Writing on Hunter Software Consulting

So Your Deadline Is Slipping

Fri, 09 Oct 2026 00:03:43 -0500

We’ve all been here - and if you haven’t yet, trust me, you will be. You’re a leader. You’ve set expectations. This year Black Friday is Nov 27. It’s the biggest day of the year for your customers. That means, it’s the biggest day of the year for you. Your Superbowl, your Olympics, your World Cup. You’ve got several projects in the works before them, but let’s call one Project X. Project X is your darling, and stands to both help your customers and make you a lot of money - especially on Black Friday.

One Line, 25M Revenue

Fri, 02 Oct 2026 00:03:43 -0500

“I mean, it’s one bug, Michael. What could it cost, ten dollars?”

Do you have some wild war stories? You know, the unbelievable ones. Today I want to share one of those stories. On the scale of the economy and enterprise, it’s not small potatoes, but it is potatoes. I live in Startup Land, where every dollar counts. I do not think I would be here writing this blog if not for this event. I might not be running my business if not for this event. This event was the one where I went from “I think there’s an opportunity” to “I was incredibly pessimistic and naive about the available opportunity”.

You're holding AI wrong

Thu, 24 Sep 2026 07:46:43 -0500

I’d like to lead with - I use AI very actively. As we speak (well, as I write), I have 7 discrete projects running X (X being a large number) of agents doing forward development. Am I moving faster than I was pre-AI? Sure. Somewhat. Probably not double - I feel like the 2-30% increase seems about right. 20-35% in what matters most, but it does allow me to do more “stuff” in parallel. But today, I’m not talking about that. I’m here to talk about how I actually used AI to get stuff done, very rapidly.

The Release Tax

Sun, 20 Sep 2026 07:46:43 -0500

I’ve had a rough week. One organization I’m associated with really struggled with their release. In my humble opinion, you’re moving slow if you don’t ship every day - calibrated across orgs, that means I want to scale deployment roughly linearly with headcount. Some people will be on large initiatives that take a while to ship, others will ship 15x in a day, so I’ve observed it to be a valid baseline. This client takes about 9 days average to ship, for a team of 9. This is also obviously complicated by AI velocity - each release is large and bulky (we are fixing this), and as a result high risk, with many changes. But the release itself was a 4 day marathon, shared among a set of engineers responsible for the release.

3 Side Hustles In a Trench Coat

Sun, 13 Sep 2026 07:46:43 -0500

So, I listen to a lot of pitches. One thing that I notice frequently is that there are pitches that are a business, and pitches that are 3 side hustles in a trench coat. This is a little hard to describe, but you know it when you see it. It’s the B2B company that nonetheless has a B2C social media application tacked on. Or the product that does everything.

Bigger is not Better

Sun, 06 Sep 2026 07:46:43 -0500

Bigger is better, right? Bigger company, more engineers, more product managers, more output. That’s unequivocally good, right? Well, I think we all know that this isn’t the case at large, but we still desire more and more and more.

Your Product Threatens Your Product

Sun, 30 Aug 2026 07:46:43 -0500

I periodically return to several ideas, as I talk to founders. One central one, that seems especially relevant with AI, is that your product threatens your product. This of course sounds a bit nonsensical when put that way, so let’s dig in.

Ship to Burn

Sun, 09 Aug 2026 07:46:43 -0500

Recently I had the pleasure of taking a workshop by Jeff Patton. We covered a ton of concepts and practices that I had experienced firsthand. Two pithy terms he uses are “build to learn” and “build to earn”. You might notice this is the foundation of Dual Track DevelopmentYou can read some writing about it here for example. I’ll cover the two for you briefly before covering a third category that I consider the bane of my existence - one that keeps recurring too.

Straight-Jackets and Exoskeletons

Sun, 19 Jul 2026 07:46:43 -0500

It’s overly pithy to reduce things to a binary, but I’ve done it before and I’ll do it again. I think there are two major categories of businesses - exoskeletons and straight-jackets. This is to say, businesses (and more broadly, organizations) that largely empower the people within them, and ones that disempower them.

Scissors, Glue, and AI

Sun, 12 Jul 2026 07:46:43 -0500

So, there’s a story I frequently tell about my first “real” job. I’ve thought about a pithy title for this, including something like “Efficiency is a Lie” or something otherwise provocative, but I’m not sure I believe that enough to use it. That said, I’d like you to keep AI in mind as I tell this story. I’ll return to it at the end, but I hope you see the through-line.

