Tim Dickey | Agile Practitioner, Product Development & AI Fluency

Most days, I am treading water.

The AI tooling landscape changes daily, and I coach software development teams for a living. That means the background noise is a professional risk, not just a distraction.

Here’s how I stay on the leading edge without falling off:

1. I participate, not just observe.
   When dev conversations happen, I show up with a business lens. My job isn’t to write the code; it’s to help the people who do.
2. I’m transparent about what I am.
   I call myself “the actor playing a software developer on LinkedIn.” It’s not self-deprecation. It’s clarity; I respect my colleagues enough to be honest about where my expertise ends and theirs begins.
3. I treat learning like a commitment, not a hobby.
   Time in uniform taught me that. Special operations don’t tolerate stagnation. Neither does this industry. Bookending my day with learning isn’t motivation; it’s maintenance.

---

To some people, I probably look obsessive about generative AI and product development. I prefer to think of it as being selective and hyper‑focused.

There’s a fine line between being seriously committed to a craft and getting fixated to the point of burnout. Here’s how I stay on the rails and keep AI as a powerful ally instead of a constant distraction.

1. Timebox everything

I don’t spend hours a day doom‑scrolling AI news. Instead, I give myself 1–2 hours each day to:

• Scan headlines and releases
• Pull a few relevant articles or papers
• Have a focused conversation about how AI applies to the work I actually do

When the time is up, I stop. The constraint forces me to prioritize signal over noise.

2. Find passionate collaborators

You might be surprised what you can learn from people outside your primary discipline.

At Improving Dallas, we have a weekly AI Roundtable where we demo experiments, share wins and failures, and ask questions in front of peers. I don’t fully understand everything my colleagues are doing, and that’s the point. Having a trusted community lets me:

• Borrow perspectives from different domains
• Get unstuck faster
• Stay accountable to actually trying things, not just reading about them

3. Run small experiments

“Small, safe‑to‑fail experiments” are just as valuable for AI as they are for product development.

Rather than chasing every new model or tool, I define concrete experiments:

• One specific workflow to augment
• A clear success metric (faster, better, clearer, cheaper, safer)
• A limited time window to test, reflect, and decide whether to keep or discard

Applied learning beats abstract theory. The more I run experiments, the more I learn what actually works in my context.

4. Set learning goals—and keep promises to yourself

This connects back to my earlier post on learning goals. I keep a short list of specific AI‑related skills I want to build (for example, “prototype one AI‑assisted research workflow this month”), and I treat those commitments as seriously as I treat commitments to clients.

Closing the loop matters:

• Set a goal
• Make a plan
• Do the work
• Reflect on what changed

That cycle reinforces the psychological benefits of continuous learning: confidence, momentum, and a sense of agency in a disruptive time.

---

Practical Examples to Start With

1. Timebox everything
I give myself 1–2 hours per day. No more. Here’s how I structure it:

• Morning scan (15–20 min): I use a single aggregator — right now, that’s a curated Inoreader collection — to skim headlines without opening rabbit holes
• Focused pull (30–45 min): I use NotebookLM to ingest 2–3 relevant articles or papers and surface connections to current work
• Applied conversation (15–30 min): I use a persistent Claude or ChatGPT thread as a “thinking partner”. I ask how something applies to my specific context, not the world in general

When the window closes, I stop. The constraint is the practice.

2. Find passionate collaborators
This point is worth mentioning again. Improving Dallas has a weekly AI Roundtable that includes live demos, shared experiments, and honest failures. It’s a structured community of practice, not a passive Slack channel.

If you don’t have one, you can build the equivalent:

• Start a recurring 45-minute working session with 2–3 peers; no agenda, just experiments
• Use a shared Notion or Obsidian space to log what you tried, what worked, and what you killed
• Follow practitioners in adjacent fields (designers, researchers, ops leads); cross-domain signals consistently outperform staying inside your own lane

3. Run small experiments
I don’t chase every new model. Instead, I define a structured micro-experiment:

• One specific workflow to augment (e.g., summarizing retro notes, drafting coaching frameworks)
• One clear success metric — faster, clearer, cheaper, or safer than before
• A 2-week time box to test, reflect, and decide: keep, modify, or discard

I log each experiment in a simple markdown file: hypothesis, tool used, result, and decision. Over time, that log becomes a personal evidence base, not just a collection of impressions.

4. Set learning goals you’d be embarrassed to abandon
I keep a short list of named AI skills I’m building — something like: “Prototype one AI-assisted research workflow this month.”

The mechanics that make it stick:

• Write the goal in public (a post, a note to a colleague, a calendar event with a title that makes it real)
• Use a weekly 15-minute review; I run mine on Friday to check: Did I do the work?
• Close the loop: set → plan → do → reflect. Each completed cycle builds the confidence and agency that sustained learning actually requires

None of this is effortless. But intentional engagement with AI (structured, bounded, and grounded in real work) is what separates professionals who adapt from those who just watch.