I'm glad to see the first thing you mentioned was the value of purchasing software.

For some reason software developers are particularly averse to paying for software, even when it pays for itself.

My current company spent over $10k + at least a month of developer salary building an OKR app for the company, so probably in total $25k. One look at the SaaS landscape there would have revealed plenty of $5/mo/seat solutions that would take DECADES to break even with the investment we already made. To add insult to injury, the app was vibe-coded and barely works.

In my experience, software that wants to help you is cheap. Software that wants to _trap_ you (Salesforce, Oracle, the Cloud) is expensive, and any software you write yourself falls into the latter category by default (you own it, you maintain it).

Dealing with SaaS doesn't just mean accounting for the subscription cost of the software. There's also the fact that said software can change its terms, pricing or functionality at any time (or disappear entirely) - often a time most inconvenient for your business.

In contrast, your own solution can be built and maintained at your own schedule and the only changes will be those you decide.

Indeed, the article mentions Wikipedia experiencing similar scraping pains, even though they already DO have bulk data available.

Who are running these bots? I presume developers at all of the frontier labs know (or at least would know to look for) Wikipedia has bulk APIs for automated access. Unnecessary scraping increases their workload/costs too, so why in 2026 is this still a problem?

Black market and gray market data. All the firms want data. All the other firms want data. The banks want data. The other criminals also want data for their crimes and schemes. Oh insurance companies, and the ATS systems. Everybody wants as much data as they can get and they don't care how they get it.

This data selling also happens with leaks of all kinds like medical data, often to current or future employers, health insurance companies, etc.

The maps haven't changed at all, this is a feature at destructuring sites.

Sorry if I wasn’t clear, I was referring to the maps that are being destructured.

> The mess just moves

That's exactly the point. The author recognizes that people will extract components, and that is exactly why tailwind is attractive.

Plain CSS offers a componentization layer over (mostly) classes, but modern SPA frameworks offer their own componentization layers over functions. It's annoying and error-prone to keep these 2 layers, and lots of devs just abandon CSS for utility class mess. It's just a different choice.

I like the framing of organizational slack as immune system, instead of fat, and I emphatically agree with the author.

> ...“slack”: unbooked time that isn’t already spoken for. It’s always the first casualty, because it’s treated as a nicety... because it looks like waste...

> But that’s not what slack is. Slack is the only reason anything that comes up ever gets acted on. It’s what lets people take that signal that something is wrong, figure out why it’s happening, and then solve it... Without slack, the signal still arrives, everybody’s just too busy to pick up the phone or listen to the backlog of voicemails.

> [The AI playbook] gets this infuriatingly backwards: it treats slack as the thing AI lets you eliminate. If the tools make us faster, why do we still need buffer time?

> Getting rid of it at the time you’re flooding the system with machine-generated code you don’t fully understand removes the mechanism to absorb something we know will happen: it will surprise you.

> You didn’t cut fat. It was never fat. You cut your organization’s immune system.

Nope, splitting green wood is much more difficult than splitting dried logs, so I often cut a tree in the spring, stack the rounds, then split those rounds in the fall.

People overestimate how dry wood needs to be to burn correctly. Just have some ultra-dry kindling (seasoned for 2+ years) and you won't have any problems.

On the contrary, I know some folks who let all their wood dry too far, and it burned way too hot and ruined their stove (and almost burned their house down).

It’s an equation. If you have dry firewood, you need less of it at once. Some folks don’t understand that.

More water in the wood means less efficient combustion, more smoke and harsher smoke, which may irritate your neighbors downwind, or everyone around on still days.

Something every pit master learns along the way. People can tell you, you can read about it, but until you actually try using wood of different dryness, they are just words.

Not to mention a cooler stack temperature and the increased creosote condensation that brings.

Or that it's just pain heavier to lug into the house.

Yikes. I hope you got your chimney swept annually.

Seasoned firewood will burn cleaner, longer, and more efficiently.

> for reasons that nobody can articulate

This is only true if you ignore the role of America's greatest ally, in which case, the motivations become very obvious.

You assume the stone is released along a circular path, but that isn't quite right. When you release the stone you extend your arm forward, changing the arc of travel.

Ah the power of elliptical motion!

Nice blog! However, when I hear timeline, I want to see a timeline. This is an article ordered by time, which is much less interesting, and very long.

I'm a big fan of Dustin, I think the programming interface electric presents is very compelling, and the benefits are really impressive.

The killer feature, for me, of electric is that it makes the network _transparent_ to me, the programmer, and makes solving complex problems concise.

I see electric as a kind of emacs - an excellent operating system that lacks a decent front-end. Reactive programming is hard, and it really ought to be way more visual, and electric should give me those tools.

I find the frontend side of the electric somewhat obtuse, and it's easy to go off the rails, needing to understand the underlying reactive layer https://github.com/leonoel/missionary. Why should I target this bizarre DSL when there are millions of React, Angular, JQuery apps out there that work, and would be too onerous to adapt to this style of programming, and too annoying to carry a second framework in prod (even if you bundle an electric app in your existing one, the runtime can be very big by web standards). I would love to see a framework built on top of electric that interfaces with any existing frontend as easily as it does now for any backend (Datomic is just one target of many for the spreadsheet Dustin built). I may try to explore this idea myself.

But is the business of Hyperfiddle building instances of this new hypermedia, or the hypermedia system itself? I don't understand exactly what Dustin is selling me, at the end. Trying to sell this particular spreadsheet as a solution, or is it access to the runtime (electric 3)? How am I supposed to buy into a new programming model if I cannot use it?