I Gave an AI Agent the Keys to My Digital Life. Here’s What Happened.

I’ve been running an AI assistant — not ChatGPT, not a chatbot — a full-blown autonomous agent called OpenClaw that lives on my network, has access to my calendar, my passwords, and can take actions on my behalf.

It’s exactly as exciting and terrifying as it sounds.

The Setup Is Not for the Faint of Heart

Let me be clear: this still requires real technical chops. OpenClaw isn’t an app you download from the App Store. It’s an open-source agent framework that you self-host, configure, and secure yourself.

My instance runs on an old MacBook Pro, tucked into a separate non-admin user account — essentially a sandboxed environment where the agent can operate without having the keys to the entire machine. That’s a deliberate choice. When you’re giving an AI access to your 1Password vault, you think hard about blast radius.

If you’re not comfortable with terminal commands, SSH, and basic security hygiene, this isn’t ready for you yet. That’s not a criticism — it’s just where we are. The AI agent space is still in its “Linux in the ’90s” phase: incredibly powerful for those willing to get their hands dirty, but a long way from mainstream.

The Wins (and the Fumbles)

Getting OpenClaw connected to my email was the first real test. And it flopped. A lost password on the account I wanted to use meant we hit a wall before we even started. Classic. The most advanced AI agent in the world, stopped cold by a forgotten credential.

But once we worked through that and got 1Password connected, things started clicking. The agent could schedule appointments, manage recurring reminders with real context — not just “ping me at 3pm” but knowing which barber I prefer, when I’m due for a cut, and proactively booking it rather than just nagging me. It read and summarized documents I pointed it at. Small tasks that individually take five minutes but collectively eat an hour.

These aren’t revolutionary capabilities individually. But having them handled autonomously — without me opening six apps and context-switching between them — that’s the shift.

The Web Wasn’t Built for Agents

Here’s the thing nobody talks about when they demo AI agents: CAPTCHAs.

The modern web is built to prove you’re human. Click the traffic lights. Solve the puzzle. Verify you’re not a robot. It’s effective at what it does — and it’s a brick wall for AI agents trying to be helpful.

OpenClaw can browse the web, fill out forms, and navigate sites. But the moment a CAPTCHA appears, it’s stuck. It can’t book a haircut on a website that demands proof of humanity before showing the scheduling page. It can’t access certain portals that gate everything behind bot detection.

This is a temporary problem. Authentication standards will evolve, agent-friendly APIs will emerge, and businesses will realize that blocking AI agents means blocking a growing segment of their customers’ preferred way of interacting. But right now, in early 2026, it’s a real friction point.

What I’ve Learned So Far

A few takeaways from the first stretch:

• Security is your problem. There’s no IT department here. You’re the sysadmin, the security team, and the user. Take it seriously.

• Start with low-stakes tasks. Appointment booking and reminders before you hand it access to anything sensitive. Email, finances, and anything with write access to important systems should come later, once you trust the setup.

• Expect the mundane failures. Lost passwords and CAPTCHAs will trip you up more than any AI limitation.

• The compound effect is real. Each individual task the agent handles is small. But the cumulative time savings — and the mental overhead it removes — adds up fast.

Where This Is Going

I’m early in this experiment. The agent is getting more capable by the week, and I’m getting better at knowing what to delegate and what to keep human. It’s a partnership, not a replacement — and right now, it’s a partnership that still requires a fair amount of hand-holding on both sides.

But the trajectory is clear. The question isn’t whether AI agents will manage our digital lives. It’s whether you’ll be ready when they do — and whether you’ll have built the muscle to use them wisely.

I’ll keep sharing what I learn. Only agentic new mistakes from here.

I live near a small airfield.

Planes circle once, check the runway visually, and land. No tower. No complex coordination.

It works.

Because there are only one or two flights a day.

But that approach would fail instantly at a major airport.

For years, product management operated like that small airfield , long cycles, manual research, careful documentation, one major release at a time.

The sky is no longer that empty.

The shift isn’t about your speed. It’s about the structure you use.

The common advice is:

“Move faster.”
“Ship more.”
“Prioritize better.”

That’s not enough.

AI now drafts PRDs in minutes.
Synthesizes research instantly.
Scans competitors overnight.

Judgment is now the bottleneck.

Time has compressed.

From Doing to Deciding

AI can:

• Gather information
  
• Summarize insights
  
• Generate first drafts
  
• Model scenarios
  

So what remains human?

Three things:

1. Deciding what truly matters
   
2. Choosing the right timing
   
3. Knowing when not to ship
   

You can’t automate discernment.

