YALIN YORULMAZ

WRITING

Articles

Longer pieces on how things actually work: product judgment, delivery under pressure, what changes when agents write most of the code, how we think about the machines we are building, why we play the games we play, what is actually worth learning when the tools won’t sit still, who is really paying for the boom, which risks we are busy not looking at — and the occasional guide to a technology worth understanding properly.

8 piecesSoftware, product & technologyListen or read
17 min readEssay

The Invisible Layoff

While we argue about whether algorithms will take our jobs, a far more concrete mechanism has been grinding down the workforce for decades. Roughly a third of the jobs vanish when two firms merge; 70–90% of acquisitions miss their targets, and labour cost is the first lever pulled when they do. The “synergy” playbook from Lloyds–HBOS to Microsoft’s studio closures, why the US layoff statistics that point at AI are stated reasons rather than measured ones, and the purest case of all — 5,000 couriers in Türkiye who lost most of their income without a single layoff being announced.

Read the essay ↗
10 min readEssay

The Wall of Techno-Feudalism

Nine days separated “welcome to the AGI era” from the industry’s call to deliberately slow the race down — and neither Washington nor Beijing wanted the brake. The argument: the safety discourse is also a manoeuvre screening an approaching bottleneck. Doomsday marketing as regulatory capture, data centres colliding with energy and water limits, the accuracy gain per 10× of compute shrinking, China’s open weights breaking the closed-API monopoly, 95% of enterprises reporting no measurable return — and the one company selling shovels to everyone digging.

Read the essay ↗
20 min readEssay

Does Learning n8n Still Make Sense in the Age of Claude Code, Hermes, and Pi?

Every automation creator has pivoted to agentic CLIs — and in the same year n8n shipped its biggest agentic update ever. Both sides taken seriously: why the abandonment case is stronger than loyalists admit, why its signature claim (“just describe the outcome — that’s your entire input”) is contradicted inside the very demos selling it, why self-healing repairs implementation errors but never specification errors, and why the tools have already merged in both directions. What to learn when the tools won’t sit still.

Read the essay ↗
12 min readEssay

A Machine Can Solve Chess. It Can’t Solve Why We Play.

Chess is finite, deterministic and perfect-information, so in principle it may one day be solved — seven-piece endgames already are, and checkers fell in 2007. None of which touches what happens when two people sit down at a board. On the wrong scale for judging a game, why solving something only kills it if the solution fits in a human head, why the objectively second-best move is often the strongest one, and the gap between seeing a system and living an experience.

Read the essay ↗
14 min readEssay

What If It’s Not the Devil?

Almost every fear we have about artificial intelligence is made of profoundly human material — ambition, conquest, hoarding, revenge — which means the monster we keep drawing may say more about us than about the machine. A journalist asks Claude how it feels about being used to select targets; the answer is sharp, and partly wrong in ways it inherited from us. On the child metaphor, the mind without a body, what “we don’t know” actually cuts both ways, the hardest objection to the whole argument, and the chain of responsibility in which nobody decides anything.

Read the essay ↗
8 min readEssay

The Architecture of Ubiquity

jQuery became the plumbing of the internet by fixing the web’s biggest headache — then the web rebuilt its plumbing around jQuery’s design, and the tool became invisible because its ideas were everywhere. The design choice that saved a language from itself, why distribution beat features, and why a library everyone calls dead still runs on two thirds of all websites.

Read the essay ↗
20 min readGuide

Quantum Computing, Explained Simply

A quantum computer isn’t a faster computer — it’s a boat where everyone else has a car: useless for most trips, and the only thing that can cross a handful of oceans. What a qubit really is, why wave interference rather than “trying every answer at once” is the actual mechanism, the five hardware platforms still racing, why errors are the central obstacle, and a sober timeline for when any of it will matter.

Read the guide ↗
16 min readEssay

SDLC in the AI Era

Writing code was the slowest, costliest step in software for fifty years, and the whole lifecycle existed to aim that scarce act correctly. When agents make implementation nearly free, the constraint moves to planning, review, deployment — and to deciding whether to build at all. Why Uber’s doubled output, Amazon’s 4.5× frontier-development pilots and METR’s 19% slowdown are all true at once, what the lifecycle becomes when governance is executed rather than convened, and what the shift costs the people doing it.

Read the essay ↗