New Year, Same Foundation: Why I Don't Do Resolutions

Another year begins, and I find myself less excited about “new year, new me” proclamations. Maybe it’s age, or maybe it’s finally understanding that real growth doesn’t reset with the calendar. The person who went to bed on 31 December is the same person who wakes up on 1 January, and any plan that ignores that starts from a false premise.

The resolution trap

I stopped making New Year’s resolutions a few years back. Not because I don’t believe in self-improvement, but because I realised I was setting myself up for failure by treating the new year as a magical reset button. Technical goals with milestones are fine; I set mine out in setting technical goals that actually work. A goal with a plan behind it is a different thing from a midnight promise.

The problem with a resolution is its shape. It’s usually a big, binary promise (“get fit”, “write a book”, “learn a new language”) made at the moment of maximum optimism and minimum information. The first missed day feels like the whole thing has failed, so it gets quietly abandoned by February. The fix is in the design of the commitment, not in more willpower. What I’ve dropped is the all-or-nothing personal promise.

What does change from year to year is my awareness. Each year I become a little more conscious of my patterns, my strengths, and my limitations. I’m learning to work with who I am rather than against it, and that’s a far better starting point than pretending to be someone else for a few weeks.

Building on what’s already there

This year, I’m not starting from zero. I’m building on:

• The parenting lessons learned from Andriel and Archael, and what I’ve written before about the importance of a father figure
• The technical expertise accumulated over years of working with Azure and AI
• The relationships I’ve nurtured in the tech community
• The habits, good and bad, that define my daily routine

That last one matters most. The bad habits are part of the foundation too, and pretending they don’t exist is how resolutions fail. I’d rather name them and work around them than declare them gone. My rule of thumb: if a habit keeps winning, change the environment rather than relying on willpower. If the phone charges outside the bedroom, late-night scrolling has nothing to grab onto; if the gym bag is packed the night before, the morning session doesn’t need a decision. Making the good option the easy one does more than any promise made at midnight.

So instead of revolutionary change, I’m looking at incremental improvement. Can I be 1% better as a father? As a developer? As a husband? For me, 1% means the size of change I’d still be doing in March, like the 300 words a day below.

What I’m carrying forward

These are principles I’ve already been living by, imperfectly, and want to keep. None of them started on 1 January.

Consistency over intensity. I’d rather write 300 words daily than promise myself a book and write nothing. A small thing done every day beats a big thing planned forever. The same is true of exercise, learning, and showing up for people.

Depth over breadth. I’m done chasing every new AI model or framework. There’s a new model or agent framework most weeks, and trying to keep up with all of it leaves you knowing a little about everything and not enough about anything to ship it well. I want to master what I use rather than dabble in everything. When something new genuinely changes how I’d build a solution, I’ll learn it properly. My test is simple: would I use it on a client project in the next six months? If not, it goes on a list, not into my evenings. I plan to review that list once a quarter, on the expectation that most items will have either matured or disappeared by then.

Presence over productivity. My kids won’t remember how many pull requests I merged. They’ll remember if I was there. Productivity is easy to measure, which is exactly why it tends to crowd out the things that aren’t.

The AI reality check

Three years ago, getting a model to call your code meant parsing free text and hoping; OpenAI only added function calling to its API in June 2023. Structured tool calls are now a baseline feature, and agents are built on them. But I’m not caught up in the hype anymore. A model can draft the answer; it can’t decide whether the question was worth asking. The thinking, the strategy, the human judgement: that’s still on us.

The trap I want to avoid is confusing access to a capable model with having solved a problem. Anyone can call an API. Knowing which problem is worth solving, what “good” looks like, and when the answer is “don’t use AI for this” is where the actual work sits.

I’m particularly interested in how AI agents will evolve this year, but I’m approaching them with measured expectations. My bar for an agent is the same as for any system: a narrow job, a human checkpoint where mistakes are expensive, and an evaluation I can rerun. I’ve written up my more detailed take on where AI agents earn their keep in 2026.

How I’ll judge the year

Resolutions are measured on 1 January by how ambitious they sound. I’d rather measure the year on 31 December by how it actually went. At the end of 2026, I’ll ask myself four questions:

• Did I show up consistently for my family?
• Did I create value for my clients and community?
• Did I maintain my mental and physical health?
• Did I learn something meaningful?

None of these has a number attached, and that’s deliberate. Each is something I can check against the year as I lived it: the weekends, the work I delivered, the health habits I kept or let slip, the things I can now do that I couldn’t in January. If the honest answer to all four is yes, it was a good year, regardless of what happened to my calendar, my inbox, or the latest model leaderboard. The rest is noise.

If you’re tempted to write a list of resolutions today, try a different exercise: write down what you’re already doing well and want to keep. That list is your foundation. Build on it.

Happy 2026.