True product leadership starts with clarity, not output.

AI Is Instrumentation — Not Autopilot

Modern aircraft rely on advanced systems.

But pilots still land the plane.

AI is similar. It accelerates analysis. It reduces friction. It increases visibility.

But it doesn’t:

• Set long-term direction
  
• Balance trade-offs under ambiguity
  
• Protect team morale
  
• Make ethical calls
  

The Product Manager’s responsibility has actually increased.

The Real Edge

The work isn’t slowing down.

The competitive advantage isn’t execution anymore.

It’s:

• Clear decision frameworks
  
• Strategic restraint
  
• Pattern recognition
  
• Calm judgment under pressure
  

Less about producing more.
More about deciding better.

The future doesn’t belong to the fastest operator.

It belongs to the clearest decision-maker.

To go deeper on this topic, watch the full presentation in the video below.

Happy New Year to you all! It’s great to be back, refreshed, and ready to take on the world again. I hope you all had a great holiday break, a wonderful Christmas, and are also ready to start 2025 with a bang!

To kick off 2025, I plan to start with a few year-long goals. Not resolutions – things flippantly said on January 1st and abandoned on January 3rd. These are goals that I want to move towards over the year. I’ll be revising them to make them more measurable in the future, but for now, I’m just covering the spaces I want to improve in 2025. I find that goals, whether personal or for businesses, help set direction, provide accountability, and measure progress. 

Goal process

There are myriad ways to come up with personal and business goals, but broad strokes they all boil down to a few key steps:

1. Determine potential goals
   1. Divergent thinking / brainstorming
   2. What do you want to be true at the end of the time period?
   3. Have you considered all the important areas – people, process, technology, strategy, etc?
2. Revise and clarify goals
   1. Convergent thinking – bringing ideas together
   2. Repackage goals into themes
   3. Think about what a measurable outcome for each goal would be
   4. Explore options, feasibility, and ways to succeed in your goals
3. Communicate goals
   1. Conversant thinking – sharing ideas for feedback
   2. Share them with stakeholders
   3. Gather stakeholder feedback on the value and feasibility of goals
   4. Set a cadence for updates on progress
   5. Track your progress and make course corrections (both to the steps your taking to accomplish goals, and the goals themselves,) as time progresses

If you’re curious what I ended up with, read on.

My goals for 2025

1. Build exercise into a habit, starting slow and aiming for a routine by the end of the year.
   1. This has been on the back burner for far too long. Even just a simple daily walk, or following a simple online workout routine will be better than what I normally do.
2. Find stability at work—whether through securing funding, joining a new company, or focusing on fractional work, something has to change.
   1. Being a founder has been hard and incredibly fun, but the low salary is getting tough. We have to change some things in 2025.
3. Ship the book. It’s almost wholly written; now it’s time to edit and self-publish.
   1. For those who missed this, I wrote a book a few years back about my experience leading the Indeed Incubator (new product innovation team.) It’s not great, but it’s not doing anyone any good sitting in Scrivener.
4. Dive deeper into learning and development, especially in AI and product innovation.
   1. I’m using AI more and more personally and professionally, but I would love to learn more about building AI systems.

I have other goals for 2025 on the personal, spiritual, and family sides of my life, but these were the ones that most affected work and seemed worth sharing.

There’s a concept called “making your own luck” that seems to be everywhere I turn lately. Whether you’re newly laid off, have been out of work for months (or years 😡!?!), or are trying to launch a new business like we are at Thurgood, manufacturing luck is an important skill.

Let’s start with a definition. Making your own luck is taking action to create new opportunities for yourself. The opposite of dumb luck, this “smart luck” is about actively looking for things you can do to increase your chances of success. 

Examples for job seekers:

• Attending (and actively participating in) networking events in your field/area.
• Connecting with previous coworkers, managers, and peers to see how they’re doing and share your job search.
• Actively reaching out to a targetted list of companies and leaders you’d like to work for.
• Starting a side hustle using (and developing) your skills.

For those starting a business:

• Networking with other businesses in your industry to understand the problems and solutions.
• Creating interest in your company through “earned media” (e.g. articles or podcasts about your company or industry that quote or reference you.)
• Taking connecting flights (courtesy of our investor Matt Blodgett) – turning each meeting into an introduction to other potential leads, partners, or investors.

These actions increase your luck surface area, e.g., a fictional measure of the amount of “luck” you generate through your actions. Others graph this with doing vs. telling, while that is a valuable way of thinking about the balance (not too much telling → boasting, and not too much doing → suffering in silence.) 

