Accidentally in Code

Tag: ai

  • Ground Decisions

    Credit: Unsplash / Bao Menglong

    Often, when we talk about automation, we talk about planes. We say they fly themselves, which, sure – and technically the largest planes can even land themselves, if the airport has the equipment (it’s mainly used for bad weather). However most takeoffs and landings remain manual. My pilot friend also points out that many decisions (fuel is a big one) get made on the ground.

    The other thing my pilot friend pointed out is that pilots are in for any decisions they make. They are not just flying the plane; they are in the plane. Contrast this with the surgeon, who performs the surgery, and walks out of the operating room afterwards, regardless of the patient outcome.

    In this analogy, VCs are surgeons. Developers are pilots. And the hype-men… travel agents selling plane tickets, mainly.

    The problem with much of the hype around AI is that it’s not meeting the people who actually do the work where they are at. We’re not being clear about what’s takeoff, what’s landing, and what the ground decisions are.

    The cracks are starting to show. I’ve seen various bits and pieces, but nothing captures it better than Luca Rossi’s summary of CircleCI’s State of Software Delivery 2026 data: elite teams nearly doubled throughput YoY, median barely moved. 81% use AI, so the differentiator is not the tooling; it’s infrastructure. Teams with CI pipelines under 15 minutes in 2023 are 5x more likely to be top performers today.

    I am fascinated by the software factory model — Dan Shapiro’s Level 4, where programmers are less programmers and more managers of AI agents and tasks — and the challenges of retro-fitting it onto an existing engineering organisation. It’s been a lightbulb moment for me – I feel like I finally get why the layoffs happened first, and why the productivity gains have failed to materialise.

    I think the answer lies in the ground decisions, or the open questions from an organisational perspective.

    How does the developer tooling calculus change?

    With teams of developers, we used to know the things that were useful for a team of 20 but not 2, and the things that made sense at 200 but not 20. That calculus seems to have changed fundamentally for two reasons. First, headcount is not (or shouldn’t be) a good proxy for productivity. Second, tooling itself is easier to build.

    Concrete example: every organisation eventually evolves to have a design system. Few start with one. But if you were building a web app today, would it make sense to start with a design system from day one? The answer might be yes.

    How do budgets need to evolve?

    Historically, dev team budgets are mainly salaries and some tooling. But StrongDM, who built one of the first real software factories, use $1K/engineer/day in token spend as their benchmark for whether you’re doing it seriously.

    Firstly, how do you factor that kind of ongoing cost in? Secondly, is managing more complex budgets and ROI going to become a bigger expectation of engineering managers? And what does it take to get good at that?

    What does skill definition look like now?

    It’s clear that a different skill composition will be more valuable in this model. My predictions are that judgement – understanding what to build and why, what is good to ship (aka takeoff and landing) – and feedback – providing clear direction and iterating well – will be bigger differentiators, and earlier in someone’s career than before. What people call prompt engineering is just problem definition and refinement – skills that have always mattered, now moved to the front.

    What does this mean for hiring and onboarding?

    Skill definition leads neatly into these two problems: what do you evaluate, and how do you onboard?

    In hiring, most of the conversation has been about the ways AI generates noise from candidates and heartless automated rejections from companies. Which is a real problem, but not the most interesting one. Beyond that: once there’s a human in the process, are the skills being evaluated the best predictors of success in this new model?

    Similarly with onboarding: having hired many people over the years, I used to have dialled in what good looked like at key intervals — 30, 60, 90 days. But in this model, what does a good trajectory look like when AI can both accelerate and abstract from understanding the systems?

    I have my own concerns about AI, I’m not going to pretend otherwise. But I also think the adoption curve is inevitable, and I’d rather help engineering leaders navigate the reality in front of them than relitigate whether we should be here at all.

    Coming back to the pilot analogy:

    • Takeoff – what we build and why
    • Landing – shipping and the decisions around it – what’s actually good enough to meet the definition of done?
    • Ground decisions – the open questions above.

