2025 was the first year I meaningfully interacted with the world since 2021. All of 2022-2023 I was extraordinarily ill. I was almost entirely home bound. By early 2023, when I tried thinking about math for as little as 5 minutes I would have a multi-day-long crash. The illness was so bad that I could barely speak or tolerate light or sound. For a few weeks, I could lie in bed watching the clouds—resting enough to have energy to eat. It was a very hard time. Though with time I would improve and in 2025 I felt like I reentered the world.

The story of 2025

Going into 2025

My health started to improve in late 2023. Throughout 2024 I attempted engaging with the world a little. I tried working a little on computational tensor-rank problem with Austin Conner. I still had to avoid thinking too much to avoid crashing. However, I thought basic coding could be doable. It was, but the inability to think much at a time made it too hard to make real progress. I would also only work in 30-minute sessions, and then I had to take breaks for a few days. I found that I could not retain what I had done previously and spent a lot of time just reviewing. I’m infinitely grateful to Austin for his patience. The calls with him to talk about math and joke a little were a bright spot in an (unsurprisingly) depressing time of my life.

In the middle of 2024, I started recording some music for my sister’s wedding. It was similarly slow-going. The most challenging part of it was sitting upright at the piano. Due to my illness, my legs aren’t great at getting blood back to my head and I get drained quickly sitting with my legs down. This continues to be the hardest part of playing the piano for me. But after 10 years of improvising, it was overdue that I record something. My sister’s wedding was the perfect opportunity. I recorded just over an hour of music and I would play a rendition of the wedding march as my sister came down the aisle in October.

Going into 2025, I could feel my health still improving (albeit subtly). I went into Cambridge a couple times—to two lectures and to MIT’s crypto day. This proved to be beyond my health (I would have multi-day crashes after each relatively short outing) but it turned out to be well worth it. I got to talk to an old cryptography professor of mine: Srini Devadas. I said that I had been removed from the world for quite some time but I felt like I was healthy enough to do academic work again. All of 2025, Srini has been very generous with his time and has essentially become my unofficial advisor.

2025 Spring

For Fall 2025, I tried auditing MIT’s 6.5610 on applied cryptography with the permission of the instructors Yael Kalai and Henry Corrigan-Gibbs. My friend Luis audited it alongside me. Unfortunately, via my illness knocking me off track halfway through the semester and perhaps poor scheduling by me, we did not get to collaborate much. Before this, my last interaction with cryptography was MIT’s theoretical cryptography course back in 2018 (by Vaikuntanathan and Goldwasser). 6.5610 seemed to have a lot of overlap in material, and so I thought it was a good jumping off point.

Overall, the course went well. It did have a lot of overlap in material—maybe a little too much. Even though I am more theoretically oriented, I would have liked to see more of the applied side of things. My favorite lecture was Henry’s lecture on MPC work he did that actually runs on Apple devices to do private telemetry. Though this was fortunate for me in the end with the aforementioned crash. Although not quite on time, I would finish all the problem sets and the final exam.

Though it may not seem like much—taking and completing a course very similar to one I had done before—it was miles above what I could do the past few years. I still primarily worked in short chunks with large breaks between them. What changed during the course is that I did much better at remembering where I left off between breaks—this made all the difference.

2025 Summer

Most of the summer, and perhaps most of my energy during 2025, would be taken up by a romance. In some ways, it is surprising that I have had any romances. I have been chronically ill with ME/CFS since I was 14 and even at my healthiest since then I hardly left the house. When I did leave the house it was to go to mathematics or computer science classes/meetings. Yet somehow I have been able to meet some special people who also found me special to my luck and surprise

This romance was unusual for a few reasons. For one, it was agreed at the outset that it would be limited to the summer. She was going to grad school in the Fall, and it made sense to have that cut off. It was also long distance with her visiting and staying in the same place as me three times during the summer for a total period of around a month.

It was really a lovely time. Although I still wanted to do mathematics and computer science, I was ok if almost none of that happened because of the energy the relationship took. Having variable health for over a decade, and thus retreating from the world multiple times before, I have found that loneliness is by far the most difficult part of it. It is easier to give up on all of your dreams than it is to be socially isolated. For me it is, at least.

