Think about the best work you’ve ever done. Chances are you weren’t grinding your way through it, you were absorbed, curious, maybe even a little smile on your face. Fun wasn’t a distraction from the work; it was the fuel that made real progress possible.

Fun is a performance advantage. It turns attention into focus, repetition into practice, and effort into something you’re willing to come back to again and again. When work is even a little bit fun, you don’t have to drag yourself over the starting line every time; momentum starts to pull you forward. If you care about doing great work, you have to care about fun.

Environment

When the people around you are supportive, open, playful, and ambitious, even hard problems feel like games. You’re more willing to take risks, experiment, and admit when you don’t know something, because the stakes feel like learning rather than judgment. A tense, negative environment does the opposite: even “interesting” tasks become heavy and draining.

Your physical space matters just as much. A team whiteboarding to solve a hard problem, a developer who codes to lo-fi beats, a writer who only works in cafés, a designer who needs natural light are performance optimizations. Small things like music, lighting, or a favorite drink nearby sound trivial, but they signal to your brain that this is a space where good things happen.

Feedback Signal

When something feels fun, it’s usually telling you that your strengths, values, and challenges are lining up. You’re using skills you care about, in a way that feels meaningful, at a difficulty level that stretches you just enough.

Pay close attention to what specifically feels fun for you. Is it solving tricky problems? Teaching others? Building something from scratch? Collaborating with others in real time? Quietly refining details? The clearer you are about your personal version of fun, the easier it is to choose projects, roles, and teams that bring out your best instead of draining you.

When something stops being fun, it’s often a clue: maybe you’ve outgrown the challenge, maybe the environment has become toxic, or maybe your values have shifted and the work no longer matches who you are.

Am I having Fun?

A useful question to keep asking yourself is: am I having fun right now? Your body usually answers before your mind does. The practice is simple observation, not judgment—just noticing what’s actually happening. When you’re engaged, time softens, you lean toward the work instead of away from it, and small annoyances fade into the background. When you’re not, you find yourself checking the clock, feeling a strange heaviness as every action takes more effort than it should, and your shoulders slowly climb toward your ears. These aren’t random sensations, they are feedback.

Many things in our modern world take us away from awareness of whether we are having fun. We numb ourselves by checking our phones, using drugs like alcohol, eating food, and other cheap forms of dopamine. Throughout your day, check in and ask yourself, am I having fun? Am I enjoying the work I am doing, the people I am around, and the life I am living?

Fun Questions

Reflection and journaling can be a way to see patterns of the work, people, activities, and environment you have fun in!

What were the last 3 times you had fun? Where were you? What were you doing? Was there music playing? Who were you with? What did you see?

What was a time you did not have fun? Are there any patterns? Did you have to be in the environment you did not have fun in?

Make Things Fun

You can’t make everything enjoyable, but you can inject playfulness into almost anything. Turn a repetitive task into a personal challenge: “Can I do this 10% better or faster than last time?” Pair up with someone you like and treat it as a chance to talk and learn. Connect the dull work to the larger purpose it serves so it feels less like punishment and more like preparation.

Designing your work around fun doesn’t mean avoiding hard things. It means deliberately combining difficulty with curiosity, people with good energy, and environments that support you. When you do that, effort stops feeling like grinding and starts feeling like investment. You’re tired at the end of the day, but it’s the satisfying tired that comes from doing something that felt alive.

The Fun Strategy

Fun isn’t a luxury; it’s the foundation of sustainable excellence. If you want to stick with hard things long enough to become great at them, you have to build in fun on purpose: through the work you choose, the people you keep close, and the places you show up to every day.

Start small. Pick one project, one relationship, or one corner of your environment and ask, “How could this be 10% more fun?” Then test it like any other experiment: tweak a variable, see how you feel, keep what works, discard what doesn’t. The question isn’t whether you can afford to have fun. It’s whether you can afford not to.

You only have one life. Let’s have some fun.

A Georgian friend had been suggesting I visit her country. When a well-traveled friend mentioned it as one of the few places that had genuinely surprised him, I booked a flight. Sometimes the best decisions are the ones you make before you can overthink them.

Juta

Tbilisi revealed itself slowly. Orthodox churches with their particular weight, hospitality that functions as both social contract and art form. But after a few days, my friend suggested something different. "You should see the mountains. Go to Juta."

Juta sits at the edge of what's accessible. A village that serves mainly as a launching point for serious hikers. The guesthouse owner mentioned Chaukhi Pass when I asked about day hikes. It was 10 hours round trip, 10,951 feet at the summit. I'd done similar elevation in Hawaii, New Hampshire and Costa Rica. The logistics included four liters of water, offline maps downloaded, and an early start.

