Summer 2026: What Actually Needs to Happen
I have a game sitting in a folder on my hard drive that I have been meaning to finish for the better part of three years. Every summer I tell myself this is the one. Every summer something comes up and I end up closing the folder around September with a vague promise to get back to it. This summer, I genuinely mean it.
Writing this out publicly because apparently that is the only accountability mechanism that works for me. It worked well enough with the reading list. Most of it, anyway.
The Game
I have the skills to do this now. That is the part that makes not finishing it a personal defeat in a way it wasn’t a couple of years ago. A few years back I could have made peace with the gap between what I wanted to build and what I was capable of building. That gap is smaller now, and that is no longer an excuse.
This is the top priority. Everything else on this list is conditional on this happening first.
I am not going to go into specifics yet. Mostly because I want to save that for when there is something to show alongside the post. The last thing I need is to write a whole announcement and then be back here in September with another extension.
Thesis: Finding the Lane
Graduate thesis looms. I have been genuinely enjoying the process of finding an area of focus in graphics and right now the thing pulling me in most is Neural Rendering.
The DLSS 5 reception was mixed, to put it generously. I have seen the dunks. I understand the arguments. But I think a lot of the criticism was aimed at the wrong target. The question is not whether you want neural reconstruction replacing rasterized geometry in every situation, it is whether the technique earns its keep in specific cases. Frame generation on constrained hardware feels like one of those cases. Arm were already pointing in this direction at GDC, specifically for mobile and lower-end platforms, and that is a problem space I find a lot more interesting than chasing the high-end benchmarks.
The goal this summer is to find the niche within neural rendering that is worth spending the next year on. Not the applications that have already been covered to death.
Siggraph
Got accepted as an SV (Student Volunteer) for Siggraph this year and the timing honestly could not have worked out better. I am actively getting into the weeds of graphics research and being surrounded by the people who are actually doing it is exactly the kind of environment I need right now.
The lesson from GDC was clear though: if you go in without a plan you end up wandering. I am not doing that at Siggraph. I need to be proactive about which talks and sessions I want to get to before I am actually standing in the convention center. More on this as it gets closer.
Getting Off Windows
This one gets its own post because there is too much to say and some of it is still actively in progress. The short version: I am done. The Copilot button on my laptop was the final insult.
More to follow.
A Bit of Leetcode
Lowest priority on the list, but it is on the list. I have noticed that leaning heavily on AI for most of my implementation decisions over the past year has quietly dulled some of my instincts around algorithm design and problem planning. Not dramatically, but enough that I catch myself hesitating in moments where I should not be.
The goal is not a grind. Just enough to keep the muscle from atrophying completely.
Five things. One summer. I will check back in on how many of them actually happened.