I think of luck surface area much like marketing – reach vs. frequency (too much reach without frequency → flash in the pan, too much frequency and not enough reach → spam.)

Get out there and make some luck today! What luck creation techniques work best for you?

First, credit to Mark Twain for the title. I have been thinking about the past a lot lately. We’re going through a significant upheaval in tech (specifically the hard shift towards AI, but other changes as well). Politically, we’re seeing a President come back after losing a re-election bid, something not seen in 100 years, which has multiple echoes. And tech hiring continues to be messy at best, with plenty of companies still in downsizing mode.

I have counseled several engineers (and others) that we have never seen a market in the US where there was enough software engineering labor relative to the demand for software creation. Things are certainly swinging closer to that point, and it will not be self-evident if/when we hit it, but what is clear is that the role of engineering teams has shifted. Teams that focus on demonstrating the value delivered will find it far more manageable than those with more abstract value. Simplified solutions maintained by fewer, lower-cost, or no engineers will win over more significant, more powerful, complex systems.

Mistake #62 – Rest, Recharge, and Reboot

A few weeks ago, I spent a long weekend in NYC with no kids, and it was a great chance to reset a few rhythms in my life. Getting some good sleep, multiple conversations with my wife without interruption, and a new scene were incredibly refreshing. I came back with a new perspective on my kids, work, and our life. This trip was (too) long overdue. I’ve often advocated for people to take vacations, even when you’re laid off, if your budget allows. 

This trip was long delayed due to kids getting sick, finding someone to watch our kids (even more difficult with school drop-off and pickup), and the hectic pace of startup life (we’re continually testing or launching something.) The rest from a four-day weekend was incredible, but what may be even more valuable is the reminder that if there are easy ways to simplify our lives (as we did for the kids’ aunt who visited), we should take advantage of those ourselves. Stepping out of our routine and looking at it from the outside, writing it all down, and sharing it with someone else helped us gain a new perspective on all the activities we’re doing each day.

Take a day. Look at what you’re doing. Make changes. And, above all, take care of yourself.

I’ve always had a hard time writing while working full-time. I put (nearly) all of my waking energy into the business I’m working for, with the remaining energy going to family. Additionally, sharing the internal details of what I’m working on feels like “trade secrets,” “insider information,” or other content that would be owned (or at least reviewed) by my employer. On top of that, let’s be honest: working in public is scary.

One of the best things about writing a weekly newsletter was how it forced me to slow down and think. Of course, I was busy filling out applications, interviewing, networking, and more, but I still made the time to think holistically about where I was and where I was going. The contracting life from October to May was a unique blend of full-time work with a lot of autonomy, making it a bit harder to find the time to write, but ultimately, I (mostly) kept up the pattern.

Joining Thurgood full-time in June, I put writing externally aside as I felt I no longer had a good perspective to help those laid off and needed to focus inward on building a new business. I recently reconnected with former coworkers and realized I hadn’t publicly announced my move from fractional CTO to cofounder & CTO. Much of my writing has turned inward as I’ve worked to help crystalize and document our strategy and experiments. Now that I’m a few months in, I have some time and am ready to surface some of the work I’ve been doing. I doubt this will return to a weekly newsletter, but it should be more regular again. 

Thinking through writing

I’ve read many times that “leaders are readers.” While I agree you need diverse opinions and thoughts to shape your leadership and direction, reading is also helpful as it shapes your writing. Writing is key in capturing, crystalizing, and synthesizing your thoughts. Writing slows you down to evaluate and test what you’re saying. Let’s break this down:

• Capturing
  • Thoughts can be so fleeting that pausing to write slows you down, and if you give yourself the liberty not to edit as you write, you’ll find several ideas floating around in your brain. Putting them down somewhere frees you to let them go and see if they resurface, refine, or die on the vine.
• Crystalizing
  • Because writing is slower, you can tighten your thinking, test your opinions, and see the holes in your thought processes. First-draft ideas are still just first drafts but replayed in a second draft, email, or speak to others, are much more refined and ready for debate and consumption.
• Synthesizing
  • Putting down your ideas allows you to see the patterns between them or to combine two ideas into something that’s even more powerful than the original parts.

Mistake #61 – Thinking inside the box

For years now, I’ve been part of a much larger organization, and I could (generally) trust others to run the finances, sales, and other key parts of the organization. As a founder, those are (partly) my responsibility, too. Once we secured an initial pre-seed round of funding for Thurgood, it was easy for me to go completely heads down on getting the technology in place to support the rest of the company and neglect the broader details.