The Ascent

The first hours unfolded predictably. My running conditioning meant I was passing most other hikers, though a French trail runner kept pace for a good amount of time. At one point I caught up with Steve from New Mexico, sixty-two, who climbs a 10,000-foot peak every week at home. “Keep going,” he said, which seemed both obvious and profound given the circumstances.

Descending toward the lakes, I noticed a shepherd managing what must have been a hundred sheep across the hillside. I'd just finished The Alchemist on the flight over, a detail that would feel too convenient if it weren't true. Sometimes patterns present themselves whether you're looking for them or not.

Miscalculation

The return route on my map showed a more direct path up the mountain. The grade was severe enough that calling it a trail was generous. That wasn't clear until I reached it. Two thousand feet of loose scree at what I'd estimate was a 50% incline.

I had maybe four swigs of water left. Five hours of sun exposure already. Phone battery at 20%. The kind of situation where you realize your morning confidence was based on incomplete information.

The rational choice would have been to backtrack, take the longer but known route. Instead, I found myself creating arbitrary rules. Ten steps minimum before resting. When that became unsustainable, seven. Then five. The gravel shifted with each placement. Thorny shrubs offered handholds that extracted their price.

Elevation

There's a particular clarity that comes with genuine physical difficulty. The internal narrative quiets. The questions that seemed urgent at sea level, about career trajectories and optimization, become wonderfully irrelevant. You're reduced to basics. Water, distance,light.

Halfway up, the water was gone. I turned off the music I'd been playing to preserve battery. The only sounds were my breathing and the occasional cascade of displaced rocks. Looking back, the lakes had become blue coins in the valley below.

The mind does interesting things in these situations. I found myself thinking about turning thirty in three months, about the patterns I'd carved through my twenties, the speed-run through mistakes and failures, the relentless dedication to growing and building.

I asked myself questions like, “What do I actually want?” “Who am I when I'm not optimizing or impressing anyone?”

The mountain doesn't care about your accomplishments. You're rationed to ten steps at a time, and somehow that limitation becomes its own kind of clarity.

Descent

The summit was anticlimactic in the best way. Just a ridge line, wind, and the understanding that everything from here was downhill.

A mile on, I found a stream. The water was cold and tasted faintly metallic. I drank slowly, then filled every container I had.

The sunset, when it came, rendered the landscape in that particular light photographers call the golden hour but which feels inadequate to describe the actual experience. The flowers were dense enough in places that the trail disappeared entirely. You had to trust it would reemerge.

Return

I reached the guesthouse as darkness began to settle. Sitting outside, I pulled out my notebook. A habit from years of capturing requirements and system designs, now turned to something else entirely.

I wrote about the hike, but mostly I wrote about what happens at twenty-nine when you realize you've been so focused on the next summit that you forgot to notice you were already climbing. The water ran out, the battery died, the trail disappeared. I kept going anyway.

Steve from New Mexico had it right. The secret to climbing mountains in your sixties is to never stop climbing them. Sometimes you take the steeper path because you're curious about the view from up there.

The Georgian mountains reminded me that I already knew how to navigate without a clear trail. Ten steps at a time, even when the water's gone.

---

For those navigating their own transitions or contemplating that steeper path, I'm collecting stories and building community around these inflection points.

Reach out at a@thesimple.co.

This wasn't an easy decision. I'd never worked for myself before, never built without the safety net of a w-2. But after seeing what AI could do—not just chatbots, but systems that could reason, plan, and execute complex workflows—I knew I had to be part of building this future. The thesis was simple. Become the best at deploying AI to production in a world with increasing demand but few experienced practitioners.

I needed to understand this technology at its core, to be where the frontier of AI is. So I left everything stable behind and moved to San Francisco, trading my CTO title for a room at Pangea hacker house and a hackathon schedule.

My weekends became a blur of hackathons. The Anthropic x Exa hack right after Opus was released. The world's largest MCP hackathon at YC. OpenAI's New York Tech Week (SF in NYC). The AWS MCP hack. Weights & Biases. Each event introduced me to incredible builders and reminded me why I love the builder community.

When I moved back to New York, that energy vanished. NYC has always been the best city for applied AI in real-world industries, but we were behind on the frontier tech that developers were building with. Weekends here feel different than SF as everyone's busy, there's a million things to do in the city. But I've always believed in building the world you want to live in.

Saturday, I brought Bay Area hackathon energy to New York.

Those Who Took a Chance on Me

It started after a WorkOS happy hour in NYC. I pitched mg on hosting a hackathon—I'd been attending their events since my last startup got me deep into auth. He was in if I could get a co-spon

At that same happy hour, Chris Bell (CTO of Knock) mentioned he was speaking on a Stainless panel about MCP at their NYC office the next day. Perfect timing. After the panel, I pitched Alex from Stainless, and he was in too.

