Back to home

2025 yearly review

December 26, 2025

This is the 7th installment of my yearly review. See 2019, 2020, 2021, 2022, 2023 and 2024 for previous years.

(No AI was used to write what you will be reading, apart from spell-checking. Enjoy my non-native English and human rawness.)

2025 has been a strange year. Deeply satisfying on a professional level, but personally daunting, for reasons I will not get into here. Despite this, there were several notable highlights.

Work highlights:

  • Surpassed $2M ARR.
  • The quality and chemistry of our team at Typefully is the best it has ever been. It genuinely feels exciting and a privilege to show up to work every day.
  • We tackled some of the scariest refactors ever and managed to successfully land them in production.
  • Agents have made software engineering exciting again and massively expanded the scope of the problems we can tackle as a small team.

Personal highlights:

  • Fixed my knee and shoulder after months of physiotherapy.
  • Got back into gym and tennis.
  • Read more than in 2024.
  • Found joy in working from home.
  • Met some new friends in Lisbon.

Stick around to the end for my carve-outs for some recs on the best media I consumed this year!

Work

$2M ARR

We’re not publishing MRR tweets anymore, but this is a milestone that feels worth reflecting on. What a crazy journey it has been since May 2021 — almost 5 years ago. I honestly would not have believed that we would get this far.

$0 to $100k took 6 months. $100k to $1M took 1.5 years. $1M to $2M took 2.5 years. It’s not the “usual” progression that you see shared, where growth accelerates as the numbers increase, but a few things happened in between those milestones that materially impacted our momentum and hiring plans.

I both feel extremely proud of this milestone and kind of impatient because the opportunity ahead is huge and there is so much more to do. We have a kick-ass roadmap planned for 2026 with projects that will materially affect our growth curve and propel us to new heights in 2026.

The Team

We are 5 full-time people now: 3 engineers, 1 product, 1 support. The quality of our team has never been so high, and neither has our resolve in keeping it that way.

We have made the deliberate choice of operating as a small team of talented ICs that love the craft and want to do their best work while having the ambition of scaling up and taking a bigger chunk of the market. I want more of that.

We will start next year with a new Head of Content and plan to hire at least 2 more engineers in the first 6 months of 2026, while still remaining proudly bootstrapped.

Tackling the hardest problems

I am proud we landed some huge improvements to Typefully this year, like completely reworking our auth stack, successfully decoupling from X. Now you can log in with any provider you want and use Typefully as a LinkedIn-only scheduler, if you so wish. That took an ungodly amount of work considering Typefully was born as a scrappy MVP to publish threads on X, and the tech debt built up over the years on top of the initial abstractions.

Another huge one has been our recently announced API v2 (with Zapier, MCP, and webhooks) that massively boosted how much you can automate Typefully.

These days we are working on a cutting-edge AI agent that lives in the right sidebar and has the promise of revolutionizing what writing on social media looks like, and on completely reworking our pricing so it scales more with usage. I am incredibly excited to release both of these changes to the world because of how transformational they could be, for our users and our business.

Agents

I buried the lede here. Autonomous AI Agents are undoubtedly what got me the most excited as a technologist and entrepreneur, and looking back on the year, it’s crazy how far we have come.

Back in January up until May, I was still mostly writing code by hand and accepting tab completions, then something changed. I believe this is when I switched to Claude Code and my workflow changed from writing code to commanding an army of agents towards a goal and reviewing their output. Output has been code, but also plans and ideas. Agents have been sparring partners more than code monkeys — or well, both!

It has been exhilarating trying to keep up with the pace of new models and new tools released almost every week. My workflow shifted from prompting in Cursor using the sidebar and operating mostly at the file level, to creating high-level plans with Claude Code, reviewing them carefully, and then letting agents implement them end-to-end. Research, Plan, Implement is the workflow I swear by these days, resulting in mostly one-shotted implementations that are also well-architected and will stand the test of time. This works in complex codebases and is the best workflow I have found so far to work with agents.

Zooming back a bit again, all these developments make me extremely excited to show up every day and work on Typefully. It feels like we can do more, and agency and taste have become the limiting factors when building products. Don’t get me wrong, you still need to put in some effort on some hard engineering problems, but the fact that menial tasks are almost fully automatable has been a complete game changer as a small team.

I am really looking forward to further improvements on this front in 2026, to try new tools, workflows, models, and to stay at the frontier of knowledge work as we figure out how to harness our steel and steam engine moment for a new Industrial Revolution.

Personal

Fixing injuries and my new approach to training

Physically, this year was a turning point.

Last year was pretty bad: knee tendinitis, a shoulder issue, and months of frustration. I ended up doing physiotherapy for about six months, twice a week. It was boring, time-intensive, and completely dull — but it worked. Physiotherapy is kind of magic.

More importantly, it changed how I approach training. I warm up properly now. I cool down. I shut my ego down when lifting weights. I use sensible loads, move slowly, and focus on time under tension instead of chasing numbers. Training stopped being about pushing limits and started being about staying healthy long-term.

I also brought proper cardio back, mostly through tennis. Tennis has easily been one of the highlights of the year. You sit in that 130–140 BPM range, fully present, unplugged. You can play for an hour or more without even noticing the time passing. It’s great for the heart, and I’m sure for the brain too.

Mixing up my routines and trying something new

At the beginning of the year, I was pondering an apparently smaller choice concerning whether to leave the coworking space I had been at for the past year, start working from home, and change gyms too — effectively moving to a completely different part of town. This would change the fabric of my day, my habits, and make my day-to-day life significantly different. I would also have to use a car to get to the new gym.

I was hesitant to make this choice and fairly sure I would not like it because of the compromise of having to use the car and the fact that working from home had troubled me in the past. Despite this, I pushed through, knowing the decision was easily reversible and it would not cost me a lot of money.

Anyway, long story short: this decision turned out to be crazy good. I am loving working from home. It’s more efficient, my days are compressed with less time wasted “commuting” to the cowork, fewer distractions with people having calls next door, more focus and higher quality work. This efficiency allows me to do things like go play tennis at 4pm and still get my 8 working hours in for the day, if I do an early start (8am) and cut my lunch break short. I don’t do that every day, but it’s really great to have the option. I miss people at the office and the social aspect, but my current setup feels like a better compromise for where I am at in life right now.

Anyway, this apparently banal section about switching gyms is significant because I think sometimes in life (and work too!) you need to switch things up for the sake of it, and change might bring unexpected benefits, even when you wouldn’t predict so, as it happened in this case. Fuck around and find out turns out to be excellent advice.

Carve Outs

One star means “recommended”, two stars means “highly recommended”.

Books

I read a total of 12 books this year.

Articles

Matter remains my app of choice for this after a brief flirt with Readwise.

TV

IMDB tracking this year has been consistent, so I have a very detailed list. I suggest you use their “check-in” feature to keep track of what you watch.

Movies:

TV Shows:

Podcasts

I am listening to fewer and fewer podcasts these days, mostly on Spotify and via the Il Post app (Italian newspaper):