New year's resolutions for 2026

The answer to the question “should I use an LLM for that?” is increasingly “hell yeah!”, for almost any given “that”.

In the last three or four weeks, as resolutions for the new year were slowly taking shape in my head, it hadn't even crossed my mind to draw inspiration from a source other than my own ideas (and from a few friends, like Fidel). To be honest, I wasn't excited about the resolutions I was coming up with: they felt narrow and a direct continuation of last year's. Then, three days ago, in the middle of the night, I suddenly realised that I am so present on the interwebs, and I have shared so much information over the years about interests, hobbies, accomplishments and failures of mine, that a chatbot should be able to “think” about me and expand my boring set of 2026 goals with some personalised suggestions that I could use!

Yesterday's beautiful morning, hiking with the family not far from home

And so I told Claude, ChatGPT, Copilot and Gemini:

“Help me come up with good new year's resolutions for 2026.

This is me: [demographics and personal information]. Use the information on my personal web page to find out more about me: https://tripu.info/

My resolutions for 2025 were these: https://blog.tripu.info/2025-resolutions . And this is how well I did on those resolutions at the end of the year: https://blog.tripu.info/2025-recap

These are the resolutions I have already decided for 2026: [short list of existing ideas].

Suggest other resolutions to complement these, based on who I am, my situation, and my hobbies. Cover different aspects of life (health, finances, family life, learning, experiences, work, etc). Suggest resolutions that align well with the ones I did last year and with the ones I've already decided for this year -- but suggest also resolutions that are outside my usual ideas and that might give me fresh perspectives, add variation to my routine, or expand my goals, experiences or skills in some way.”

The results were quite interesting, in general. Gemini was the least helpful. I give Claude the highest mark; it was the most thoughtful and accurate in its assessment. These were some of its ideas:

• Claude decided, on its own, to read my site Vīta “to understand more about [my] projects and interests”. Very nice.
• It understood my failure(s) to study Japanese and provided some useful ideas (switch textbooks for daily Anki; aim at JLPT N4 in December, not July).
• It told me to visit Japan, “if finances and family logistics allow”. Sigh. That is extremely unlikely this year, but spot-on as an aspiration for me.
• “Your 2025 mid-year checkpoint was useful but came too late to course-correct […]. Set calendar reminders for April 1, July 1, and October 1 to review progress. Adjust goals that aren't working rather than carrying them as dead weight all year.” So true.

A few suggestions were offered by all, or most, of the models:

• Three out of four models told me that my kids are at the “golden age/window/years” and that I should make the most of that; eg: go hiking with them more often (Gemini told me they're “old enough to walk decently, young enough to still think hanging out with you is cool”. Ouch!).
• The models advised me to take up manual, tactile work or hobbies, for a change: woodwork, bread baking, calligraphy, etc.
• Participate in some event like Ludum Dare (“you have an account!”; that is right), NaNoWriMo (now defunct, but they apparently don't know that), or Inktober (like Koalie).
• Complete a “deep dive” into some specific topic; eg: history of Madrid, Stoicism, Japanese aesthetics. Read two or three books, and write a synthesis post.

Other good ideas:

• Contribute regularly to an open‑source project.
• One screen-free day per week. One no‑screens evening per week.
• Declutter 100 items. Optimise one system each month.
• Meditate.
• Stretch.

In the end, incorporating some of those ideas, here go my resolutions for 2026.

Reading

This year, for a change, I resolved to read mostly short books.

I have already collected eight physical books that I wanted to read for one reason or another and that are ≤200 pages long. I'll always find something lovely in the paper format, and these are books that I had already bought or that were lying around at my parents', so zero euros and no DRM crap. I'm looking forward to reading a bit of original Italian again (Italo Calvino), something I don't get to do often.

On Goodreads, I committed to reading at least ten books this year.

These are quite modest goals, given my track record. They don't even feel like goals, more like indulgence.

I will probably also read a few more issues of Saga, and I might start with some of the heavy items on my to-read list, like Dostoevsky or Saint Augustine — but I make no commitment to those this year.

I also decided that I'll download some long-form blog posts that I have been saving over the years on Pocket Wallabag and read them on the reMarkable. It's mostly long essays or complex explorations of some topic that really interest me — but I don't like reading those on a computer screen, and so they keep piling on. It's mostly philosophy, economics, science, and Astral Codex Ten posts or other similarly brilliant stuff from the rationality blogosphere.

Writing

My unfinished short story will stay short, but it will stop being unfinished in 2026. I made sure the latest draft is on my reMarkable and I'll use some of the slots I usually dedicate to reading (commute time, bed time) to write by hand instead.

I reckon I've written more than half already, and my goal is to have the first version ready to start requesting feedback and polishing it, so I intend to finish the draft during the first quarter of the year.

Drawing and painting

One of the presents I received from the family this Christmas is this sketchbook.

Last year, for some reason, I felt compelled to rescue this textbook from my parents' (is it yours, brother?)

I take those as signals. Of Something.

I should take up drawing and painting again. I used to like it. Will I have this year the discipline, and the time?

(And does this hesitancy even count as a resolution?)

Nutrition

I can't take Bryan Johnson seriously; I think he's either a clown, delusional, or a grifter. But I appreciate this post of his that Fidel shared. I like the sound of “don't die” better than the alternative.

In 2026 I'll try (again) to be a better vegetarian and

“to make a conscious effort to plan meals more often, […] to eat a little better, [and] to cut on bad food”

Exercise

Keep on running and weight-lifting throughout the year. I want to work out at the gym at least twice a week (should be thrice), and do it better this time.

I might sign up for a 10K or a 21K again, but it's not a goal.

It'd also be great if I take up a third sport, like swimming or rock-climbing.

Work

I'm starting at a new company very soon and that'll keep me extra busy, at least for a while.

And yet, I hereby commit to announcing at last the business project I've been working on in stealth mode for the past six months, intermittently. I'll publish the web site, the material I've been working on, and its associated social media accounts.

Since that's long overdue, and already close to completion, I'll do that by the end of February.

Philanthropy

This is not a new resolution, but an old one: in 2026 I'll give away 10% of my gross income to charities working on poverty and global health. It'll be my seventh or eighth year in a row.

In my case, that's usually an amount in the upper half of four figures. Or, put another way, around two lives saved per year.

In Effective Altruism circles, being vocal about one's impactful donations is not only not taboo, but even encouraged — and for good reasons. So here I am reminding my three regular readers that we are rich, that we're incredibly lucky, and that we ought to do a lot more to help the most unfortunate.

Private resolutions

Promising a few more things to myself, privately. I keep track of these codes elsewhere. I may or may not report on results at the end of the year.

• #B daily
• #D weekly
• #J
• #P

Discarded ideas

• No Japanese-learning goal this year. I should be realistic, and I have enough commitments already.
• I will not even try to publish a webcomic about Effective Altruism and the Rationality community. The idea sounded great in my head: a simple horizontal stripe with few panels, stick figure-style, poking fun at some of the tenets and tropes treasured by my fellow nerds, published every other week or so. I would love to follow a webcomic like that myself. But then I realised that, well, I actually had no good ideas or jokes. I chatted with Claude a bit and I was disappointed by the storyboards it produced: either utilitarianism and randomized controlled trials are incompatible with humour (that could be the case), or it's harder than I expected to come up with good jokes. (But then, how come SMBC is so good!?)