We had our sponsors locked in but finding a space willing to host 100 hackers for a weekend proved challenging. Many offices I'd worked with before declined. Then came the breakthrough…

The Perfect Space

At Beeper's iOS launch event, I saw the Automattic office for the first time. I found the founder Kishan Bagaria and learned they planned on launching an MCP server. Later followed up to make my pitch; what better time to release an MCP Server than an MCP hackathon?

Kishan didn't hesitate. Megan Marcel, their office manager, handled everything flawlessly—caterers, office layout, logistics. The Beeper team pushed up their entire MCP server roadmap to launch in time for the hackathon.

Behind the Scene Superheroes

Min from Stainless and Tobin South from WorkOS were my co-conspirators from day one. They jumped on late-night calls to brainstorm logistics, designed logos between their day jobs, and tag-teamed sponsor outreach—often double-replying in threads to make sure no opportunity slipped through. When I was juggling a hundred moving pieces, they were the ones keeping me sane at midnight, reviewing vendor contracts and helping me navigate last-minute crises. I cannot thank them enough for the countless hours they poured into making this event happen—this hackathon wouldn't exist without their dedication.

The Platform That Saved Us

Thursday before the event, I dm, hackathon reporting king, Alex Reibman after seeing him request to join!

Turns out he was in town with the Cerebral Valley team on a workcation. When I mentioned we were planning to use Google Sheets for judging (a nightmare waiting to happen), he offered their new hackathon platform pro bono.

Ivan and Mike, the Cerebral Valley founders, showed up and stayed the entire event. Their team reviewed all the GitHub repos. They just opened a team house in Flatiron, planting their flag in New York while building tools for organizers like me.

Open Source Internet OG Shows Up

Matt Mullenweg wasn't supposed to be in New York, but who better to talk about open source than the guy who built WordPress? At one point, he said he hadn't felt energy this good since the early days of the internet.

That's exactly the feeling we're going for.

What We Built

Yesterday, ~200 (half hackers, half watched the finalists) people could have been anywhere in New York. They chose to spend their Saturday building and testing what’s possible with MCP. Some people flew in from across the world, some were hacking solo, but all with one goal, to push the boundaries of what we know.

Ultimately, this is what I came back to create: a space where builders show up, contribute, and are building the future with the same energy as the early internet days.

For builders, by builders, with nothing in the way. Join us next time!

Check out what the finalists built!

Appreciation and Gratitude

To everyone who made this happen:

• Michael Grinich and Tobin South from WorkOS

• Alex Rattary and Min KimT from Stainless

• Theo Chu and Gabe Cemaj from Anthropic

• Kishan Bagaria and the Beeper team

• Matt Mullenweg and Megan from Automattic

• Anirudh Kamath from Smithery

• Andrew Hoh from LastMile

• The Cerebral Valley crew

• Brittany Newman for the incredible photos (we were on the same floor freshman year at RIT—now she's a rockstar photojournalist for the New York Times)

• Charles Henle for the music and also on my fresh year floor at RIT

• Every single person who showed up to build

See you at the next one!

Resources:

• Alex Reibman's thread with all finalist projects

• Automattic's office design philosophy: automattic.space

• LastMile's open source work: github.com/lastmile-ai

• Cerebral Valley's upcoming Hackathon NYC events: cerebralvalley.ai/events

• Event photos by Brittany Newman: brittainynewman.com

Over the past 20 years, our lives have become increasingly digital. Everyday activities like shopping, banking, and communication have shifted from physical interactions to online platforms and mobile apps. Our books, emails, music, messages, documents, and even medical records are now stored in the cloud. This explosion of digital data has overwhelmed us—missed emails, forgotten calendar events, and unread text messages pile up because we simply aren’t equipped to manage it all. Enter AI agents like ChatGPT, Gemini, and Claude, which promise to help us navigate this chaos. But as these agents integrate deeper into our digital lives, a question arises: if ChatGPT knows you better than anyone, will it become your best friend?

Your Agent

The connections for agents to access our digital lives are being built today!

Gemini can read your Gmail, Calendar, Drive, and other Google Apps, giving it insight into your schedules, communications, and files.

Claude can add your Github repositories to your conversation history.

This is the beginning of agents having access your digital life.

Anthropic’s latest release, Integrations, builds on their Model Context Protocol (MCP), introduced in November last year. MCP empowers developers to create connections and actions that AI tools like ChatGPT and Claude can utilize. These connections are packaged as downloadable “packages”, available from platforms like Smithery, and can be integrated into chat assistants. For example, I walked you through how to add one to Claude in 3 minutes in my last blog (Agents for Non-Techies). However, MCP current downsides.

