I don't like them. I had an ELO touchscreen monitor at work before. I used it for a few days, but it caused me severe RSI issues in my shoulders, so I have no interest in a touchscreen laptop.
How is that different than any other car? Most modern cars lock their doors when you put them in gear, making breaking the glass the only option to get in during an emergency.
Most European brands I've tried only lock the cars when the speed exceeds 20 km/h.
In addition, most brands I'm aware of will unlock the car when the airbags are deployed (Kia/Hyundai, Renault, Volkswagen, Mercedes, BMW). Because these cars still have door handles, they can be opened by a bystander after a crash.
The issue is not that the doors are locked. The issue is that unlocked doors without door handles are still useless.
I use Rust and prefer Elixir over Gleam because Elixir has Lisp-style macros (Gleam doesn't), uses OTP directly, has the fantastic IEx REPL (Gleam has no counterpart), and has Phoenix, the best web framework I've ever used.
Comparing learning one language's concurrency model with learning another language's entire ecosystem is incommensurable.
Having learned both Go and Elixir, I found Elixir easier to learn and a lot more enjoyable to work with. I'm not alone in this opinion. According to Stack Overflow's 2025 "admired" languages, Elixir scored 65.9% compared with Go's 56.5%; Phoenix was the most admired web framework of 2025 at 79% and has held that spot for the past three years.
Erlang is dynamically typed, so many of the performance costs are similar to a JS runtime and always fully deciding typing ahead of time is an undecidable problem.
In practice, this is why many such languages have JIT's (unless targeting a subset or an type-information enhanced superset like TS), there was a seminal OOPSLA paper in 1995 by Agesen and Hölsze (who worked on the JVM Hotspot compiler) that compared JIT's to AOT compilation in practice (the Agesen CPA algorithm isn't perfect but it's pretty good for the time and others have probed that it's an undecidable problem).
That said, they also had a historically bad performance story due to misjudgments in development direction, the interpreter was default and they twice tried to make "HPC JIT's", ie.. complex JIT's that tried to be "perfect" and focused on numerical code gains, they'd be good for optimizing a matrix kernel, yet fairly useless or even negative on more common code patterns.
OTP 24,25 and 26 took learnings from the JS runtimes and also added compiler hints (since they already had binary precompiled modules).
What still saves the OTP runtime is that many basic operations that would suck without a good performance story is handled by built-in functions, so like Python most practical programs works well enough even if the runtime is behind.
Well to use Elixirs concurrency model you kinda have to use the rest of the langs ecosystem and learn a shit ton more compared to familiar feeling langs like Go. For Go I dont have to learn its execution model, some VM specifics and whatnot.
I'm a writer with what may be an unpopular opinion. I have a college degree in creative writing; I do technical writing and programming for a living, and I used to publish technical articles on Linux, Python, Elixir, and other topics—complete with em dashes—long before AI was a thing.
When confronting AI as a writer and deciding how I feel about it, I've landed on the conclusion that I've never accused mathematicians of churning out "calculator slop." A calculator is just a tool to make the complex work of mathematics easier. I now view AI as a calculator for writing, among other things. If the idea is good and the information is sound, I don't let the occasional awkward phrasing bother me. The truth is, most people can't write very well anyway, so if AI helps them communicate better—even if it is obviously generated by AI— more power to them.
- At least to me, AI writing (especially from Claude) is actively painful to read. It's worse than the average person who has flawed English (or just sucks at writing) because of how verbose and stylistically repetitive it is.
- AI writing is a strong probabilistic signal that what I'm reading is going to be crap in terms of the actual ideas and content too, even if it isn't the case 100% of the time. I just don't want to waste my time on that kind of thing.
When I use it for writing I usually have an idea or thesis to explore already. I typically write it out stream of consciousness. The core idea is mine.
I'm not using LLMs to think for me, instead just edit the writing so it has proper grammar and the run ons are reduced.
Even better for something that has data. You can pull PDFs with encoded tables and extract the data into consistent units to compare. You can have it write code to do it and cross check the outputs. You can do that in a minute. Would take hours without an LLM.
There's a bit of a difference to me between a calculator and writing.. Creativity. Words aren't just a vessel for ideas. If there were different calculation structures etc and space for people to develop their own styles and approaches to calculate things, then sure.
A person's writing style is developed through what they read. And if human's writing style becomes more homogeneous through AI then that's sort of a shame.
Somebody at $formerwork wrote a "skill" that takes a poorly/briefly specified Jira ticket, and "fleshes it out", purportedly so that when an agent picks up the ticket it has more context.
Really at that point, with head in hands, rather than try to explain the foolishness, I decided I was better out.
"Zoom and enhance" used to be a joke/meme, apparently it's now "engineering".
One line of idea spread into 30 is the worse way to convene an idea. Only people terrible at writing think using AI for prose is good. Most of the times, they also hate reading it, but love asking people to read their AI slop.
Calculating is not math. Calculators make mathematicians better mathematicians by allowing them to spend their time on actual mathematics.
Writing is thinking.
LLM writing makes authors worsr authors and even dumber people overall.
Also a calculator gives the same answer as a human (except for human errors where the calculator isa strict improvement). LLMs absolutely do not give the same output as humans.
Isn’t that expected, though? If 100 of us were to write a five page essay on the chaffing mechanics of an Underoos wedgie, the likelihood that any of them are the same is pretty small, isn’t it?
reply