dgreensp

Where do you see information about the efficiency gains over AV1?

This reveals a staggering level of incompetence, if that’s really all it is, and lack of transparency.

They don’t have ANY product-level quality tests that picked this up? Many users did their own tests and published them. It’s not hard. And these users’ complaints were initially dismissed.

I don’t think the high vs medium change is really on par with the others. That’s a setting you change in the UI, and depending on what you are doing, both effort levels are pretty capable, they just operate a bit differently. Unless I’m missing something and they are saying they were doing some kind of routing behind the scenes.

If they are constantly pushing major changes to the prompts and workings of the tool, without communicating about it, and without testing, it’s likely there are other bugs and quality-degrading changes beyond the ones in this article, which would make a lot of sense.

I agree with all this.

I came to Deno because I needed a break from Node/NPM. I don’t agree with all of Node’s decisions (particularly the ES module debacle), but Node/NPM have improved over the years.

A big problem with JSR was no private packages. All your code has to be open source. But JSR is the only way to get constraint solving in Deno, besides using NPM.

I find the irreverent tone refreshing, personally.

As a founder who built all my prototypes and side projects on Deno for two years, I personally think Deno’s execution was just horrible, and avoidably so. Head-scratchingly, bafflingly bad decision-making.

I was the first engineering hire at Meteor (2012-2016), and we made the mistake of thinking we could reinvent the whole app development ecosystem, and make money at it, so I have the benefit of that experience, but it is not really rocket science or some insight that I wouldn’t expect Ryan Dahl and team to have, in the 2020s.

They were stretched thin with too many projects, which they were always neglecting or rewriting, without a solid business case. They coupled together runtime, framework, linting, docs, hosting, and packaging, with almost all of these components being inferior to the usual tools. The package system became an absolute nightmare.

If the goal was to eventually replace Node and NPM with something where TypeScript was first-class, there was better security, etc, they could have done a classic “embrace and extend.”

This piece starts off making it sound like the computer is pretty much doing all the work, while the human maybe weighs in on a matter of taste once in a while, if they like, but by the end, the list of what the LLM can actually do is really short. Implementing a sorting algorithm for you, perhaps, but not necessarily one without “egregious flaws,” and really you should still use a library for that. Replacing high-quality libraries of mature software, that have tests, etc, is obviously one of the poorer uses of vibe-slop coding.

It comes down to “adding code” that attempts to, or seems to, achieve something.

I always interpreted cathedral vs bazaar as being about the architecture of large things. Do you build to a master plan? Or does everyone do whatever they want? (Within some kind of framework, of course.) Like the cathedral of the Java SDKs vs the flea market of NPM.

This author seems to have some kind of attitude about organization in general—anything with people and process, that happens to exist around some project, that might require at least a small commitment to be a part of. Like complaining that a flea market has a form to sign.

The ability for people to functionally collaborate, with some kind of structure, is the key thing that enables building large things together.

By that logic, Microsoft’s brand means nothing when OpenOffice is free.

A curly brace is multiple tokens? Even in models trained to read and write code? Even if true, I’m not sure how much that matters, but if it does, it can be fixed.

Imagine saying existing human languages like English are “inefficient” for LLMs so we need to invent a new language. The whole thing LLMs are good at is producing output that resembles their training data, right?

I ran into this, and there was a bizarre fix—I think having Adobe apps open in the background caused it, or something.

Upvoted because educational, despite the AI-ness and clickbait.

I’ve worked at orgs that used Postgres in production, but I’ve never been the one responsible for tuning/maintenance. I never knew that Postgres doesn’t merge pages or have a minimum page occupancy. I would have thought it’s not technically a B-tree if it doesn’t.

This is some of the best writing I've read in a while, and truly fascinating.

Radix sort is not a comparison-based sort and is not O(n log n).

No, because radix sort is not a comparison-based sort and is not O(n log n).

This semi-explains why I have started to notice (sadly) serious bugs in TextEdit, not just scrolling but editing/corruption.

Never place rich UI elements within a table, list, or other markdown element.

Place rich UI elements within tables, lists, or other markdown elements when appropriate.

I think people are missing the fact that Wired has been about “vibes” since the beginning.

Wired vs. tired is literally about what’s “cool.” That’s it. It has never been rigorous about anything.

Googlebot respects robots.txt. And Google doesn't use the fetched data from users of Chrome to supplement their search index (as a2128 is speculating that Perplexity might do when they fetch pages on the user's behalf).

I met someone with SDAM who described it in a more striking way.

He said he doesn’t have any “first-person” memories. Most people, even if they forget most things they do day-to-day, and don’t have great “indices” of their memories, can think back on certain times where they remember being there, doing a thing, probably with some visuals associated with that, which can be played back. This person said that none of the things they remember that they did, they can remember personally doing, if that makes sense.

For me, I have scattered samples, like once in a while throughout my life, my brain sort of takes a snapshot and forms a memory. I can imagine myself back in each place I lived, for example, or each place I worked, or graduating from college, or walking on the beach.

I don’t think the parent was saying that everyone’s intentions were pure until recently, but rather that naked greed wasn’t cool before, but now it is.

The Internet has changed a lot over the decades, and it did used to be different, with the differences depending on how many years you go back.