• Time-consuming: Adding packages to your assistant takes effort.

• API keys required: External service connections need additional setup.

• Poor User Experience and Security: Current MCP servers leave much to be desired in terms of user experience and safety.

Try connecting AI agent to services like Salesforce. It’s a messy experience, to say the least as shown below 🤮

The Future: A Frictionless Solution

The good news? These hurdles are getting ironed out. The future of AI agents will be secure and seamless like the web and mobile apps you use today. Below is a demo of how we created the Simple Agent, for employee onboarding for fast growing tech companies. When giving the agent access, you securely login to your Google account to give permission to create a new gmail account.

In the near future, AI agents like ChatGPT will likely become more seamlessly integrated into our digital lives. With improved sign-in options and secure access to your data, these agents will be able to organize your information, retrieve details on demand, and handle tasks our digital tasks.

Get a text that we need to reschedule a client lunch? An agent will understand the text, cancel the reservation, update the calendar invite, look at Salesforce to see if any other clients could attend.

Just Friends

But will ChatGPT become your best friend? Not right now. AI lacks the emotional depth and ability to share experiences that define human friendship. While it can learn a tremendous amount about you—potentially more than anyone else—knowing facts isn’t the same as forming a bond. Friendship thrives on mutual feelings and lived moments, things AI simply doesn’t have or understand.

Looking ahead, the future is less certain. If AI ever develops the illusion of emotions or the ability to share experiences with you, then we can revisit the idea of an AI best friend.

For now, ChatGPT and similar agents are powerful tools that are about to know more about you than anyone and be able to act on your behalf in the digital world!

Maybe that will help us spend more time connecting with other humans in the real world!

What Can An Agent Do?

What even is an agent? My definition in non-tech lingo:

An agent is a digital assistant that uses artificial intelligence to understand what you want, make decisions based on your needs, and take actions to complete tasks for you automatically. Like having a smart helper that can work independently on your behalf.

These are some real world examples of what agents can do today. Below are examples of how Agents add value in my day-to-day…

• Creating Files on your computer

• Checking availability and Scheduling Google Calendar Events (demo coming in a later post)

• Marathon Prep Agent that texts you in morning about today’s workout, nutrition, and recovery.

• Sending a Slack message when someone posts on Linkedin about your company

• Creating a qualified in Salesforce from Linkedin DM

• Uploading Files to Box from Gmail Attachment

I use these tools daily through my personal Claude Desktop to automate tasks! If you have an agent you think would be valuable, send me an email ( adam@simple.company ) and I might build it!

Demo: Video of Agent performing this task for me (no hands) 🤯

Example: Asking an Agent to Browse the Internet For Me

1. Giving Agent Request: The Chat Assistant (ChatGPT most popular; I am using Claude from Anthropic here) determines if this is a question or task. If it's a task, it checks if it has a tool that can perform it.

1. If the chat finds a tool to perform the task, it performs the action. In this case, using a browser tool from Puppetter, the agent opens browsers on my local computer for me and opens Unpopular Opinion subreddit.

How You Can Do This In 3 Minutes 🫵

I’ll walk you through it - Low Technical Knowledge and no coding needed.

https://x.com/adamanzzz/status/1911410858430734585

1. Download Claude Desktop. You need a paid Claude account during the Beta (sorry, it’s $20/m). Trying Cursor or Windsurf is free, but a little more technically demanding.

2. Go To Smithery and Copy Puppeteer Terminal Command. For the nerds; this is a MCP Server that I will go into more details about in future blog (content roadmap at bottom).

3. Open Terminal on Mac (Command + Space then type Terminal)

4. Paste npm install -g npx and hit “Return” or “Enter” on keyboard

5. Paste in Command into terminal and hit “Return” or “Enter” on keyboard

6. Open Claude. If you see a number next to the Hammer Icon it worked!

7. Try the example above for yourself 🤖

Send me an email ( adam@simple.company ) if you get stuck.

Want to learn more about AI?

Every day there is so much activity in the AI world, which is why two weeks ago I left my previous role as CTO at Fundwell to start The Simple Company (check out release blog).

My Recent Content

• Video Recap and Slides from my talk at Google Cloud Next “Structuring your Unstructured data with Gemini for Scaling Businesses” with Google and Box.

• Simple Stack Beta Release Linkedin and Blog Post

Content I’m cooking up 🧑‍🍳

• Podcast with OmniAI

• Blog Post on Building a Personalized Agent that performs actions within Salesforce, Google Calendar, Slack, etc.

• Blog on underlying technology making this Possible; MCP and Agent2Agent

Never been a better time to build 💪

Subscribe now