Good science fiction predicts not the car, but the traffic jam

FWIW, I've never had a year typify "You do not rise to the level of your goals, you fail to the level of your systems" more than 2025. Emotionally and energy wise --fires and fascism factored in -- I've never been more punched out. But I worked and lined up more work based on old habits saving me.

Attacking a data center like it’s The Bastille, everyone fleeing with a bushels of RAM to desolder and resolder the chips. One peasant coded guy with a wheelbarrow of SFP+ modules. Manifesting this for 2026.

Pretty wild that you have to be like, a graphic and web designer to tell your career and work story these days I feel like we don't talk that much about the escalation of these aesthetics and expectations

25 years ago, I encouraged dev teams to take a multi-dimensional view of their business goals, balancing different - perhaps competing - perspectives. In 2026, I just hope they *have* a business goal and that somebody will tell them what it is. That's how far we've come.

How I teach software design starts with a problem. It's hard to find vegan takeaway round here. I keep forgetting to buy bin liners. I miss the best bands when they're in town. Not "An app..." or "An API..." or "An agent..." A problem. Remember those? I mean, it's not like we ran out of them.

“I’m honored to be your ancestor and believe that disabled oracles like us will light the way to the future. Don’t let the bastards grind you down. I love you all.”

4 principles of modular design: 1. One reason to change 2. Hides internal workings 3. Has swappable dependencies 4. Usage-driven interfaces I'm adding a 5th: 5. Does what it says on the tin, and says what it does on the tin It should be obvious what it does, and it should bloody well do it!

The teams producing reliable software are only able to do that because they test continuously. Do you know what else they can do because they test continuously? Ship any time the customer wants. Short lead times and reliability go hand-in-hand. Long delays and buggy software likewise.

HAHAHAHA pubmed.ncbi.nlm.nih.gov/30101594/

How I'm walking out of 2025

"The software development game has completely changed because <goes on to describe what I was trying to tell them 28 years ago that I learned from a book published in the 1970s>"

On the McKinsey report that "discovered" smaller dev teams are more productive... I coined a term: "project heat" I'd estimate how much it would cost with a team of 4 good devs. Then I'd compare that to the actual. The difference is "lost to heat". I've seen > 99% on more than one occasion.

Every year, the snowman claims another child from the British countryside.

As 2025 draws to a close, my hopes "AI" was going to force more teams to address the real bottlenecks in development fade into the distance. 95% will just ship less reliable software, and take longer and spend more £ doing it. And users will be coerced into eating the costs. Business as usual.

Dear listeners, When the days feel narrow and the noise gets loud, try a song. Not as escape — as evidence. Someone, somewhere, once felt this too and made something that still holds. If hope is hiding, it often hums first.

"Coding is no longer the bottleneck". So what you're saying is that NOTHING HAS CHANGED.

It's fun watching organisations try to reduce the queues at the checkout by making the baskets bigger.

Because of the power law distribution of examples coding LLMs are trained on - 99% Stack Overflow answers etc - well-factored code's out-of-distribution. We shouldn't expect it from them, because they've "seen" so little of it. It's their "wine glass full to the brim".

915 MHz Forecast: Rolling Your Own Offline Weather Station

Libby (the library app) just launched the best new interface for book recommendations I’ve seen in years. #code4lib

It seems to me that we all need to significantly narrow the definition of a "stakeholder" to people who have a vested interest in our work's success. Your failure must be their failure, your success their success. 1/12

If you read a book this year, im proud of you 🫶🏽 reading and literacy has become an act of rebellion in such an illiterate world

Believing that LLMs producing code faster will speed up the production of software is the same logical mistake that car manufacturers made when they thought introducing robots to an assembly line would immediately lead to more profits. We are still struggling to understand the theory of constraints.

Merry Christmas and by Christmas I am of course referring to that time a baby was born to a pair of impoverished refugees who were fleeing a tyrant

How can I put it in a way CS grads can understand? If we were to apply time complexity to software development as an algorithm, coding would disappear as a factor, and you'd be left with a Big O that's all about batch size and feedback loops.

"The bottleneck is no longer code" Have I stepped through a timewarp into 1952? Because that was the last time the bottleneck was code.

When you say "we don't need user research, I already know what users want" this is what you look like

A gov trick I've seen deployed often is to announce spending commitments that *sound* like a lot of money. Exhibit A: Labour government announces additional funding of £725M for 50,000 more apprenticeships over 3 years. 2.4 million people will reach working age in the same period.

NO MORE NIHAO ONLY NI-HOWDY

I sure would like to see the Virginia General Assembly legalize balcony solar in 2026. Let people buy a $100 solar panel at Costco, plug it into an outlet on their porch, and make electricity. Germany does it. Utah does it. Virginia can do it too.

And one of my most radical beliefs is this is really worth doing even when you don't always succeed because it will protect YOU, yes you the leader, you the infra expert, you the DevEx lead, guess what your psychology also matters and this kind of work can help you feel less completely alone

I don't have an ounce of chill, so I love this.

#ResonantComputing makes me feel hopeful resonantcomputing.org

We are about to discover the far right equivalent of "everyone at the leftist commune wants to design the ideal society and no one wants to do laundry"

They are!! Once you realise that almost everything that can be gamed is gonna be gamed… well, it opens up a lot of flexibility. Levers can be used for good, or bad, but they’re changeable! Influenceable! If people are gonna abuse metrics, make the abuse deliciously healthy and helpful to people

I'm going to make a pitch here, if anyone who runs a software org wants to partner with me to tell this story in big, quantifiable, board-resilient numbers in 2026, inbox is open. You and I both know it falls apart but your bosses struggle to see it.

Hire junior developers and mentor them. Treat them well and pay them decently. It’s not that complicated.

That’s bad, right?

NEWS: The Institute of Museum & Library Services has restored ALL previously canceled federal grants to libraries, following a ruling by a federal judge in Rhode Island last month. This is a massive win for libraries & communities in every state & territory! Learn more: www.ala.org/news/2025/12...

That executives still haven't figured this out is a testament to the Peter Principle

But the big lesson for me was just how different things can look when you take an evidence-based approach. It was my "enlightenment". It's said just how many are still living in the Dark Ages where, if it feels right, it must be true.

“AI is clearly sticking around so you have to get used to it” wrong. I don’t have to get used to shit. I am a practiced hater and I can keep this going for decades if I am required to

I hope someone draws this wonderful story about Tom Stoppard's Arcadia and the power of the arts to help us see things differently, to the attention of our Education Secretary. Do read it, it will lift your spirits.

ANYWAY Buy from small businesses Buy from independent creators Buy passion projects Buy commissions Buy art In a society that's decided that money is final word, be aware of where yours goes and to support what Yes it sucks, I don't like it either, but pretending it's not a thing does not help!

the ones who walk away from LLMs

If you asked me to characterise software development in its essence, I'd say it's a process of reducing uncertainty. Otherwise known as "learning". codemanship.wordpress.com/2025/11/28/w...

I am once again pointing to the imperial boomerang.

To echo what others are saying, I am also seeing this play out in real time. Daily, I get calls from people I haven’t seen in a while, asking if I have any job leads. The job market is bleak. But I am also seeing so much mutual aid, that part is inspiring.

"It's really good at summarising code I don't understand". But, weirdly, it's pretty bad at summarising code I *do* understand.