Feedback Loops upon Loops upon Loops

Fri, 03 Jul 2026 07:46:43 -0500

I keep returning to loops lately, with my work with engineering organizations (and with how popular AI loops have been across LinkedIn). When we look at much of nature, cycles are a fundamental constant. Earth moves around the sun (giving us seasons). Earth spins on its axis (giving us days). The moon moves through the lunar cycle (giving us the tides and such). Life has cycles as well, from infant to adult back to infant.

Cowboys, Frontiersmen, Settlers, Townspeople, Cityfolk

Sun, 28 Jun 2026 07:46:43 -0500

“What got you here won’t get you there” applies not only to founders, but to the organizations they build. An underappreciated aspect of this is risk and chaos, and how it appeals to discrete personalities. I consider most of these archetype to respond mostly to certain levels of organizational chaos - and as such, there are areas where they thrive, and areas where they are out of their depth. I’ll hearken back partially to Wardley’s Pioneers, Settlers, Town Planners as part of this.

The Fractional CPTO

Sun, 21 Jun 2026 07:46:43 -0500

I’m often asked “what does a fCPTO, really”? That is, fractional Chief Product Technology Officer. I thought I’d break it down into some real activities that I’ve done recently.

AI & The Red Queen

Sun, 14 Jun 2026 07:46:43 -0500

“My dear, here we must run as fast as we can, just to stay in place. And if you wish to go anywhere you must run twice as fast as that.” ― Lewis Carroll, Alice in Wonderland

A Very Brief History of AI

Sun, 07 Jun 2026 07:46:43 -0500

One thing that strikes me recently is how disparate AI adoption has become. To that effect, I’d like to recap some of the recent eras of AI we’ve seen. As a note, I’m not trying to capture the entire history of machine learning or artificial intelligence. I’m more focused on “AI” as we currently refer to it - LLMs. I hope to taxonomize and identify a few key concepts that have appeared as a result. Do note that these periods are rough - some people were ahead of the game and many were behind.

AgentEx and Cognitive Load

Sun, 31 May 2026 07:46:43 -0500

I’ve said it a few times, and I’ll say it again - AgentEx is DevEx. What scales engineering success scales agentic performance. I usually talk about feedback loops, but that’s one part of it. Today I want to talk a little about cognitive load.

Why Does AI Scale With Engineering Skill?

Sat, 23 May 2026 07:46:43 -0500

I’m frequently asked why some people see more success with AI than others. The following are largely hypotheses and anecdotes (not data!), but I’ll try to list several with some conjecture. I’m personally looking forward to more hard scientific research in this area I can cite one day.

Local Maxima

Sun, 17 May 2026 07:46:43 -0500

One concept I keep returning to is that of local maxima. If you recall, local maxima occurs when a graph swells to a particular point, dips, and then ascends to a prior point. It’s a recurring problem in optimization - naieve optimization will always reach the local maximum and stop. There are many techniques to avoid this - picking a random starting point, jolting forward when you think you’re at an apex, and retrying, among others.

Agentic Tasks: a Few Samples

Sun, 10 May 2026 07:46:43 -0500

What models are capable of and what they aren’t capable of continues to astound me. I’ve been pushing myself to use pi for a number of tasks these few weeks and wanted to share some learnings.

About Weirdness Budgets

Tue, 05 May 2026 07:46:43 -0500

What is a weirdness budget? You might have heard about this concept as an “innovation budget” or a “risk budget”. I personally don’t like the term “risk budget”, as I think that underestimates risk itself.

From Anthropic Toward Pi (for now)

Mon, 27 Apr 2026 07:46:43 -0500

I really see this moment as an inflection point (as so many recent moments have been). And actually that’s overstating it - the moment was a few weeks ago with Opus 4.7. I see us moving towards another period of tool flux.

On Slop: April 2026

Tue, 21 Apr 2026 06:46:43 -0500

I’d like to preface with a disclaimer: I do think you should be using agentic tools and engaging with them. This is less a discussion about the tools themselves, and more about how you should go about using them (and a particularly sharp edge I see).

The Career Ladder

Mon, 20 Apr 2026 16:46:43 -0500

One interesting failure I see in many startups is a lack of a career ladder. The common refrain is “We’re early and scrappy, we don’t need hierarchy!” - and I’m inclined to agree! But you also need to hire the right talent, which starts with a ladder.