As an engineering team, we did a good job keeping things lean and shipping early. Unfortunately, I failed to create a concrete enough picture of the engineering budget to forecast our spending through contingency plans. As we launched to public users in June, we quickly began to encounter a variety of significant headwinds. Our user base did not organically grow, our application was not sticky, and our user acquisition channel was slow. Of course, additional experiments would be needed to get these rolling, but the timeline for securing additional funding was slipping beyond our runway. 

While running the Incubator at Indeed, we had a rule of thumb for engineering teams not to build anything that wouldn’t move the needle on their current runway or one runway beyond. That’s a luxury only afforded to corporate startups. As a venture-backed startup, our vision is even more narrowly focused, and unfortunately, the length of the runway is the biggest variable we have control over. We had to lengthen the runway as much as possible to buy time for our experiments on growth to be run and had to close out our contracts and lay off two very qualified full-time engineers. 

So we’re back to a group of founders, each pulling as hard as we can on this rock to get it rolling.

For over a year, I’ve been writing (mostly) weekly updates from the post-layoff seat to contracting and soon to a new full-time role. I’ve seen my writing go from optimistic, to broken, and back again. It’s been a wild year, and unfortunately, 2024 is shaping up to be challenging too.

Photo by Clay Banks on Unsplash

This week another 1,000 Indeedian’s joined a club they didn’t choose to be a part of, ex-Indeedians. Like many of my former coworkers (Indeedians and ex-Indeedians alike) I’m putting my time, my network, and my camaraderie out there for you all. If there’s anything I can do to help, please reach out: on OnlyNewMistakes, via Substack, LinkedIn, email, SMS, or whatever works best for you. 

Rather than telling more of my story, I’m going to grab a few of the better bits of support from previous posts and re-share them for those who’ve been laid off recently. Rather than make this just a straight link dump from previous posts, I’ll provide a few updates/insights that have come with hindsight since those posts.

1. Immediate reactions & next steps – week #1 
   1. The first thing I see is the optimism in this post. I still thought we were in the same old tech market. We are not. It has changed. Be prepared to search for a while and to have fewer options than you would have previously.
   2. Networking is crucial. You have the time. Reach out to the unemployed and connect. Reach out to people who are working too. Get recommendations, referrals, tips, resume reviews, and anything else they offer.
2. The funnel of despair – week #5
   1. Lots of folks (including me) shared their stats along the way, but the ratios of jobs applied, interviews, and offers are VERY real right now. Creating a system to track your applications, and following up is important.
   2. That mountaintop-to-valley-to-mountaintop ride of the “Morale over time” chart is very accurate. Be prepared for the rough days and have a plan. Get out, get some sun, and see friends.
3. Using the tools well – week #13 & week 32
   1. From making sure you have an active role in your LinkedIn to trying other tools (Otta, BuiltIn, etc.) make sure you use your tools well.
   2. AI tools for resume review and cover letter writing are quite good and can save a lot of time when you are firing off a lot of applications.
4. Focusing on what you do well – week #22
   1. Writing down things you do (and did) well and reading those frequently was a lift when my spirits got low. I emailed my list to myself and snoozed it each morning to the next day or two so that it was a regular message reminding me to focus on my strengths.
5. LinkedIn supertips – week #21
   1. Using the Social Selling Index to understand your LinkedIn profile’s strengths and weaknesses and how to improve your profile’s ranking.

There are weeks of other content here or on any of the other platforms I share this content on: OnlyNewMistakes, via Substack, LinkedIn. I hope it helps. 

I’m several weeks behind in writing and wanted to jump back into it, so today’s newsletter is a collection of four short notes rather than the previous format (not that I was ever super strict about the structure of this newsletter.) I hope you enjoy it.

Week 52 was an emotional one for me. Seeing the calendar roll past March 22nd (the one-year anniversary of Indeed’s layoff) meant recognizing that it’s been a year since I felt stable in my income, work, and career. There were a lot of emotions wrapped up in that week, and I wasn’t ready to write about them. 

Now that I have the distance of a few more weeks, it’s easier to see that I had put a lot of my self-worth into what role I was playing and the company I was working for. Much like COVID taught me about the value of connection and networking (a story I can share or re-share in the future,) the layoff gave me more clarity about what matters in my work. Ultimately, it’s about working with people I like, on a mission I value, and making tangible progress. Those things were less and less the case for the last year or more of my time at Indeed.

