This year's Time Person of the Year makes perfect sense if you're familiar with any of the past Time People of the Year. Hitler wasn't even an outlier.

Somebody needs to write at length of the unhinged nonsense of calling academics critical of a monstrous technology, esp ones with direct expertise in formalism & implementation details: technopessimist. Are climate experts called pessimists for believing in & fighting against the climate crisis? 1/2

As was true with the translators I heard from in an earlier edition, copywriters never felt that they were being replaced because the AI's work was better. It was almost always worse. It was simply cheaper and faster, and deemed "good enough" by clients.

I was reading about ad firms using AI, and they rationalize the ads unpopularity with "oh, it got your attention" which strikes me hypocritical, nobody (in ad industry or lay commentary) looked at the Dylan Mulvaney mess and justified using her for a brand deal with "oh, got your attention"

CSS text-decoration-inset landed in Firefox 146! Here's how it works:

people taking the high horse on the react vulns really piss me off.. your tribalist low-effort dunks just show that you don't understand how pushing the envelope always comes with risks for new exploits! much love from the svelte community to all the react folks working overtime to keep people save

This pattern shows up across the industry, not just in JavaScript. For example, after Log4Shell, additional CVEs were reported as the community examined the original fix. Additional disclosures can be frustrating, but they are generally a sign of a healthy response cycle.

It’s common for critical CVEs to uncover follow‑up vulnerabilities. When a critical vulnerability is disclosed, researchers scrutinize adjacent code paths looking for variant exploit techniques to test whether the initial mitigation can be bypassed.

It's heartbreaking to read of such an aimless, anchorless, design process. www.therepository.email/wordpress-ai...

It's wild to me that lateinit in Kotlin was created before it was widely used in Android. It seems like a feature laser-targeted to Fragments. I can understand why people hate it

People say AI will replace all the jobs shortly and the quality of output is shockingly amazing. Until you ask “what about your job?” Oh no it could never do that. Too nuanced and complex. I tried it last week and AI made a decently okay first draft that was almost useful.

It took me a while to understand it, but this book has been helpful for learning CSS. I use the patterns as the C in CUBE CSS

Lol, I think Denmark is also worried about this. Honestly, we all should be.

React2Shell is not an exploit due to some problem unique to the design of server components. If you're using JavaScript at all, you cannot dunk on them without your stones breaking your glass house.

It’s not just a UI/UX issue. Instead of understanding what people’s needs are and helping them bridge over the gap, or understanding what people’s actual problems are and designing solutions for them, they develop *products* and then go about coercing, forcing, or confounding people in to using them

Not sure if I agree but I can't resist reposting a take this spicy

A big part of the schadenfreude with AI is that if you've had even a modest amount of exposure to it, you can tell that the people that fall hardest for it are those that respond to glazing the hardest. Telling people they are a genius is powerful medicine

First issue of Inclusive Android Apps is live! 🚀 This one's about how Row layouts break when users increase text size - and why FlowRow is a good fix. Read it: buttondown.com/inclusive-an... #AndroidDev #AndroidAccessibility #a11y #InclusiveAndroidApps

Students, staff, faculty turning to ChatGPT? That means they are being asked to do things they don't have the time, resources, support, etc to do.

While not ideal, it would still be an improvement if I could solve technical interview problems the way I solve Advent of Code.

This is what happens to software companies who go all-in on coding with LLMs. Coding isn’t the one “good use case” for LLMs but a dozen Y2K crises in the making. And I’m saying this as a software developer who knows more about how LLMs work than most in the field.

I feel the clients now using GenAI to do stuff and going "good enough" when they used to nitpick at every last detail is due to them finding out all those nit picky changes take time and work they cant be bothered to put in.

While I wouldn’t say they are harder to learn than the rules of hooks, Svelte runes are something unintuitive that needs to be learned to be used properly. I don't think you can have as powerful reactivity guarantees without something as complicated as hooks or runes.

