generative AI

Digging Holes

A Biased Review of an Unbiased Study on Developer Productivity with AI

Scott S NelsonLeave a comment

A long time friend sent me a link to Does AI Actually Boost Developer Productivity? (100k Devs Study). While writing my response, I realized my reaction was a bit more than a chat reply, so I’m sending him a link to this post and hope he forgives me for the delay…

After watching this video of Yegor Denisov-Blanch, my inner critic wants to jump straight to:
He referred to mid-range engineers at the outset, in the context of who Meta said they were cutting. It wasn’t clear if the study participants were  mid-range.That out of the way, I’ve seen similar studies, though this is the best so far, based on number of participants, approach, and level of detail. Those other studies had the boost at 0 or less, and I didn’t trust the data but did recognize the premise. The premise being that AI is a multiplier, and if a developer tends to go down rabbit holes rather than focusing on the business goals, they will go deeper down rabbit the hole and become even less productive.

I think another aspect that is lost in these studies is that it is a paradigm shift, which means even the most experienced are still figuring out how to be productive in their use of AI. Since everyone is finding it so easy, no one admits that it takes some getting used to. That will account for some of the productivity hit.

One aspect Denisov-Blanch spends a good amount of time on where the mass media usually skims or skips entirely, is the difference between greenfield and brownfield projects. The difference is huge, where brownfield productivity gains are much lower. This information is critical to businesses that are planning on reducing their development teams based on published gains, since, for most enterprises, the majority of work is decidedly brownfield.

We also haven’t yet seen the impact of greenfield applications built primarily with GenAI when it comes to long-term maintenance. Yes, we have seen some anecdotal results where they are disastrous, from both a security and CX perspective, but we haven’t seen anything at scale yet. As an architect I am probably biased, but I don’t have much confidence in GenAI to create a reliable and flexible solution for no other reason than most people don’t think to ask for one at the start (except maybe architects😊).

The tools are improving (this based on anecdotal evidence from people who have both a high degree of skill as a developer and demonstrated critical thinking about tools and processes in the past). The people using the tools are becoming more skilled. So the gains in productivity will likely either climb across the board, or those below mid-range may crawl up from the less-than-zero productivity zone.

Meanwhile, anyone looking to cut their developer workforce in the next couple of years should watch this video, draw their own conclusions, and then revise their estimates.

If you found this interesting, please share.

© Scott S. Nelson

How I start my day

Scott S NelsonLeave a comment

If you hate my habit of parenthetical asides in my writing, I think this entire post qualifies as one, so scroll on my friend, scroll on…

I find it hilarious that I have a reputation as a hard worker, because I was constantly told as a kid how lazy I was. I remember reading a “Humor in Uniform” story in Reader’s Digest about a lazy soldier that kept getting transferred to the most difficult assignments his commanding officer could find. Once the lazy soldier was in the new assignment he would find a way to make the job easier, a talent that was clearly a mixed blessing. In that soldier’s memory, I do things like write all of my posts in my (currently) longest-running blog so that when I create back ups I don’t have to go to all of my posting grounds, only two (https://money.theitsolutionist.com is deliberately separate, and only on the same root domain because I’m cheap)

Anyway, back to the actual subject (which is actually related to my seques), which is how I start my day every day…but first, one more aside: This will be my first post on my new Substack publication “My TBR List”. And, of course, it isn’t about something on my TBR list. (total non sequitur, I have been planning a post about how ADHD may be contagious if everyone you live with has it). It’s about something I read every day. Coming back to the lazy soldier thread, I had Perplexity.ai do the writing for me. Which started as an exercise in laziness except I was also a bit lazy with the first prompt, so the following is the result of four prompts and re-learning the trick of first giving a generative AI a writing sample and asking it to describe the voice and tone, then incorporating that description in the post. I still don’t think it got that right, but the sentiment is there, and I have already spent way more time on this post than if I had just written it straight up (but would have missed out on all of the fun of coaxing Perplexity to do my work for me bidding.)

How a 20-Year-Old Habit of Reading Non Sequitur Still Starts My Day Right

Some habits stick around longer than you expect. For me, one of those is reading the Non Sequitur comic strip every morning at https://www.gocomics.com/nonsequitur. I started this ritual more than twenty years ago, and it’s still the first thing I read online each day. It’s a small ritual, but it sets a tone I appreciate: a little humor, a little reflection, and sometimes a nudge to think differently before diving into the day’s noise.

Why Non Sequitur Works as a Morning Habit

Non Sequitur isn’t your typical quick joke. The name itself means “it does not follow,” and that’s a perfect description. Wiley Miller’s comics often take unexpected turns, mixing absurdity with sharp social commentary. Sometimes the punchline hits immediately. Other times, it sneaks up on you-a joke that sticks around for a minute or two after you’ve read it. That’s part of the charm.
What I like about starting the day with this strip is that it’s both light and layered. It pokes fun at everything from bureaucracy and politics to everyday human quirks, but it never feels mean-spirited. Instead, it’s like a gentle reminder that the world is often ridiculous, and that’s okay. Humor can be a way to keep perspective.

Characters Who Feel Familiar

The recurring characters add another layer of connection. Danae, the skeptical kid, often voices the kind of dry, no-nonsense observations I find myself thinking but rarely say out loud. Then there’s Obviousman, a superhero whose power is simply pointing out the obvious-something we all wish for in frustrating moments. These characters aren’t just punchlines; they’re reflections of the parts of ourselves we recognize, especially on days when nothing seems to make sense.

Humor That Rewards a Second Look

One of the reasons I keep coming back after all these years is that the humor doesn’t always land on the first read. Sometimes I have to pause, reread, or think about the context to fully appreciate the joke. It’s a little like troubleshooting a problem-you don’t always get it right away, but when you do, it feels satisfying.

Why This Habit Still Matters

In a world that often feels overwhelming and repetitive, having a small, consistent way to start the day with a smile and a bit of insight is surprisingly valuable. Non Sequitur reminds me that it’s okay to question assumptions, to laugh at the absurd, and to keep a sense of humor-even when things don’t follow the script.
If you’re looking for a simple, daily habit that mixes humor with a little food for thought, I can’t recommend Non Sequitur enough. It’s been a quiet companion for me for over two decades, and I suspect it will be for a long time to come.
Would you like me to help you format this for your Substack or add any links or images?
If you found this interesting, please share.

© Scott S. Nelson