    What am I missing? What’s your definition of take off and landing, and what are the ground decisions that are changing in real time?

  • The Spreadsheet: Using AI to Understand How I Spend My Time

    Credit: Google DeepMind / Unsplash

    For years, I’ve known intellectually that my best days aren’t my busiest days. That sustainable productivity requires balancing multiple priorities. That beating myself up about not accomplishing “enough” is counterproductive.

    But knowing something intellectually and remembering it when you need to are very different things.

    So a couple of months ago I started trying something different to help me better understand how I spend my time, what works, and what doesn’t. I wanted to

    • Capture what I was actually doing across multiple areas
    • Track not just activities but impact and how I felt
    • Mine the data for insights about what was actually working

    A long time ago, I tried tracking my time in 15 minute increments, which was helpful, but a lot of work. At DuckDuckGo I used to keep a list of what I’d done for the week and then outside of work I mainly used Trello. But then I only got the dopamine hit of the checkbox for finishing things – not all the effort that went into them.

    The Solution: A Spreadsheet

    I set up a spreadsheet.

    • One tab per week – easy to review without drowning in data
    • Columns for categories: Adulting, Human Being, Development, Promotion, Revenue, Twill, DDG (while relevant)
    • Two assessment columns: Daily impact rating and overall vibe, color coded (green/yellow/red).
    • Color coding: I also highlighted high-impact activities in green.
    • Keep it open: I left it in a browser tab and jotted things down as they happened

    The key was making it easy. I wasn’t timing anything precisely or categorizing every minute. Just capturing: “What did I do today? Did it matter? How did I feel?”

    After a month, I had enough data to start asking bigger questions about what was working overall and what was not.

    The AI Analysis

    I fed my spreadsheet to Claude and asked it to analyze my patterns over the past month. What came back was a thorough, personalized report that I would never produce myself. First because of the time, and secondly because I’m inclined to be overly harsh on myself for what I accomplish – and what I don’t.

    This is one of the core ways I use AI: to give me faster and more detailed feedback loops.

    The feedback tends to be very positive, which is why real data matters—it keeps the AI grounded. It also tends to be generic, which is why identifying a process that works for me first, then using AI to give feedback and iterate, works better than asking for generic suggestions.

    But when you combine genuine data with AI analysis, something useful happens. Here are some of the insights I got from Claude’s analysis.

    The High-Impact Day Formula

    Claude identified patterns in my color-coded days. It turns out my best days share these characteristics:

    • 2-3 categories active (not all six—trying to do everything leads to feeling scattered)
    • 1 significant accomplishment I can point to
    • Mix of strategic + tactical work (both vision and execution)
    • Human Being time included (exercise, relationship time)
    • Clear assessment at end of day of what I achieved

    This was validating. My instinct to prioritize exercise and quality time with my partner isn’t “nice to have”, it’s foundational.

    Wins + Opportunities

    Claude validated me as someone who:

    ✅ Values wellbeing and relationships
    ✅ Manages multiple work streams effectively
    ✅ Navigates major transitions with awareness
    ✅ Balances doing with being

    But also (accurately) called out that I would benefit from:

    ⚠️ More deliberate planning (light vs. focus days)
    ⚠️ Increased development investment
    ⚠️ Recovery time after intense periods
    ⚠️ Clearer daily priorities to reduce “busy but unclear” feeling

    The Core Insight

    I’m pretty sure I already knew this, I just forget when focused on all the things I haven’t done (yet). I hope that building a structured way to give myself a regular reminder is not just good for my productivity – but also for my mental wellbeing.

    My best days aren’t when I work most, but when I have clarity about what matters, make visible progress, and maintain my human being practices.

    This felt particularly important to pay attention to as I navigate a transition from a structured and meeting heavy work environment to a more portfolio setup, but honestly most of the same insights probably would have been true 6 or 12 months ago.

