iiinova — where your digital self keeps evolving

Dragan Stepanović @dstepanovic.bsky.social · 01/10/2026

I don't know about you, but Someone in the IT Services and IT Consulting industry from Berlin Metropolitan Area viewed my profile on LinkedIn.

010

Reposted by Dragan Stepanović

Dr. Elizabeth Sawin @bethsawin.bsky.social · 28/09/2026

Some good things to work on this week as collapses and power grabs surround and the operating instructions of the industrial growth domination and supremacy worldview are further revealed and clarified (remember apocalypse means "uncovering")

3179

Dragan Stepanović @dstepanovic.bsky.social · 27/09/2026

Inherent delays, distant effects, non-linearity, path dependence and tipping points: that's what you're dealing with in socio-technical systems.

010

Dragan Stepanović @dstepanovic.bsky.social · 27/09/2026

No amount of upfront experimentation can compensate for the inherent delays in complex systems. Effects take time to feed through. By the time they do, you might already be setting up an email account at your next gig, while your previous company will be living with the consequences.

110

Dragan Stepanović @dstepanovic.bsky.social · 27/09/2026

If you think the effects of dissolving the team's mental model of the system, by removing humans (authors and reviewers) from the path to production, will show up soon enough to reverse course safely, I have news for you.

162

Reposted by Dragan Stepanović

Ursula K. Le Guin Bot @leguinbot.bsky.social · 26/09/2026

He had grown up in a country run by politicians who sent the pilots to man the bombers to kill the babies to make the world safe for children to grow up in.

015845

Dragan Stepanović @dstepanovic.bsky.social · 23/09/2026

There's a big difference between emergence (in complex systems) and stochasticity, and there's often confusion of the latter with the former when reasoning about LLMs and their properties.

030

Dragan Stepanović @dstepanovic.bsky.social · 20/09/2026

Either way I need some setters and getters in there

030

Dragan Stepanović @dstepanovic.bsky.social · 19/09/2026

Instead of trying to bend reality to what we think it should be, recognizing it for what it actually is.

021

Dragan Stepanović @dstepanovic.bsky.social · 19/09/2026

Independence as a means of scaling is often a mirage. Systems will always find a way to prove to you how interdependent things are. "Dancing with the system" starts with recognizing interdependence, not seeking the holy grail of independence.

220

Dragan Stepanović @dstepanovic.bsky.social · 11/09/2026

Believe it or not, code review stops being a "bottleneck" as soon as you stop overproducing code.

141

Reposted by Dragan Stepanović

Nick Tune @nick-tune.me · 11/09/2026

As a result, this is how I now judge LLMs: Are they a good pair programmer or assistant. Can they follow instructions, can they be taught good design principles, can they be taught not to do stupid workarounds? I don't care how genius the are, good maintainable code requires mastering the basics.

053

Reposted by Dragan Stepanović

Nick Tune @nick-tune.me · 11/09/2026

This is also why big up-front planning is not sufficient. You can't predict all the guardrails and architecture rules the agent will hit during implementation (although you should try to anticipate as much as possible IMO). Everything I see points to software engineers being augmented by AI.

121

Reposted by Dragan Stepanović

Nick Tune @nick-tune.me · 11/09/2026

And this is why you need to stay close to the code as an engineer. So you can actually understand and contribute to these discussions. You still own the design.

131

Dragan Stepanović @dstepanovic.bsky.social · 10/09/2026

Thanks, Nick. Your posts make feel sane (again). (and I do use LLMs every day)

000

Dragan Stepanović @dstepanovic.bsky.social · 10/09/2026

We indeed are. I know of Adam and his work (recommend it often) and I hear, but I also disagree.

010

Dragan Stepanović @dstepanovic.bsky.social · 10/09/2026

Lots of slower ones (as we climb the test pyramid). The more cumulative functionality the more of those and their execution time scales the worst.

100

Dragan Stepanović @dstepanovic.bsky.social · 10/09/2026

That's just one problem. A couple more due to moving up the size of the System Under Test which reduces the (valuable) pressure on the design and teasing out domain concepts which would help agents to produce higher quality output. So, marginal deterioration.

110

Dragan Stepanović @dstepanovic.bsky.social · 10/09/2026

Wrong way to go if you ask me. LLMs can speed test generation, but they can't speed up test execution. And the bigger the difference of ratios of these two, the faster the feedback loop slows down.

200

Dragan Stepanović @dstepanovic.bsky.social · 05/09/2026

You renamed way too many methods in Jetbrains IDE, didn't you?

100

Dragan Stepanović @dstepanovic.bsky.social · 31/08/2026

If deleveraging is a regression to the mean, or "what it should have been all along," then our perception of products built during nearly 15 years of very low rates was also unrealistic and unsustainable. For a lot of people, that era is all they've ever known of digital products they use today.

000

Dragan Stepanović @dstepanovic.bsky.social · 31/08/2026

What strikes me is that leverage in the financial system (eventually leading to higher volatility) also translates into leverage in products' perceived value and experience.

