I’m doing Budget Inkhaven again! (I didn’t realize last time that “Halfhaven” also meant specifically shooting for half-length posts, too.) I’ve decided to post these in weekly batches. This is the fifth of five. I’m posting these here because Blogspot’s comment apparatus sucks and also because no one will comment otherwise. Just like last time, this is the last week and because there’s only two posts in it, so I’m also going to mention a few highlights and talk about what went well and what missed the mark.
Inexpensive, Not Cheap
All of these synthetics are far cheaper and just as good as their dirt-dug counterparts. When I finally get around to making beautiful crowns someday, I will make a point of it to use only synthetic gems….
Half-ass it with all you’ve got, is what I’m saying.
The ICML 2026 Seoul Survival Guide
So you’re going to Seoul for ICML. Congrats!… Here’s everything you actually need.
I’ll see you there… in the middle of a thundering monsoon as the heat breaks and the sun returns.
Highlights! I’ve picked out the handful of posts with the ~most reads after a little while, and the handful that I was most proud of. I’ll crosspost the best couple of each category at some point.
Highlights (as judged by estimated readers/day, measured on 5/12):
The ICML 2026 Seoul Survival Guide (9.54 (!))
Measurable Cognitive Factors Contributing to General Intelligence (3.04)
Factual Truth, Mythic Truth (2.80)
Eigenfruit Pie (2.38)
Don’t Replace Your Brain (High Actuation and Prosthetic Executive Function in the Age of AI) (2.05)
The Prepared Practitioner (2.00)
Deadlock in the Parliament of the Self (1.83)
All others had <1.80 readers/day, with a cluster around 1.75 readers/day comprising If I Were Emperor of New AI Safety Researcher Training…, Two Gods of Trade (A Treatise), For Grandma Kim, Who Justly Asks Why Her Post-Singularity World Has To Be So Complicated (왜 이렇게 얼마나 복잡 해야지?), and The Clustering Theory of Gender Specification and Enumeration.
Highlights (as judged by me and what I thought were my best posts):
Once again, it overlaps pretty heavily with the reader’s choice highlights, and this time, that surprises me less. A few notable posts I was particularly proud of but which don’t seem to have got as much attention as I might have expected: The Compass I Wear Which Works Only At Night (And Why the Moon Matters When All Else is Lost), More Reasons Why the First “High Dimension” is Six or Maybe Five, Yoda’s Dance/In the Hall of the Jedi Master, Why Everything is a Spring, A Taxonomy of Wizard Power, and The Last Enemy That Shall Be Defeated is Ignorance.
I’m pretty happy with… maybe two-thirds or three-quarters of these posts? There’s a few I thought were especially weak, and most of them could have used an editing pass, but I was largely happy with what I wrote about and how I wrote about it.
Postmortem! What was I hoping to get out of this? What went well, and what could have gone better? What could I have done to get more out of Budget Inkhaven?
As before, I think it went pretty well, given that I was doing this alongside a research grant, a volunteer research mentoring program, and substantial surprise problems in my working life.
I’m happy with the extent to which I got a little braver/riskier/weirder with my topic choices.
This go-round, I think, built less muscle. It hypertrophied my poster’s gland somewhat less. Diminishing returns, whatcha gonna do.
I don’t really regret not having gone to Inkhaven proper this time. I feel like if nothing else, I’m proof positive that you can avoid spending several thousand dollars and still get most of the benefits.
Just like before, I kept emptying out my post ideas folder, including a lot of ideas that I didn’t get to last time. Some of these, I actually wrote in the meantime between November and April.
“Why Conceptual Engineering is So Important” became On Conceptual Engineering.
“Homology Chains of Leftovers” became Chain Complexes of Leftovers.
I’m pretty proud of The Con Badge Problem.
Comprehensive Zendo Rules was the only major one that I specifically wrote up during this round of Budget Inkhaven. Maybe I could have mined that list harder?
A few ideas didn’t quite make the cut this time around, just like last time.
Here’s a few working titles that didn’t make it, that exist as yet only in some other possible-world: Achievement vs Renown, What Is It Like To Be A Tindalosian Fox?, The Thing About Steering Vectors and Rotation vs Vector Addition, You Can’t Stop Being What Shaped You, I Don’t Know How to Tell You That a Solution to AI X-Risk Would Involve Overthrowing the Great Powers, Asymptotics for Whether Envy-Free Division is Worth It, Emergent Conspiracies, and Tools As Natural Categories.
As always, if you really want to see one of these written, reach out to me! Maybe it’ll even change my mind.
As ever, I accept public feedback here, and private feedback at https://admonymous.co/lorxus.
Discuss Read More