> are trained on what we told them we do. They don’t “think” at all. They’re a mixmaster of other people’s ideas, cleverly packaged in a way that we perceive as natural.

Sometimes I wonder how this is different from most of my education. Or my creativity, mixing ideas together to see if they still make sense with other things I have been told.

"what is 16929481231+22312333222?" is an easy way to test this claim. Pick large enough numbers and there's no way all the sums of that size would fit into the dataset (you don't need to stick to + either, but it's the simplest thing that works)

But if you were to ask that same question to a human with no specific math training there are exceedingly low odds they would get the right answer.

We spend hours and hours over reinforced over years to have humans that can do it.

Indeed! So, where are you going with this?

For my contribution to the conversation: Earlier/cheaper models can't do it either, they make mistakes, they need a calculator/jupiter kernel/what have you. 'Medium' models will put the numbers underneath each other and do it 'properly' in a table, checking themselves after. Claude Opus 4.6 (the current rolls royce today) just says the answer in one go sometimes (it's a monster). But all of them end up spending many seconds and thousands of tokens on a task that takes a calculator or an ALU fractions of a second.

Yeah, that’s my main beef with this article. There is not even an attempt, just waving hands and saying they are not.

Decades and decades of “turing test” talk until they can pass it.

Mind: Turing test doesn't test for actual thinking though, just functional indistinguishability. Turing sidestepped the problem way back when.

That seems a little harsh. GUI tools can give us a more vibrant and useful interface.

But, I think the main problem is that although there have been many attempts we have not gotten to a standard way to compose different GUI tools easily or repeat actions.

> but GUI based IDEs are generally useful and easier to use out of the box for development.

This is true, they are much better for discovery and affordance, but as you progress with your tooling and tool usage there is a much higher ceiling on your productivity with other tools and their composability. In my opinion, not putting effort into learning tools ultimately holds a lot of people back from their potential.

I use both and mostly agree, but for me I don’t think the ROI for learning terminal based tooling is there.

They make some parts of text manipulation faster, but those parts of text manipulation take up less than 1% of my time spent working.

Things like debugging, which take up a large portion of my time, are not so nice in terminal based environments

Yes, for things like Node, I do use tools like the chrome dev tools for debugging and such.

But find a terminal first approach leads me to other tools like curl and jq usage as I go. I see coworkers using a ton of time trying to repetitively execute the code to see those spots in really inefficient ways. And end up completely lost when they could be using tools like git bisect.

Or another good example devops type support is if one web server out of many seems to be misbehaving, I can use aws command line to get internal ips behind the lb to curl to grep and find it in minutes after others have tried for hours. It makes it second nature if your mind goes there first.

I work 99% in a terminal and fire up a JetBrains IDE when I need to do deep debugging. It’s so rare for me though that it’s worth more for me to get good at the terminal stuff. I’m sure this depends heavily on the type of work being done, game dev for example really needs a good debugger. That being said, gdb and others have perfectly fine text mode interfaces, albeit with a steeper learning curve.

As always, the “best” tool is the one your most familiar with that gets the job done. Text vs GUI doesn’t really matter at the middle of the bell curve.

> there is a much higher ceiling on your productivity with other tools and their composability

What exactly is the "ceiling" for GUI based IDEs?

Composition. I don’t think there’s any GUI that can be used for the git email workflow.

Versatility. In most TUI editors, running an external commands is easy, including piping a part of the text and receiving its output. As this is a fundamental unix principle, there’s basically no barrier between what you’re editing and other data on your system. In GUI, everything is its own silo.

Presentation. Code is not linear, but most gui forces use to use one window to see your text files. And when they allow splitting it’s cumbersome to use. Vim and Emacs has a easier way to divide you screen so that the relevant information can be presented at once. And there’s terminal multiplexers for simpler editors.

You’d also probably be surprised about how subjective and unevenly applied the law is… by design, to allow appropriate outcomes and discretion.

Edit: Consider the following words included in law.

“reasonable” “reckless” “due care”

The last line was really good. I may have differing opinions on some specifics but I loved this line: "The personal computer movement was about empowerment not dependency"

Long term contracts are routinely broken in bankruptcy without some sort of surety bond if things go sideways. This leaves localities footing the bill on maintenance if things do not turn out.

> Angular, React, Vue, whatever - as if solving business problems just became too boring for software developers, so we decided to spend our cycles on the new hotness at every turn.

They often do solve business problems around responsive design, security and ux.

Currently working maintenance with one foot in a real legacy system and the other foot in modern systems the difference is immense.

It is a bit different than what we are discussing, but intent plays a huge role in Western justice. The same physical action can lead to vastly different outcomes.

A high profile interesting example of this is the assassination attempt on Brett Kavanaugh. If you look at the details none of the actions would have been an attempt if not for the intent.

It is an interesting thought experiment as to how many actions you have to take for a crime that you don’t commit to be charged as an attempt or more broadly as conspiracy and at what point people are allowed to change their mind. We see this in terrorism cases pretty frequently.

> The same physical action can lead to vastly different outcomes.

Well, yeah, that’s kind of obvious once you realize that tools can be used for multiple purposes. A hammer can be used to pound both nails (legal) and smash a person’s head in (not).

But the notion that “thought crimes” where people are being punished merely for their feelings and where no act in furtherance of the outcome has taken place is just baloney. At least in the West.