100

Dragan Stepanović @dstepanovic.bsky.social · 31/08/2026

The more expensive borrowing is, the tighter funding gets, the harder investors push for profitability, and the more aggressively companies try to monetize previously subsidized products. And remember that runway is a stock, not a flow, so rate hikes show up in product decisions with a lag.

100

Dragan Stepanović @dstepanovic.bsky.social · 31/08/2026

It's interesting to think about the dynamics between product enshittification, which I feel has been on a sharp rise over the last couple of years, and central banks raising interest rates.

120

Dragan Stepanović @dstepanovic.bsky.social · 12/08/2026

One more day, one more "reading code when LLMs are spitting out so much code is impractical, so you should give it up" LinkedIn post.

020

Reposted by Dragan Stepanović

Ursula K. Le Guin Bot @leguinbot.bsky.social · 11/08/2026

The world is, no matter how we think it ought to be. You have to be with it. You have to let it be.

0428

Dragan Stepanović @dstepanovic.bsky.social · 11/08/2026

And I can understand that a mental model that reduces collaboration to that goal can be drawn to the idea of an LLM dark factory, with harnesses and constraints that attempt to rein in LLMs' code-generation output. But let's not confuse the output for the process that created it.

000

Dragan Stepanović @dstepanovic.bsky.social · 11/08/2026

Collaboration - of which throwing PRs over the wall is likely the most suboptimal way to go about it, if it can be called collaboration at all - was never only about making sure the code is correct.

130

Dragan Stepanović @dstepanovic.bsky.social · 09/08/2026

It's the mental model behind "more things in progress and higher utilization = more things delivered" that is the problem, not the lack of capacity.

000

Dragan Stepanović @dstepanovic.bsky.social · 09/08/2026

If teams' utilization is so high that it causes a non-linear increase in delays, adding more people while keeping the same utilization rate won't help reduce the delays you wanted to reduce by adding more people.

111

Dragan Stepanović @dstepanovic.bsky.social · 30/07/2026

Too many of the same speakers at the same conferences.

130

Dragan Stepanović @dstepanovic.bsky.social · 30/07/2026

Software delivery measurement vendors would like you to improve the metrics, but not to the point where you change the way of working (XP) to the point that you no longer need the vendor or the measurements.

031

Dragan Stepanović @dstepanovic.bsky.social · 30/07/2026

And nobody with positional power likes when the team is so good that they become a bottleneck.

010

Reposted by Dragan Stepanović

Jason Gorman @jasongorman.bsky.social · 30/07/2026

The risk of doing great work in software development is that it can look like the team isn't trying as hard as the teams heroically fighting fires (that they started)

32710

Reposted by Dragan Stepanović

Natia Kurdadze @natiakurdadze.bsky.social · 19/12/2024

The Toyota Way isn’t just about small gains. At its heart? Kaizen.

152

Dragan Stepanović @dstepanovic.bsky.social · 17/07/2026

The purpose of all automation and cost reduction we got as a result of technological improvement since the start of the Industrial Revolution was to spend more time on social networks and consuming polarizing and addictive content. The purpose of the system is what it does. (POSIWID).

020

Dragan Stepanović @dstepanovic.bsky.social · 14/07/2026

Is writing code nearly free, or writing _plausible-looking code_ nearly free?

140

Dragan Stepanović @dstepanovic.bsky.social · 13/07/2026

The right time to intervene in systems where an exponential, reinforcing feedback loop is dominating the system's behavior is when it seems it's too soon. "Let's wait and see" is a guarantee that it'll be too late.

010

Reposted by Dragan Stepanović

Tom Geraghty @tomgeraghty.bsky.social · 15/12/2024

“A numerical goal leads to distortion and faking, especially when the system is not capable of meeting the goal. Anybody will meet the quota (goal) allotted to him. He is not responsible for the losses so generated.” W E Deming, in “The New Economics”

073

Dragan Stepanović @dstepanovic.bsky.social · 27/06/2026

The problem is that people deciding if developers should be writing code are very often the ones who have never had enough experience with writing code to have a good idea what the value of writing code is and the purpose it serves, but they are convinced they know.

174

Reposted by Dragan Stepanović

Gale Sinatra @galesinatra.bsky.social · 27/06/2026

I think we just passed oops.

451173235800

Dragan Stepanović @dstepanovic.bsky.social · 26/06/2026

The counterintuitive idea about managing flow is that the leverage is most often much less about speeding up slow parts and much more about slowing down parts that are running too fast when taking into account the whole system.

020

Reposted by Dragan Stepanović

Ted M. Young (he/him) aka @JitterTed @ted.dev · 21/06/2026

I've seen comments about accepting LLM-generated code in that we don't/can't understand it and excusing that because: "There's always code in the system that we don't understand"

131

Reposted by Dragan Stepanović

Ted M. Young (he/him) aka @JitterTed @ted.dev · 21/06/2026

Except, LLM-gen'd and human-written code are different. One was written by someone with a mental model. Modified by humans with another mental model (correct or not), etc. The other is generated by statistical, stochastic methods with constraints. These are not the same.

174