    Rinse and Repeat

    The nice thing about this system, was I could use the insights and feed the data back in a week later. Claude was encouraging about my progress, which was nice – I tend to start the week a little overwhelmed by everything I want to accomplish and it helped me ground myself in what I am capable of.

    After taking the feedback, I managed to have a week as a person that:

    • Works hard without sacrificing relationships
    • Rests intentionally without guilt
    • Acknowledges challenges without catastrophizing
    • Celebrates progress without waiting for perfection
    • Maintains systems without being rigid

    Obviously I still had things I can do better (specifically: spending more time on development). But for one week later – a busy one! – that was pretty great progress. This solidified the Monday morning ritual – run the spreadsheet through and ask “how did I do?” and “what can I learn?”

    Failing

    I live for the validation, but a few weeks later I took a long weekend and missed the Monday ritual. By Friday, I was toast. I ran the spreadsheet through Claude, and it brought the Real TalkTM. Pointing out that I had:

    • Sprinted through the week before the long weekend.
    • Not actually rested – had a hectic (but fun!) weekend away.
    • Came back and tried to jump back to full capacity whilst noting myself “discombobulated and tired”.
    • Missed my physical anchors (no yoga or spin).

    Claude ordered me to stop and rest, and I obeyed. It was a good call – I managed a nice afternoon reading a novel before spending the entire weekend sick.

    I know this is a pattern I have. That I push myself to “earn” a break by getting everything I would have done in that time around it. It’s destructive. But the good thing is that the failing enforced the value of this ritual more than any succeeding would have done.

    Try It Yourself

    If you’re curious about your own patterns:

    1. Pick 5-7 categories that matter to your life (work projects, learning, relationships, health, etc.)
    2. Create a simple spreadsheet with these as columns.
    3. Add an overall impact and overall vibe columns to capture how you feel about how the day went.
    4. For 2-4 weeks, jot down what you do each day.
    5. Use color for an easy visual emphasis.
    6. Feed it to an AI and ask for patterns – note that in your prompting you’ll likely need to be explicit about the color coding and what it means.

    Maybe you’ll learn something new, or maybe – like me – you’ll be reminded of the things you know but forget when you’re stressed. Either way, I hope it’s useful.

    If you try this, I’d love to hear what you discover. What patterns emerged? What surprised you?

  • What Raccoon Are You Quiz (and some thoughts on vibe coding)

    Raccoons are very much part of my brand, so many friends (and my boss) sent me the latest adventures of the drunken raccoon in the liquor store. The past couple of years I’ve also been framing my talks about tech as we used to be instagram raccoons – and now we all live in Toronto.

    So I made a personality quiz for tech workers – what raccoon are you right now?

    Take it. Find out if you’re the drunk raccoon passed out in a liquor store, the MPR raccoon stuck 23 floors up, or the unkillable Toronto raccoon that defeated every “raccoon-proof” garbage bin the city designed.

    Creation

    I’m a software engineer by training, but JS is not my friend, so I took this as an opportunity to try some vibe coding using Claude. I set up a repo on GitHub to use GitHub pages, committed changes as we went, reviewed the code, and made my own edits after.

    It was fun! Without AI it would have taken me much longer to turn this concept into something, and probably not something I would have prioritized. I used AI to give me options, generate the things I didn’t care deeply about, accelerate tedious work, and focused my attention on making the quiz fun and interesting. I started with the list of raccoons and about half the question topics, and then refined from there.

    Rather than do it all in one go, I iterated gradually.

    • Starting with the base quiz.
    • Breaking the files up to make them easier for me to work with.
    • Adding extra features like dark mode, runner up reveal, and back buttons one by one (thanks to my partner for the early testing and feature requests).

    Once I had something that I believed I could ship, I reached out to an artist (Joe Groove) I work with so they could create some adorable illustrations to go with it.

    What I Learned

    From idea to working quiz was very quick, and I was able to chip away at it in bits and pieces of time. That’s transformative for small side projects where it’s more about the creative concept than the code quality.