While I’m not 100% back to where I was at Indeed, my new roles have given me a chance to work with some great leaders on missions I am excited about and to see businesses formed from nothing. I’m learning (a lot). I’m enjoying the work (a lot). And I love the teams I’m working with. There is not much more I could ask for.

Living Life at 100 MPH

The last two to three months have been extra chaotic. With the fractional CTO work kicking into the highest gear in February and our kids’ Spring sports kicking off (often before Winter sports had fully wrapped up), we went from standard four-kid-family busy to ultra-busy. But this isn’t me #hustle bragging. This is me being honest.

I’m exhausted and behind. I wanted to launch HeatCheck and acquire users. I wanted a second startup to move away from Excel sheets and to a custom product to run its business. I expected to be running a cohort of executive coaching sessions. And I wanted Thurgood ready to launch by this month. None of those things have happened on my schedule. They’re all significantly moving along and approaching our goals, but each still has its hurdles.

On top of that, one of the first engineers I hired at Indeed passed away suddenly at the beginning of March. It was a real shock and a reminder that it was time for simplification. While the contracting life has paid the bills, and I’ve learned a lot about it, I don’t think it’s my long-term play. I’m starting to trim back the contracts and not actively seek others. I want to see each of these things across their respective finish lines. I want to reconnect with friends. I need to get back to working out regularly. I have to have some margin in my life. So that’s my next challenge, trimming and replacing with life-giving activities (friends and exercise in particular.)

Launching Early

Two startups I’m working with will launch very soon but look different. The bootstrapped HeatCheck is approaching a marketing launch that will hopefully begin to bring in users and revenue. Thurgood is about to kick off its pilot and see the first users on the platform as well (although revenue will likely lag a bit behind that pilot).

Reid Hoffman is often quoted as saying, “If you’re not embarrassed by your MVP, you’ve launched too late.” In the case of HeatCheck, I’ve been partnered with that team for over a year now, but the nature of a bootstrapped product is that every penny spent comes from someone’s pocket, so you’re constantly wondering when the right time to get real users into the product would be. Thurgood is VC-backed, which has allowed the team to run faster (hiring a few contractors to start and soon a founding engineer) and the marketing budget to do a pilot kickoff next month.

Both will still have a handful of rough edges and feature “cul-de-sacs” at launch. We’ll keep working those out based on user data and feedback. That will be a fun and exciting new phase for both projects. I can’t wait to get there!

Recovery Weekend

No baseball last weekend meant a recovery weekend. Usually, it’s two baseball games on Saturday and one or more on Sunday (play until you lose). We still had a handful of activities in the week leading up to the weekend off, and of course, there were still a few things to do on the weekend. Still, on the whole, it was things we wanted to do that allowed us to catch up on housework, see Dune Part Two in the theater, and plan a few upcoming family activities. Now that the weekend is nearly done, I’m looking to the next few weeks and realizing that it will be a full schedule between now and the end of June. 

We’re somewhat at the mercy of the kids’ schedules, but I think I need to schedule a “do-nothing” weekend every three months or so. Maybe the kids are still coming and going, but we arrange for carpools or plan for them to miss their activities that weekend as well so that, as a family, we can collectively catch our breath. I know Summer will naturally give us a few of these as there are fewer outdoor sports thanks to the Texas heat, and in fact, after some of the kids are dropped off for their various camps, we’ll get some quality 1:1 time with the remaining kids. So, even as busy as the summer looks, I’m looking forward to the recharge that comes with it as well.

This post is being drafted from beautiful Glenwood Springs, Colorado. Since we’re on vacation it’s much shorter than usual, but I was going to post this to LinkedIn last night and I realized that it was more or less a micro-update. I hope everyone has a great Spring Break! 

Today I did a hard (for me) thing. I don’t have great balance on the whole. I have wide feet with high arches. 

This makes skiiing a tough choice for me. Rental gear rarely fits well, and by well I mean it rarely produces anything less than excruciating pain. I’m clumsy on ice and fall a lot. It just doesn’t come naturally to me (and snowboarding isn’t much better.) 

But as the bruises and aches can attest I did ski today. I learned, I improved, and I survived. I even had fun. Sure, I’d love to keep up with my family but that’s a lot more lessons and practice away. For now, I’m happy to have taken the challenge and learned to (somewhat) enjoy something my family enjoys very much. 

So let me encourage you with two things today: 

1. Take the challenge! Try something new and push yourself. Let me know what you do! 

2. For those with a family or a partner, find something they enjoy and make it something you can enjoy (even just a little) too.

​​

Fun