In another installment of the series Surely This Can't Be The Right Way To Do This In Android: how to have a per page top bar that doesn't flash on navigation change. What I have so far:

The thing that makes React slow isn't the Virtual DOM. What makes React slow is that it is flexible enough to let you shoot yourself in the foot with avoidable performance mistakes. Relatedly, the thing that makes Svelte faster than React is that it's a highly opinionated framework

I say I'm anti-genAI but actually it's a really cool, novel piece of technology. It just doesn't have any commercial applications. I wish we could focus on what it's useful for, but unfortunately we can't because it's mostly being used for exploitation and crime.

When you write the code yourself, every subsequent pass improves the quality. When you generate it using AI, subsequent passes can wipe out previous progress or touch completely unrelated things

The fact of the matter is that knowing CSS properly does not factor into hiring web developers. Unless you are working alone, the styles you’ll be working with will be mostly written by people who don’t understand what they’re doing.

Need me a sugar daddy to buy Josh Comeau’s CSS for JS devs course css-for-js.dev

When people talk about genAI improving their productivity, they mean that the first pass is generated in seconds rather than handwritten in 10-15 minutes. They're not counting the amount of time it takes to reprompt for something passable

I think I’m finally starting to understand the principle behind using CSS to hint to the browser the direction you want layouts to go to. The things promoted by Every Layout and CUBE CSS

So it turns out that if I don’t learn something new related to programming at least once a month, I start to chew at the bars of my enclosure

What has changed in the AI market that would induce a change in the opinions of people here?

I've worked on a dark/light themed site and went all-in using `light-dark()`. So far, I've thought that I can use it as a `prefers-color-scheme` replacement... I learned that there are some differences in how they treat `color-scheme`. 👇 www.stefanjudis.com/today-i-lear...

🚨 A new wave of the Shai-Hulud supply chain attack has hit npm, impacting packages across widely used projects from AsyncAPI, ENS, Postman, PostHog, and Zapier. Attackers added a malicious preinstall script following account compromise. The investigation is ongoing: socket.dev/blog/shai-hu...

the big problem with that Adversarial Poetry versus AI paper is that they did half the work with chatbots and not humans. they translated prompts into poems using chatbots. then they judged how well the poems jailbroke the chatbots using other chatbots to do the judging?! >>

Sticking “JavaScript gotcha" in the drawer next to “object-oriented programming" and keeping it locked until you learn a functional programming language

This is how I see stuff like LLM-based autocaptions. They are inaccurate, and disabled people deserve better. But they're better than what disabled people were getting before. It's not truly inclusive

I probably shouldn't stick my nose in this, but I keep seeing it and I want to say my piece. On Bluesky, I see the anti-genAI folks saying only AI bros say AI “slurs" are racist. On Tumblr, I see the anti-genAI folks saying that AI “slurs" are always racist. Both are kinda wrong in their own ways.

I was at a conference on Weds about use of AI in public services and noticed how AI as sticky tape over broken systems is becoming the norm, designing or implementing better systems is becoming more and more of a luxury.

every discussion i see on here that bears on banning gen ai in some fashion includes a code-is-speech guy who is like “good luck when anybody can run a model at home lol” and im like go for it. all 7 of you. ai would be relegated to a hobby on the level of say 3d printing

Another study that shows the dangers on genAI concluding that we should ignore the results and use genAI anyway. Starting to feel like there's some sort of mandate or censorship requiring people add these contradictory conclusions to their research

"Should AI have human rights?" is a question often asked by people who aren't clear whether humans should have human rights.

Museums should stop displaying their collections as if patrons will view them, and start designing them for robbers who will be stealing them.

I don't think we should purity test. In the interest of the greater good, we should be prepared to endure unpleasantness and get over our personal distaste. In unrelated news, I've been fired from my job as drinking water treatment plant supervisor for "negligence" and "hundreds of people died"

So I'm reviewing a PR from a colleague that was written for like 95% by AI, and while it might have made his job easier, it's making mine a lot harder because it's riddled with mistakes and issues that I have to now fix, and I feel like I have to thoroughly check every single line of code twice

"An LLM helped me with some task" is not a good argument for these things increasing productivity. Did you also factor in all the cases where it made you slower? Where the output created extra work? And how much of your work/activity is that task?