Since Thanksgiving last week I've been doing a top-secret experiment.

The early results are promising so I've decided to share them with you!

Here's the experiment:

• I created a list of questions to ask myself each morning and each night. The questions I chose help me plan my day, work on my goals, and reflect each evening.
• I stored them in ChatGPT's memory as a personalization instruction, with two trigger phrases 1) "Let's do the morning prompt" and 2) "Let's do the evening prompt"
• Each morning or evening I use my iPhone's text-to-speech to say "Let's do the morning/evening prompt." into the ChatGPT app
• It will reply "ok!" and list out the questions
• Then I start dictating into my phone, using the built-in microphone, which goes into ChatGPT's text window.
• I answer each of questions as best I can for that day, not worrying about how well the dictation comes out.
• ChatGPT responds to my dictation, helping refine my answers and set me up to have a good day

Here are screenshots of the questions (click to make it larger):

The results so far have been fascinating. The AI has helped me: Set up a 3-hr "focus block" each morning to work on high value tasks, letting me get more done this week than any week last month Got me to go to CrossFit when I was tired this week Offered me a deep breathing protocol to improve sleep (I will sometimes wake up at 3 am and can't fall back asleep) Listened to me describe all my tasks and my energy level one day, and generated an entire schedule customized exactly for what I want to do Offered me powerful mental reframes for handling daily situations (e.g. win the morning, win the day) Helped me schedule and actually study for the Certified Development Lifecycle and Deployment Architect exam, which I've been procrastinating on It's like I hired a coach and advisor with a PhD in every field to known to mankind. The AI never gets tired, angry, or bored. It remembers every detail. It's always trying to help me. It's like a friend with unlimited energy and unlimited intelligence, and committed completely to working with me to have a good day. An excellent friend to have! You can also "flavor" its responses. ChatGPT has a small memory where you can store instructions (called Personalizations) about how to interact with you. I think most people use that for things like "I like reading books." but I've turned it up a notch. Check out this personalization (click to make it larger):

Now, when I'm planning my day, the AI will ensure I'm lined up with these principles for success. It's crazy! And the funny part is I didn't even write that personalization myself.  I asked another AI tool (Claude) to write a personalization instruction that I could upload into ChatGPT to help me do deep work (I did give it two or three principles as a start). I took things even farther and asked Claude to generate 1) an identity statement and 2) summarize the philosophy and principles of some of the most brilliant people I know of. People like:  Ray Dalio Charlie Munger Elon Musk Ben Franklin Eckart Tolle Next, I took the identity statement and principles, and asked Claude to turn them into a personalization instruction for ChatGPT to use when talking to me. Here is the one Claude generated for the Roman Emperor Marcus Aurelius, which I uploaded to ChatGPT:

I made the Emperor of Rome my mentor! Ha ha. After doing this experiment for a day or two, I noticed the AI never criticized me and generally agreed with me on almost everything. Interesting... So after my next evening prompt I asked: "Is there anything you would criticize or give me feedback on based on what we've talked about? What painful truths might I not want to hear but are essential for me to grow?" AI doesn't play around. It immediately zeroed in on some of my biggest weaknesses: You're stuck in over reflection at times You protect your comfort zones in subtle ways You may be underestimating the power of single focus And listed out reasons why it thought this way.  I felt like it read my soul. It was completely correct. How could I argue?  My super intelligent friend with unlimited energy just highlighted my sticking points (and gave me solutions). Like Ben Franklin said: "The sting of a reproach is the truth of it." If you try this yourself it's important to do it after you answer lots of questions in a longer prompt. Answering a bunch of questions before asking for feedback lets you build up a fairly big context window in the current chat. ChatGPT's memory gets wiped between chats, so the longer you talk to in a single session, the more "data points" it will have on you to assess. Other questions I found interesting to ask:  What is the most limiting belief I appear to hold? What's one thing you think I may not know about myself? Based on everything we've discussed, please estimate my IQ in a 10-point range If I could receive one piece of advice every morning to help me grow, what would it be?

And that's as far as I've gotten in a week!

It's been a fun experience and I think I'll extend the experiment out 90-days to see what it's like.

Working on this got me into a real flow state so I was happy to push it pretty far.

I wouldn't say everyone needs to go all-out like I did, but I would definitely encourage you to try a modified version.

The insights are both practical and profound.