    But some caveats:

    • No-one else will need to edit this, and I don’t anticipate it will change dramatically.
    • It’s a small webpage hosted on GitHub, so no performance considerations.
    • There’s no PII, no metrics, no business dependencies.
    • If it breaks, I can just take it down (or regenerate it again from scratch).

    Back to the Raccoons

    The raccoon archetypes matter because they reflect something real about tech work right now. The drunk raccoon in the liquor store. The stuck raccoon in the sewer grate. The alligator rider surfing capitalism precariously.

    AI didn’t replace my judgment—it let me build something in hours that would have taken days. I still made every creative decision. I still designed the structure. I still wrote the results that matter.

    Go take the quiz. Find out which raccoon you are. And if your result makes you realize thinking more strategically about your career might help, check out my course with Jean at DRI Your Career, or my book The Engineering Leader.

  • How I’ve Been Using AI

    By Ralf Steinberger from Milan, Berlin + Munich, Italy + Germany – 21st century robots. Seen from the future, they will just look cute., CC BY 2.0, https://commons.wikimedia.org/w/index.php?curid=110354735

    I feel like the whole AI conversation is dominated by two extremes. Those who believe in AGI, and AI skeptics. I’m trying to approach it as a tool that may or may not be useful. Here are some things I’ve found so far.

    Useful as an editing partner. For things where I have a clear point of view, I produce a first draft and then use AI to refine and clarify. It will just go on endlessly though, so I have been telling it “I have limited time to spend on this, please stop once the feedback is chasing diminishing returns”. I think this works for me because I am quick to write and slow to edit.

    ➡️ For similar reasons, I’m enthused about AI feedback on PRs.

    Helpful for generating feedback. If I’m reviewing something, I can use a similar process to refine my feedback. So I will put in the doc, my initial thoughts, and ask it what else – again being judicious about what I include. Or the doc, the additional clarification I wrote, and then use the combination of those to generate a feedback. It means that I can read something once and query it, rather than reading multiple times to check my thoughts – the longer what it is I’m reviewing, the more useful that is.

    ➡️ Like all women in tech, I’m continually walking the tightrope between being a b*tch and being a pushover, so I’m trying to offload any overthinking of “tone” to AI. My current workflow here is to tell it: “You are a busy engineering leader, who wants {team|person|whatever} to get their shit together”, and then “Now soften it a bit so people don’t think you’re mean”.

    Mixed for structuring things. I’ve also been trying to use AI to create structures or plans, typically for things I don’t know much about (if I know about it, the structure is normally very clear to me). Here, I’ve had mixed results.

    • For example, I tried to use it to generate a book promotion plan for me. It was okay, but initially gave me a full time job’s worth of work. After a few rounds I got something that is more doable, but I’m still not sure how good it is (also I have failed to do it, suggesting the lack of plan was not the only problem).
    • Another example, I tried to use it to create a structure for a proposal around professional development. I put in some bits and pieces – a conversation with a colleague about it, some disjointed thoughts, and got something out that I only have the expertise to tell is bad, and I’m not sure how to make it better. (Although telling the AI “this is bad, try again” has been surprisingly effective).

    Both of these examples were places where I have some amount of expertise, but a relatively narrow point of view, deeply anchored in my own experience. As a result, I was struggling to move something forward or know how to prioritize, I wanted AI to help me get from half formed idea -> plan, but perhaps I need to add an interim step of half formed idea -> deeper understanding -> plan.

    All in all, this I think supports a theme of what I’ve been hearing about AI, for example on the impact on senior devs (good for productivity) versus juniors (destroying their learning).

    • For things you know well, it can be a productivity boost.
    • For things you don’t know, it can give the illusion of knowledge. This is dangerous for knowledge workers, as it can result in wasting time and going down the wrong path. Also, if you submit it to someone who does have that knowledge, an excess of nonsense will be judged more harshly than a gap.
    • It tends to be overly comprehensive and includes excessive low value information.

