ArXiV. They know category theory.
-Lorxus believes the application questions to have been well within their abilities, and genuinely believes their answers were good, to the best of their ability.
-Lorxus has spent a lot of time in the ratsphere, and has a standing interest in AI alignment.
-Lorxus applied to work with both Evan Hubinger and John Wentworth. They are working on a third application to Vivek Hubbar. (EDIT: Finished the application to Hubbar last night.)
-Lorxus will answer further questions of fact on their relevant personal details in the comments in exchange for enough marbles. (Enough might be zero.)
Insider trading is welcomed on this market.
1,000๐ Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | แน122 | |
| 2 | แน84 | |
| 3 | แน31 | |
| 4 | แน19 | |
| 5 | แน18 |
People are also trading
@AlexbGoode I generally don't - only on the rare occasions when I'm writing something for anyone else where I'm the topic.
@Lorxus at least last time John was not as selective for the online program as the other Mentors.
@Tassilo +you seem more than qualified. On the other hand I don't have a good impression of how many applicants there are, so if you hadn't applied to John's stream I would not have been as confident.
@Tassilo I will pay you 50 marbles for a finer breakdown of your current genuine best guess of p(Wentworth) vs p(Hubinger), with the reminder that they need not sum to your best guess of p(Lorxus gets accepted to SERI MATS).
@Lorxus hm.. .I think I might have been a bit under confident before. Maybe p(Wentworth) = 85%, p(Hubinger) = 40%.
@Tassilo Oh right. I got into John's stream, but there was no interview. I was not entirely sure how you would resolve this due to technicalities.
@Tassilo
> I got into John's stream, but there was no interview
That was in the last program.