Hope this helps!

Best,

Nick

P.S. Here is a quick start guide to doing this yourself:

15-Minute Quick Start Guide

Setup Phase (10 mins):

1. Open ChatGPT
2. Copy paste the following. The formatting may not translate. That’s okay.

Update your memory about me to include exactly the following. Do not change the text in any way.

Please create two trigger phrases for me to start my morning and evening prompt

• "Let's do the morning prompt"
• "Let's do the evening prompt”

My morning questions are as follows:

• What's my most important task today?
• How will I make time for it?
• What potential obstacles should I plan for?

My evening prompt questions are as follows:

• What went well today?
• What could I improve?
• What's my first priority tomorrow?

When I ask for the morning prompt please only list the morning prompt questions.

When I ask for the evening prompt please only list the evening prompt questions.

I will answer each question individually. Please evaluate my responses, refine my thinking, and help me have a good day.

Daily Practice (5 mins each session):

• Morning: Use text-to-speech to trigger morning prompt
• Evening: Use text-to-speech to trigger evening prompt
• Speak your answers naturally, don't worry about perfect dictation

Pro Tips:

• Start simple, add complexity later
• Use longer sessions to build context

Musicians practice scales to captivate audiences.

Basketball players sharpen their shots with drills.

Chess players master tactics to outthink their opponents.

Actors rehearse to deliver unforgettable performances.

But how do Salesforce developers level up?

The answer might surprise you.

In The Pragmatic Programmer by Andy Hunt and Dave Thomas, the authors take the concept of a "kata" from karate and apply it to writing code.

In martial arts like karate or kendo, katas are sequences of movements practiced repeatedly to perfect form, discipline, and technique.

The idea is not just to memorize the sequence but to internalize the principles behind the movements.

This concept of deliberate, repetitive practice aligns well with the iterative and skill-based nature of coding.

And so, the code kata was born.

Code Kata: A few lines of code that are practiced repeatedly to perfect form, discipline, and technique.

Below is an example I've been working on this week:

This Apex class is a mere 18 lines of code, and only 14 if you take away the line breaks.

Now, look how many concepts are present:

1. Class Declaration
2. Method Declaration
3. Collections (Set, List, Map)
4. Control Structures (For loop, Ternary Operator)
5. String handling
6. Counting Logic
7. Duplicate Detection
8. Debugging and logging
9. Type Initialization

Nine distinct coding concepts by my count. That's a lot!

By simply rewriting this Apex code once a day, you gain repeated exposure to the building blocks of writing Apex.

Why Code Katas Work

• Muscle Memory for the Mind: Just as physical katas build muscle memory, code katas train the brain to recognize patterns, solutions, and pitfalls in programming.
   
• Safe Environment for Experimentation: By solving the same problem repeatedly, developers can experiment with different approaches, frameworks, or optimizations without fear of failure.
  
  
• Focus on Fundamentals: Code katas emphasize core principles like loops, recursion, testing, and algorithm design, which are critical for all programmers.

Now that you understand the power of code katas, it’s time to take action. Pick a code kata to practice—whether it’s the example shared here (available to copy paste from this Github repo) or another small, focused exercise.

Commit to rewriting it daily for a week and notice how your skills sharpen with each iteration. Repetition is the path to mastery.

And if you don’t write code but work with Salesforce flows, the principle still applies: find a repetitive practice to internalize flow-building fundamentals.

For instance, take a common automation scenario—like updating related records or sending notifications—and recreate it from scratch daily.

Over time, you’ll develop the intuition and skill to tackle even the most complex flows effortlessly.

Practice doesn’t just make perfect—it makes you better every day. So, start your kata, and let’s see where deliberate practice takes you.

Best,

Nick

Alex hadn't meant to discover that reality was mutable. It happened during a late-night coding session as he prepared for an upcoming interview, when the lines between Apex triggers and consciousness began to blur.

He'd been visualizing system architectures - a technique he'd developed for tracking complex dependencies. But that night, exhausted and deep in flow state, he noticed something odd: the visualizations were starting to write themselves.

It began with simple patterns. Customer objects linking to opportunities linking to tasks - standard Salesforce relationships. But as he held these structures in his mind, he started seeing them everywhere. The way his neighbor's cat moved through space suddenly looked like an elegantly optimized query. The flow of traffic outside his window resembled a perfectly tuned workflow.