Why Core Values?

Sun, 12 Apr 2026 16:46:43 -0500

Core values are a concept I return to time and time again, across the companies I advise. And core values are surprisingly controversial to many founders. Many would rather just work on something tactical and tactile, like building the product or making sales. But even in a team of 1, culture and core values tend to raise their head.

Onboarding: Time to First Release

Mon, 30 Mar 2026 12:14:43 -0500

I’ve joined many engineering organizations, and there’s one metric I return to that differentiated successful organizations from unsuccessful ones - time to onboard, as measured by time to first release. I should add that “success” here is subjective - the “successful” ones were able to extract much more objective value (in terms of ROI in increased ARR) from my fellow engineers and I, while the “unsuccessful” ones struggled or failed to extract similar value.

Process or Outcome Orientation is the Wrong Framing

Mon, 23 Mar 2026 14:14:43 -0500

Process orientation vs outcome orientation has been debated endlessly over the years. In fact, aspects of this discussion can be traced back to Marcus Aurelius’s Meditations. The current thread seems to draw back to Peter Drucker’s “Management by Objectives” in The Practice of Management. But what if this is the wrong framing? Process vs Objective frames this discussion as a focus on ROI (outcomes), or on “mindlessly” executing a playbook. But how do you reach repeatable successful ROI? With a repeatable process. So process and objectives are actually more tightly bound than it seems.

On the Iron Triangle

Mon, 16 Mar 2026 14:14:43 -0500

The Iron Triangle is a classic model of constraints. According to the model, there are Quality, Time, and Cost. Pick one, at most two. But what if this model is inherently misleading?

Your First Engineering Hire

Mon, 09 Mar 2026 14:14:43 -0500

Many companies come to me with questions about early technical hiring. Do I need a cofounder? Do I need a CTO? There’s several major recurring personas I see in this process. I’ll attempt to at least list them, so you can see which one is most relevant to your situation.

The Full Rewrite: AI edition

Mon, 02 Mar 2026 14:14:43 -0500

I’ve had full rewrites come up several times recently, and so I wanted to write about them. The full rewrite is a kind of siren song - beckoning on the horizon is a future with higher velocity and lower technical debt. But actually doing one is one of the riskiest types of projects and organization can take on. The topic has been covered extensively, and yet it recurs, often, for good reasons. One fundamental question is - has AI & AI assisted coding changed the space? I’ll look at different types of risk in 3 areas to try to unpack this.

AI & Shadow Engineering

Mon, 23 Feb 2026 12:53:43 -0500

At the current moment, “Shadow AI” and “SaaS is dead” are two popular topics. I realized that there was an interesting intersection here. With vibe coding, anyone can replicate a web product’s functionality in the weekend (or so the theory goes). The reality is a little more complicated

DevEx is now supreme in AI Enablement

Sun, 15 Feb 2026 16:38:43 -0500

I have frequent conversations (and frequently see conversations on LinkedIn) focused on whether Codex, Cursor, or Gemini is the best too. This is the wrong question. At the present moment, model choice matters significantly less than Developer Experience, or DevEx.

Go Out and Build

Mon, 09 Feb 2026 12:53:43 -0500

Do you have an idea for an app? For a website? For a chrome extension? For a solution to a problem?

Go and build it! Today!

This moment, now, is unprecedented in human history. We have incredible creative technology, which makes computer use significantly easier. It’s subsidized to a level that I don’t believe we’ve seen before. It’s accessible for literal dollars.

Our Culture & Engineering Integration

Mon, 09 Feb 2026 08:53:43 -0500

Our corporate culture today has a recurring message that I think is extremely problematic: that Engineering should be special and different from other organizations. This has a number of knock-on effects across departments that are problematic.

AI Velocity: Planning Vs Experimentation

Tue, 03 Feb 2026 12:58:43 -0500

In talking with founders about their AI strategy, I’ve noticed consolidation into two major camps of AI assisted acceleration: Planners and Experimenters. Each of these has strengths and weaknesses in their delivery execution.

The Value of PRDs (Product Requirements Doc) in the Age of AI

Mon, 02 Feb 2026 12:58:43 -0500

I had a few recurring conversations recently in one of my communities: Engineers were preparing high volumes of code in pull requests, that were then getting jammed up in review. This is a common problem that has been exacerbated with AI.