ethmarks

But if you don't need to systematically examine every element one-by-one, lane layouts are pretty good. Sites like Pintrest use lane layouts because their content isn't meant to by systematically examined, but rather absorbed at a glance. If your content is meant to be systematically examined, using a lane layout would be a bad UX choice. But just because lane layouts can be misused doesn't mean they're a bad layout.

What?

The defining feature of masonry is that it supports mixed aspect ratios. That's its whole thing. If you aren't mixing landscape and portrait images, you shouldn't be using masonry layout.

You can't hold -5 apples any more than you can hold 5i apples, and yet we don't call numbers below zero "pretend numbers". I'm not arguing that imaginary numbers aren't imaginary, I'm arguing that calling them "imaginary" was dumb and creates a psychological barrier for students.

I think that calling numbers with a sqrt(-1) component "imaginary numbers" was a terrible decision. It makes a lot of people, especially students, think of them as pure-math nonsense without any practical applications. I think that "lateral numbers" would have been much better.

I love Zed editor's git UI for this. If you click on the left gutter of a hunk, it opens a little interface in the top right corner that lets you either Stage or Restore that hunk.

Exactly. When was the last time you heard HTML called "HyperText Markup Language"? When was the last time you heard CSS called "Cascading Style Sheets"? We should stop saying "JavaScript" and fully switch to JS.

And if each cell was a cubic micrometer (which is a side length 200-300 times smaller than a pixel on a typical screen and 50-100 times thinner than a human hair), it'd still stretch 3.7 kilometers, which is about the length of a commercial airport runway.

The average ChatGPT user isn't utilizing agentic workflows.

And even for the people that do, just because an LLM isn't absolutely state of the art doesn't invalidate it from being useful.

I noticed that the boing sound gets deeper and lower with smaller-magnitude boings. Is the boing audio generated procedurally/realistically in response to the physics of the boing, or is just playing a premade boing sound effect that's dynamically pitch shifted?

Building datacenters in the arctic also has the added benefit that sysadmins would have to take polar bear safety lessons, which would be pretty funny.

> I wonder if it is enforceable.

I can't imagine that it wouldn't be. If a company has explicit written permission from the copyright owner granting permission to use that copyright, then they can use it.

Also, it wouldn't be a special license. If you wanted to do a "For my friends everything, for my enemies the law" thing, you'd just set it as all rights reserved and add special note encouraging people to ask for permission to use it.

Plus, copyright enforcement typically goes in the other direction. It's not about who you can sue, it's about who you can't. Licenses are just a way of specifying who you cannot sue. If you want everybody to use your project but don't want to bother with a license, you can make it all rights reserved (the legal default) and just not sue anybody. You could sue them if you wanted to (which is why nobody would ever use your code: because of the risk that you change your mind and sue them), but nobody is forcing you to.

Fair point. Perhaps I'm just too pessimistic or narrow-minded, but I don't believe that LLMs will progress to that level of capability any time soon. If you think that they will, your view makes a great deal of sense. Agree to disagree.

Well, as a college student planning to start a CS program, I can tell you that it actually sounds fine to me.

And I think that teachers can adapt. A few weeks ago, my English professor assigned us an essay where we had to ask ChatGPT a question and analyze its response and check its sources. I could imagine something similar in a programming course. "Ask ChatGPT to write code to this spec, then iterate on its output and fix its errors" would teach students some of the skills to use LLMs for coding.

> My recommendation would be to encourage the tutor to ask the student how they use the LLM and to school them in effective use strategies

It's obviously not quite the same as programming, but my English professor assigned an essay a few weeks ago where we had to ask ChatGPT a question and then analyze its response, check its sources, and try to spot hallucinations. It was worth about 5% of our overall grade. I thought that it was a fascinating exercise in teaching responsible LLM use.

You can still argue that LLMs won't replace human programmers without downplaying their capabilities. Modern SOTA LLMs can often produce genuinely impressive code. Full stop. I don't personally believe that LLMs are good enough to replace human developers, but claiming that LLMs are only capable of writing bad code is ridiculous and easily falsifiable.

> most students will learn a lot less than say 5 years ago while the top 5% or so will learn a lot more

If we assume that AI will automate many/most programming jobs (which is highly debatable and I don't believe is true, but just for the sake of argument), isn't this a good outcome? If most parts of programming are automatable and only the really tricky parts need human programmers, wouldn't it be convenient if there are fewer human programmers but the ones that do exist are really skilled?