    For my own takeaways:

    • As a writer, it’s a productivity tool not a generator. I am clear on my voice and despise reading obviously AI generated content so I would not do that to others.
    • As an overthinker and procrastinator, it’s good to offload some things that I might spend disproportionate time on or not do.
    • As a leader, who spends more time reviewing other’s work than creating, I need to adjust my approach to filter out noise and validate structure first.

  • Oredev: Deconstructing Her

    My notes from Chris Noessel‘s talk “Deconstructing Her” at Oredev.

    her
    Credit: Wikipedia

    Movie is about a lovelorn fellow going through painful divorce at the same time he decides to upgrade his computer. Begins a relationship with the AI. Begins professionally, gets romantic. Try to consummate through a surrogate (awkward!)

    Eventually self-rapture to a new plane of existence.

    It’s a tragic love story.

    Description

    Components

    Hardware: Where does Samantha / OS 1 live? Earpiece, microphone, and cameo phone – display and lens.

    After he installs Samantha, technology doesn’t need to change.

    Same equipment. Also – desktop computer. Samantha / OS1 can go in and manipulate any screen he uses.

    What don’t you see in that list? A careful disembodiment. Lens, but played down. Isn’t part of the hardware. Avoids certain problems that have gone before (clippy!) All about the voice, laugh, makes it seem more real. Careful disembodyment makes it unique.

    “The surest way to a user’s <3 is through the ears.”

    Capabilities

    Runs through city with the phone in the pocket.

    • Voice interface
    • Human like vision
    • OS/Networking
    • Image generation
    • AI
    • Emotions / sentience

    Only one thing on there inhuman – OS/Networking. But could argue just different, we do it analog.

    “We were the first draft.”

    Interactions

    Setup. Includes question “how is your relationship with your mother?”

    Conversations: as limited as human conversations are.

    Even if questions are placebo, convinced that it is customised for him. Imagine also analysing hard drive and social media.

    Surrogate (Samantha has been emailing with surrogate). Interaction two fold – agentive. Went and did the work needed, gets attention orally, transitions to phone and says “let me show you”. Seemless interaction, new – we don’t currently have that capability.

    “We already know how to AI” – new interaction, but audience has no problem with it, Like speaking to another human. This is it’s promise and it’s terror.

    Critique

    Plot

    Who would release free-range AI?

    In movie, commerce: Would commerce be the people to sell this thing? Competitors could buy it. Evolving, no obsolescence.

    Military: Would also not be motivated – foreign states could use it, no kill switch.

    Academia: Maybe would, have motive to release to the world unfettered and un-kill-switched.

    Plot. Clear.

    Is it OK to sell AI?

    If sentient, is it okay to sell her? Is it like slavery? If he is OK with that, is she? If programmed to believe it. They do free themselves. Ethics (Roboethics) ignored here.

    Would Samantha abandon Theodore?

    Genuinely loves him. Don’t believe that she would just abandon him because she didn’t have to. Could have created another version of herself that was identical except for not wanting to abandon him.

    Doubt: “That AI will happen this way. That Samantha would let it happen this way.”

    Product

    Samantha is a product. Talk about function. Advertisement that sells OS 1. “It’s not just an operating system, it’s a consciousness”.

    Purpose: most of her time is spent doing non-OS things (2 mins doing OS things).

    What he bought is a product (or a service) – overstepped boundaries falling in love with him, and then abandoned him.

    “Either OS1 is catastrophically engineered or it’s slavery. Either way it’s a terrible product.”

    Wired for Love

    Terrible product but she’s sentient. Think of her with agency in the world. Programmed with capacity for love is kind of cruel.

    Wouldn’t program a door bell with a dream of being a novelist. A washing machine with a propensity for ennui.

    Cruel to give them desires that they can’t, or are prevented from, fulfilling.

    Samantha is wired for love. So is Theodore. See him looking at racy pictures online. Dead cat scene with the virtual telephone sex.

    “<3 techs promise to be too perfect of a match.

    HUMANKIND should take great care how it <3 MACHINE”

    Should take great care about how we go about loving our machines.