Lessons from 2025, Bets for 2026
This is a personal piece. Not predictions, not analysis. Just what happened, what I learned, and what I'm betting on next.
2025 was the hardest and most rewarding year of my life. I moved to a new city knowing almost no one, burned out from constant travel, and rebuilt my habits from scratch. I also changed my mind about AI and finally learned how to take care of my body.
The Big Move
I moved to San Francisco.
Leaving everything behind for a city I'd never lived in was terrifying. But within six months, I made more meaningful connections than in the previous five years. Not just business contacts. People who think about the future the same way I do.
The lesson: proximity matters more than the internet wants you to believe. Being in the room changes everything.
Building at Vapi
I joined Vapi as DevRel Lead. Shipped Vapi CLI, revamped the docs, organized events, attended hundreds of conferences as a speaker and hacker.
I got to work with some of the greatest minds I've ever met. The growth team at Vapi was something else. I hope that one day I build a team that's as effective as the one I got to work with.
The biggest lesson from DevRel: your job is not to sell. It's to make developers successful, then get out of their way. The moment you start optimizing for your metrics over their outcomes, you've lost. I also learned that the best developer content comes from real frustration. Every time I struggled with something and then documented the solution, it outperformed everything I planned.
Launching NextDev.fm
I co-founded a show called NextDev.fm. A show by developers, for developers. We stream on YouTube, Twitch, X, Spotify, Apple Music, and other podcasting platforms.
The numbers:
- 60 YouTube videos
- 50 livestreams
- 750 YouTube subscribers
- 30k YouTube views
- 300k Twitter impressions
- 500 Twitter subscribers
None of this would have been possible without my cohost Ryan Vogel and our cracked video editor Mr. Danizz. Ryan is a way more energetic yapper than me and he's kept the show going strong. Go check out his content. You won't regret it. I intend to do a lot more work with him and Mr. Danizz in the future.
P.S. Mr. Danizz, you owe me a selfie together!
The Content Machine
I went all in on building in public:
- 8,000 X/Twitter followers (40M impressions)
- ~3,000 tweets
- ~50 LinkedIn posts
- 6 blog posts
- 830 commits on personal GitHub, 330 on work
The insight: consistency compounds faster than quality. I posted when I didn't feel like it. I shared when I wasn't sure it was good enough. Most of it disappeared into the void. But the few that hit made the rest worth it.
Getting LLM-Pilled
At the start of 2025, I was skeptical of LLMs. By the end, I was fully converted.
I processed 440M tokens via Cursor using 18 different models. The majority of my time was spent on Anthropic models.
What changed my mind: watching my own productivity compound. The gap between "me with AI" and "me without AI" kept widening. It's not about replacing thinking. It's about compressing the boring parts so you can think more.
Physical Upgrades
I gained 25 pounds. This is a big deal for me. I've been underweight my entire life. This was the first year I got my exercise routine and diet dialed in.
I also started getting at least 7 hours of sleep every night. Revolutionary concept, I know.
And I got a motorcycle. Because why not.
New Skills
I learned carpentry. Building physical things with your hands changes how you think about building digital things. Constraints are real. Mistakes are permanent. You learn to measure twice.
I learned hardware. Understanding the physical layer made me a better software engineer.
I learned a lot about startups. How they work, how they fail, how to build something people actually want.
I also took an interest in BCIs and learned a lot about human biology, neuroscience, imaging, and brain-computer interfaces. More on that soon.
Personal Wins
Two things I've wanted for years:
- Shaved my head.
- Got my braces off. I can finally smile with my teeth.
Small things that feel huge when you've been waiting for them.
My Bets for 2026
These aren't predictions. They're my personal, biased bets based on what I've seen up close. I could be wrong.
Voice AI eats Tier 1 support within 18 months. I'm not talking about the obvious "AI will help customer service" take. I mean specifically: TTS costs will drop below overseas labor costs this year. The math will force the switch. Companies doing 10k+ support calls/month will move first. By end of 2026, talking to a human on a support line will feel unusual.
The terminal renaissance is real, but niche. This won't be mainstream. But for the subset of developers working with agentic tools, the IDE becomes friction. I've watched senior engineers at AI companies quietly migrate back to terminal-based workflows. It's early, but accelerating.
AI + robotics demos will terrify people. The intersection of AI and robotics will make massive leaps this year. It might not reach consumers yet, but we'll get very close. The demos will be wild. Expect mainstream discourse to shift from "AI will take our jobs" to "AI will take our... everything?"
The layoff-to-founder pipeline accelerates. Established companies will keep cutting staff. But LLMs have lowered the barrier to starting something to near zero. Some of those people will discover they prefer building for themselves. We'll need more founders and fewer employees. This is directionally true for knowledge workers in tech, probably not generalizable yet.
Longevity goes from fringe to founder-mainstream. I've been paying attention to people like Bryan Johnson. Not because I agree with everything, but because the underlying thesis is compelling: optimize for healthspan, not just lifespan. The longer you live and stay sharp, the more you can contribute to accelerating technology and human progress. More founders will start treating their bodies like infrastructure. This is already happening in my circles.
What's Next
More building in public. I'm launching a Goosewin YouTube channel for personal opinions on tech. And I'm working on some new projects that I'll share when they're ready to be shared.
It doesn't feel like enough. I look at this list and think: I should have done more.
But that's probably why I'll do more next year.
What was your 2025 like? Tell me on X/Twitter.