The tradition continues!

How were my predictions for 2025?

Terrible. Maybe my worst year yet, unclear.

Bun will encounter another controversy

I suppose this did happen. There were quite a few Bun controversies in the last year but we ended the year with quite an interesting twist. Bun got acquired by Anthropic. Given the layoffs that happened at Bun after the acquisition, my personal bet is that Anthropic is building some sort of Lovable clone with Claude using Bun as the runtime / api layer. I suspect they will fine-tune Claude for Bun for this purpose.

AlphaXiv will hit 100k users

It did! 130k to be precise. I think we probably could have been pushing more and I suppose we really should be. It's been great to see adoption in various spaces and general product improvements.

I will have more free time

Wrong. Most wrong I've ever been. Founding a startup means you're working like two and a half full time jobs. That's just life.

That being said, I have a lot of fun working on stuff at alphaXiv and some of the projects we're cooking up internally are quite exciting.

I will continue to use my m1 macbook air as my primary computer

Nope. As the company grew it became evident that we needed to get work laptops¹, and given that I'm always working I almost never use my m1.

What happened in 2025?

Looking back, I realize quite a lot actually happened this year. Small changes over a long time or whatever.

• I graduated from Berkeley. Not a lot of fun to do a full-time masters while founding a startup.
• I became a founder for alphaXiv! It's a really long story but I now fill the CTO role, and the team is now quite large at around 10 people. We also raised our $7m seed round, got featured in the new york times, and migrated our codebase from legacy slop to a really nice modern codebase (working on a blog post about part of that work!).
• I moved to San Francisco, with one of my good friends!
• I created my first linux userspace! (repo to be made public)
• Wrote my first minecraft server from scratch in JS (repo to be made public)
• Learned machine learning fundamentals, and learned some of the fundamentals of MLX
• Trained my first large language model... on a laptop
• I wrote a goated agents library, for alphaXiv. It's not publically launched yet or documented but when we get around to it it'll be really awesome.

What about my goals for 2025?

Have 2000 Github commits (33% more than my current amount)

I think I hit this. Unclear, GitHub's UI is not great but I think I was tracking it based on the contribution graph last year. By that definition, I did it!

Have walked 50 miles in a row

I have unfortunately not gotten to do a long walk in the last year. You hate to see it.

Create and release a new card game

It didn't happen. I think this depended on having vastly more free time than I ended up having, but it's still something I'm interested in doing. Perhaps more in the board game rather than card game direction.

Continue the journaling

I unfortunately locked out here. Journaling is super important but I haven't been able to do it consistently. I'll get it some day.

What are my predictions for 2026?

Okay time for the yearly round of random predictions.

Deno will have some sort of AI product

I still like the Deno runtime quite a lot, and I think a lot of the stuff they're working on, like Deno Deploy sandboxes are quite interesting. I think it's a matter of time before they ship some sort of AI (specifically LSMOL²) integration into Deno Deploy.

We will see the collapse of a major LSMOL company

It's well recognized in the space right now that we are pouring in way more money than can possibly be captured. I think we'll see the collapse of SOME major player next year (though you could argue that this already happened with WindSurf).

Zed will get acquired

Kind of a random pull but I've been using Zed since last year, and at the end of the day it is a VC-backed platform. I have somewhat strong conviction that it will be acquired by someone at some point, and I think it's somewhat likely it'll be in 2026.

alphaXiv will hit 1m users

Frankly, it was feasible for us to grow much faster than we did last year had we decided to focus on that. I think in the next year it's likely we will be pushing more and more for these numbers as well as some internal projects.

Bun will have development halted or very slowed

My suspicion is that Bun will not continue development at the pace that it was going at pre-development. Not even close, if I had to guess. I think Bun will likely become more unstable over time and I suspect users will begin to notice.

Goals for 2026

For 2026, I think I have different personal metrics than last year. Aspirational stuff:

• Release my linux userspace
• Create a new board/card game and release it
• Get alphaXiv to 1m users or have one of the other projects we're working on get received well
• Get good at journaling

Acknowledgements

Big shout out to the many goats I've met this year, as well as the people who have made the difference both personally and professionally over the last year. As always a huge thank you to my family, my friends, and my mentors.

Really want to shout out Angela <3, Angela, Bryan, Oleks + Gus + the rest of the boys, and the growing alphaXiv team. It's been an amazing year and I'm excited to see how each of you grow next year.

Conclusion

2025 was fun, I learned a lot. Cheers to a good 2026.

Footnotes

1. m4 macbook pro w/ 48gb ram (for next.js originally, now it's just for fun) ↩

2. I prefer the term Large Statistical Model Of Language over the standard LLM because it more accurately describes the process. I also think LLM is kind of over-talked about so I'm LSMOL maxxing. ↩