At first, he attributed it to too much caffeine and too little sleep. But the patterns persisted. More importantly, they responded.

When he visualized success paths through his code with perfect clarity, bugs would resolve themselves in unexpected ways. When he held the image of system architecture in his mind while falling asleep, he'd wake to find solutions emerging fully formed.

Then came the stranger observations. During technical discussions, if he maintained a clear visualization of positive outcomes, conversations would flow toward those results almost magnetically. It wasn't forcing or manipulation - just reality optimizing itself along paths he'd compiled in his mind.

"Intent," he muttered one night, though he couldn't remember where he'd heard the term. The way a compiler converts high-level code into machine language, his visualizations were converting pure intent into executable reality. He began experimenting. Before client meetings, he'd compile complete success scenarios - not just hoping or wishing, but rendering them in his mind with the same precision he used for technical architecture. The results were statistically improbable yet consistent. But the real breakthrough came when he stopped trying to force specific outcomes and instead focused on compiling states of clarity. Like refactoring cluttered code into elegant patterns, he found he could refactor his own perception. The interview process that had once seemed daunting became a series of patterns to optimize. Each question an opportunity to demonstrate not just knowledge, but understanding. Each interaction a chance to align intent with execution. Some nights, working late, he'd catch his reflection in the monitor and see not just himself, but threads of possibility - like stack traces in reverse, leading not from errors but toward optimal outcomes. He kept no notes about these discoveries. They weren't the kind of thing you could document in Asana or capture in acceptance criteria. But they compiled themselves into his understanding, like well-written code becoming part of a system's architecture. Reality, he discovered, wasn't fixed. It was more like a continuous integration pipeline - constantly building, testing, and deploying based on the patterns you compiled into it. The key was precision. Vague hopes produced unpredictable results. But clearly compiled intent, rendered with the same attention to detail he brought to his technical work, seemed to optimize reality's execution paths. One evening, still preparing for the crucial technical interview, Alex found himself thinking about quantum compilation - how observation collapses probability waves into specific states. He smiled, remembering that observer and system couldn't be separated. The compiler was itself being compiled. He closed his laptop, stood, stretched. Outside his window, the city flickered like a billion runtime operations executing in parallel. Each light a query, each moving car a thread, each human interaction a complex data exchange. All of it mutable. All of it responsive to properly compiled intent. Ready to execute.

Hope you enjoyed this story! 

Best,

Nick

Quick tip this week!

In the most recent Winter '25 release, Salesforce enabled Conditional Field Formatting.

This is a fun new feature that I think we'll see used more often over time.

In the Lightning App Builder, when working with dynamic forms, you'll notice there is now the ability to apply Conditional Formatting to fields.

Basically, you set up an if/then rule, and when the criteria for your rule is met, an icon appears on the record page next to the field of your choice.

In this example, we show a green happy face when the stage of an opportunity = Closed Won.

Which looks like this on the page:

There are a ton of icons to choose from (more than in the screenshot) so the sky is the limit in terms of how we can use this functionality.

Delighting end users is always our goal, and I think this is one of those low effort / high reward items that let will let us do just that.

Hope this helps!

Best,

Nick

Quick tip this week.

You can pass filter values into the URL of a report in order to dynamically change what the report shows.

Pretend you have an Opportunity report with one filter on the AccountID field.

So all the Opportunities on the report are related to the Account you filter on.

Normally your report URL would look something like:

yoursalesforce.com/lightning/r/Report/00Oak000009CJUjEAO/view

By adding a “?fv0=” at the end of the URL, you can pass any AccountId into the report for it to filter on dynamically when it loads. 

Like this:

yoursalesforce.com/lightning/r/Report/00Oak000009CJUjEAO/view?fv0=001ak000009juju

When you click this modified URL, the report will load specifically for the AccountID you added to the URL.

You can even pass in values for several filters by using fv0, fv1, fv2, etc. as long as you separate them using the “&” sign.

Like this:

yoursalesforce.com/lightning/r/Report/00Oak000009CJUjEAO/view?fv0=001ak000009juju&fv1=accountName

You can use dynamic filtering in hyperlink formulas for list views, flow emails, formula fields on record pages, and more.

It’s a useful thing to know, and I had to relearn how it worked this week, so I thought I’d share.

Hope it helps!

Best,

Nick