Does your transportation system also have a risk of exploding catastrophically mid-flight? 'cause otherwise no deal. /s

There is some medium in low Earth orbit. Not all vacuums are created equal. However, LEO vacuum is still very, very sparse compared to the air and water we use for cooling systems.

The main way that heat dissipates from space stations and satellites is through thermal radiation: https://en.wikipedia.org/wiki/Thermal_radiation.

There's no air and negligible thermal medium to convect heat away. The only way heat leaves is through convection from the extremely sparse atmosphere in low Earth orbit (less than a single atom per cubic millimeter) and through thermal radiation. Both of which are much, much slower than convection with water or air.

Space stations need enormous radiator panels to dissipate the heat from the onboard computers and the body heat of a few humans. Cooling an entire data center would require utterly colossal radiator panels.

Not having used anything except for Firefox, I don't have any experience with migrating to different browsers. However, my understanding is that Chrome shows a little pop-up that lets you import from previous browsers rather than relying on the user to do a data export. Correct me if I'm wrong about this.

I suspect that Claude couldn't make an "import from ChatGPT" button because OpenAI would make it difficult, so they'd have to rely on user initiative and technical capability (exporting to JSON and importing from JSON is enough technical friction that the average user won't bother).

I think it matters to more than you might think. A significant portion of the non-technical ChatGPT userbase get really attached to the model flavor.

The GPT-4o controversy is a good example. People got attached to 4o's emotional and enthusiastic response style. When GPT-5--which was much more terse and practical--rolled out, people got really upset because they were treating ChatGPT as a confident and friend, and were upset when it's personality changed.

In my experience, Gemini and Claude are much more helpful and terse than ChatGPT with less conversational padding. I can imagine that the people who value that conversational padding would have a similar reaction to Gemini or Claude as they did to GPT-5.

Fun fact: that's called a generic trademark

https://en.wikipedia.org/wiki/Generic_trademark

> Note also that the triangle has infinite perimeter but finite area.

How common is this property in geometry? I know that fractals like the Koch Snowflake also have infinite perimeter over finite area, but I don't know what else does.

And the mushrooms would be powered by cosmic radiation...

Perhaps you misunderstood my comment. I'm not advocating for vibe-coded AI-generated PRs, and I do think that blanket banning them is pretty reasonable for the reasons you stated.

However, I don't think that banning all AI-generated code is reasonable. Having an LLM generate a couple of functions or a bit of boilerplate in an otherwise manually coded PR should not invalidate it from being accepted if it's helpful.

Gemini 3 is a 10 trillion parameter model?

> A blanket ban on LLM-generated code is at least arguably a reasonable policy.

No, I don't think it is. There's more nuance to this debate than either "we're banning all LLM code" or "all of our features are vibe coded".

A blanket ban on unreviewed LLM code is a perfectly reasonable way to mitigate mass-produced slop PRs, but it is not reasonable to ban all code generated by an LLM. Not only is it unenforceable, but it's also counterproductive for people who genuinely get value out of it. As long as the author reviews the code carefully before opening a PR and can be held responsible, there's no problem.

Does it really make sense to describe Athas as an editor with "agentic AI editing" if, by their own admission in the roadmap, they haven't even begun implementing it? Seems like they should be using the future tense or listing it as a planned feature.

I beg to differ. Tactical use of a scientific or graphing calculator can absolutely replace large parts of the thinking process. If you're testing for the ability to solve differential equations, a powerful enough calculator can trivialize it, so they aren't allowed in calculus exams. A 10-digit calculator cannot trivialize calculus, so they are allowed. That's the distinction. LLMs operate at the maximum level of "helpfulness" and there's no good way to dial them back.

As someone who had a college English assignment due literally just yesterday, I think that "the vast majority" is an overstatement. There are absolutely students in my class who cheat with AI (one of them confessed to it and got a metaphorical slap on the wrist with a 15 point deduction and the opportunity to redo the assignments, which doesn't seem fair but whatever), but the majority of my classmates were actively discussing and working on their essays in class.

Whatever solution we implement in response to AI, it must avoid hurting the students who genuinely want to learn and do honest work. Treating AI detection tools as infallible oracles is a terrible idea because of the staggering number of false reports. The solution many people have proposed in this thread, short one-on-one sessions with the instructor, seems like a great way to check if students can engage with and defend the work they turned in.