The relationship did, in fact, end up taking all of my energy. Should the relationship have gone beyond the summer, something would have had to be worked out. But the plan was for it not to, and it didn’t. I don’t regret the relationship in the slightest, but I do regret how it ended. It was hard to navigate being regular friends again after our scheduled breakup. I wish I had handled things better in that respect. I do hope we are friends again sometime.

Fall 2025

This past fall was the only period in the past four years that I would say I was thriving. My health had continued to improve and on returning to academic work I found I had more capacity. I saw friends more regularly and was not as lonely as I expected to be. I started self-studying information theory and meaningfully engaging in computer science research.

Frankly, everything was easier. Socializing was easier. Sleeping was easier. Doing mathematics was easier. My piano improvisation was better. Even my Chinese somehow improved (yet it’s still very poor). The latter two things improved with no practice put in. Becoming healthier has made more or less everything easier.

I found time for a few other extracurriculars. I put up the website you are likely reading this post on, and I put up some videos on my math YouTube channel. I even spent some time helping out at an education nonprofit that is trying to do work similar to MIT OCW.

I was, and still am, thinking with much greater clarity than a year ago. This is still not the level of health I had in 2021, or 2019,… or 2016. I still spend the vast majority of my day simply resting. However, it is a big improvement. It has been quite a shock how different things are when I am healthier though. I remember the searing headaches and the dreadful emptiness sensation of crashes post thinking about math for five minutes. I remember carefully making sure I did not think too much for fear of those crashes. To be curious yet have to avoid thinking so much was my torment. This is not something I worry about anymore.

So, where am I at now?

Well, as I talked about before, I am in much better shape, but I still have a long way to go. I track more or less every activity I do every day so that I have a good idea of how I spend my time. To give some context, last semester (counted as September 3rd - December 20th), which was by far my most active semester in four years, I averaged about 90 min/day “active tasks.” This includes research, self-study, socialization, work on personal projects (YouTube channel, personal-site, etc.), and maintenance (shower, shave, brush teeth etc.). I have a good idea of how all these categories break down, but I’d like to leave some things to the imagination. For further context, about half of the active time is social which the vast majority of is FaceTime calls. I spent 80 min/day eating.

Health

In terms of quality of life, things are much better than a year ago. This is almost entirely downstream effects from having more energy which allows me to participate in life. Being severely chronically ill, even at my healthiest life is a part-time experience. Though part-time is much better than no-time. Here are where things stand from last year.

Energy level: Very low compared to a normal person but improved a lot.
- Should keep most activities below 45 minute mark
- Longest should be around 1hr15 minutes.
- 30 minutes is a pretty good session.
Mobility: Mostly home bound but can walk around for <=1.5 minutes at a time. Certainly improved. Standing still is very hard. Sitting with legs up helps a lot. I spend most of my time half lying down/half sitting.
Medications: I’ve been trying two new medications and their contribution is hard to tease out.
Major crashes of 2025: Two large crashes amounting to ~80 days out of commission in total: ok. One was from sleep falling off and the other was because I caught strep which lasted a full month and needed two rounds of antibiotics.
Pacing: pretty good; improved
Sleep: pretty good; improved a decent amount.
Mental health: great; improved a lot
What happens in the unlogged rest of the time: Mostly watching YouTube, movies, Naruto, some meditation, and I have a sleep routine that lasts from 10:30PM to 9:30am.

How come health improved so much?

Frankly, it is hard to give any definitive answers. So little is known about ME/CFS, and it is affected by so many things. However, I do have my best guesses. These down to three to four factors: starting low, slow improvement with time, getting healthier makes it easier to get healthier, and pacing improvements.

Starting low: A great way to have huge improvements is to start from a very low point. In Fall 2021, I was very overtaxed with projects, academic work, and a terrible prolonged breakup. This would culminate in me withdrawing from Harvard that semester. Early 2022, I would try Intravenous immunoglobulin (IVIg) and had a case of aseptic meningitis. This led to some of the worst pain I’ve had in my life along with an up-all-night visit to the emergency room. I would spend the rest of 2022 trying SubQIg—which I tolerated better—but which further drained me. It made me completely out of it about half of the week. My lowest point came in early 2023 after going off a medication and finding that my Crohn’s flared up. That was when I lost my ability to speak and most of my consciousness was looking at clouds and dreading the drone of the water pump in the house.

All of this is enough to drain anyone. But for someone with ME/CFS, it is enough to cause devastating losses in energy envelope. With my heightened state of illness, I was also more isolated and thus very depressed. It may not surprise you that I did not have the best sleep either.

Slow improvement: One of the most compelling explanations for my improvement is just time and rest. It would take a lot of time to recover from these setbacks. Prolonged crashes can have years-long shadows and I might still be in this crash’s shadow now. I started Low-dose Naltrexone (LDN) in 2024 and that could also have an effect—although it is hard to tell.

Improvement begets Improvement: The effect I am most convinced of is that becoming healthier is a virtuous cycle. As everything is more difficult when you are sicker, everything is easier the healthier you are: getting healthier, included. I have seen this happen first hand. Say my energy levels improve a little and I can call a friend for thirty minutes. Then my mental health improves a little which leads to my sleep improving. Then my energy improves from better sleep and I become more aware. When I’m more aware I make better pacing decisions which improves energy levels. The better energy levels mean I do more academic work, which improves mental health, and on and on.

I have seen this feedback loop play out a lot this past year. It is an effect I am most confident has made a difference—along with simple time since overtaxing events. I don’t think this kind of virtuous loop can lead me back to the health of a normal person though. I’m not sure how cheering up will allow my legs to pump blood back to my heart more effectively, stop my immune system from dismantling my small nerve fibers, or fix the mitochondrial dysfunction increasingly found in those with ME/CFS. Though, it might be able to get me back to where I was in 2018.

Pacing improvements: Pacing is the discipline to limit one’s energy-output before hitting your limit and thus avoid crashes. It is the essential tool in ME/CFS care. In talking with relatives, a few people have attributed my health improvements to my improvement in pacing. While I do think I am better at pacing now, I think this has made less of a difference than others think.

For one, it is much easier to pace well when you have a larger energy envelope. It is easier to budget well when one has a larger budget. This has certainly been an effect. When my energy envelope is so low that even short conversations can tip over the limits, it becomes a dangerous game of choosing what you go without. Do you chat a little more to be less alone but crash hard? Do you cut it off early to avoid the crash but feel more isolated? At the budget I was previously working with, oftentimes I would both be far too lonely, and I would crash. I was kept in this desperate mode wherein making conscious or wise decisions was very difficult.

I think I have matured though. I think I am more aware of my energy and less likely to let things go over the limits they should. Although, this certainly still happens. I used to always let social interactions go on just a little too long. I felt that because I had less time than most people, that if I fully accommodated myself others would not find it worth their time to interact with me. I have mostly shed this feeling… mostly. Either way, I think some pacing improvements will be what I prioritize in 2026.

Academics/Social/Other

Worked on some research projects.
Finished a course in 6.5610 for the first time in four years
Self studied some information theory
Read on some various mathematics
Romance: The biggest thing was the summer romantic relationship
In last semester I’ve started getting social life together more
Started putting up math YouTube videos aimed at advanced undergraduates and up.

What to make of all this?

After seeing 90 minutes of “active time” per day, half of which is low-energy FaceTime calls, it is hard to reconcile that with the great boom of activity I had last semester. Its just another reminder of how little energy those with ME/CFS have I guess. I’m still happy to have had this improvement but 90 minutes/day is not something I can be fully satisfied with.

Going forward, I’m avoiding the trap that so many who track their own data fall into—that of Goodhart’s law: “When a measure becomes a target, it ceases to be a good measure.” Active minutes in is hardly a good target anyway. I’d rather have the quality and quantity of work improve as well as the depth of the relationships I have. In my next post, I’ll detail what my immediate next goals and priorities will be. Hopefully, my 2026 update will upstage 2025.