Exactly. Trump is practically illiterate and is being handed things to sign. His original ideas that were pushed back on by his advisors in his first term were a different sort of idea, things like, "Why can't we just force that country to do what we want, we're the USA, we're the most powerful, we could just bomb them."

Yeah, it's a funny argument because while Apple has certainly put a lot of money into WebKit and JavaScriptCore over the years in absolute terms, they already don't prioritize Safari or treat web technologies as an alternative to native app development.

That’s my understanding as well. Duplicating the bytes of a file when you don’t have the rights to the content is technically infringement and grounds for an infringement claim, and then you have to explain in court why it’s “fair use.”

When casually-inflicted trauma and indifference to using you effectively, let alone your needs as a human being, are a constant, while other responsibilities come and go and are taken less seriously, it feels like the core of the job.

You’re absolutely right. If there are corrections, it’s only because people have gone a little crazy. People sometimes get enamored with the idea of one country or company dominating in a winner-takes-all situation. OpenAI has been trying to spin narratives in which the only rational move is for everyone to invest all of their money in OpenAI, immediately.

I don’t think the Nvidia CEO is losing any sleep. He doesn’t have a planet-sized ego, like some. They are still selling pickaxes during a gold rush, as the saying goes. And the stock is still up 100% in the past year. CEOs don’t control the stock price (much as some try), anyway, the market does. Also, the market is insane. You also can’t control the competition, or what new innovations come along. You just have to run a good business, which they are doing.

Innovation is good for all players. Reality-checks are good. Nothing here is unexpected. There are lots of smart researchers in the world. Software innovations that make better use of hardware are expected. People just like drama.

Like if Tim does a better job at a skateboard trick that has always been Tom’s thing, people want to be that kid who is the one to say, “Oh, snap!!” and won’t stop talking about it at school, because they were there. And how it’s so mind-blowing and previously inconceivable.

I had to use Docker at a job or two, I think around 2018. I hated it.

One class of issue: It made interacting with the file system slower, sometimes by orders of magnitude. Stuff like watching files, or statting a large number files, didn’t have the same performance characteristics. So you have a situation where you (probably) already have too many components that are too complicated or poorly understood to install them all on a developer’s machine, but they work on this exact machine snapshot, but now you have to figure out what process dared to stat a few thousand files.

Docker was also just always… there. In the menu bar. Doing stuff. Running system-wide. Updating itself, constantly. Like it’s Steam or Battle.net (which for some reason downloads updates to Warcraft III, an old game, multiple times a day on my kids’ PC, and sometimes breaks and you can’t play the game; this is the level of enshittification we are at).

The command-line experience… similar to git (that is, poor). There’s an underlying conceptual model that’s sort of half abstracted away by the tools and hard to find a good explanation of.

Developer tools like this have a tax: You spend at least half a day a week Googling for issues with them, forever. Same with NPM. All it takes is five such tools in your stack and every weekday morning is gone. And that’s disregarding the fact that you were probably in the middle of actually trying to get something done.

Thanks for posting it!

Remotion does look really cool for in-browser video editing. People should note that you can’t make free-to-use software with Remotion. It is a paid product, and you can download and run the code for free in very limited circumstances that let you experiment with it. You can’t use the output of your own software built on Remotion for any for-profit purpose, nor can anyone else, without a license.

While it’s just one small section of the article, I notice there is the sort of industry-standard erasure of the entire concept of good vs bad UI or engineering, in the “Web apps are bad” section.

A “good” app, in this article, is one that is “successful, versatile, [or] capable.” Any McDonald’s kiosk is an example of a “good” app, because it’s so “successful.” If you think a much-used app has UI that isn’t “good” and would be improved by using native technologies, well, first of all, “the market” is the arbiter of what’s good, not you, and second of all, even if you are right, your criticism lacks empathy for the “requirements and constraints” developers face, you are ignorant about the fact that things have tradeoffs, or you are some perfectionist who thinks everything needs to be a work of art.

Software engineers are so preemptively defensive about quality.

I can’t say it’s not possible there is symbolism such as “the scarecrow represents farmers,” so when the scarecrow is elevated to the position of Ruler of the Emerald City, that symbolizes something.

At the literal level, though, the government in these books is an imperialist monarchy. We read all the books to my ten-year-old, and he had criticisms of this without us saying anything, pointing out it didn’t seem right for entire races/cultures that had never even heard of Princess Ozma to be considered her subjects and subject to her rule.

By this article’s own standards, I can’t really fault it for jumping to half-assed conclusions, as long as a casual reader might not notice or care. I’m not being glib. Devaluing quality is a choice, and if it’s fashionable to say it’s fine that everything’s being enshittified (because then we can feel smart, rather than disappointed), I’d rather it wasn’t.

The flaw is in extrapolating from the fact that different situations require different levels of craftsmanship or attention to aesthetics, standards, etc—and something fancier or more meticulously or expertly made, like a meal from a Michelin star restaurant, might not even be more enjoyed by the particular person who experiences it—to make the point that the things only professionals and those versed in their field know to do, which a layperson might not even notice, don’t really matter.

The reality is, whether we enjoy the products and services and experiences that come our way or not is largely due to design decisions beyond what we can consciously attend to and appreciate. It’s all those little details.

There is a limit to how low-quality content can be, and it’s actively being explored with the help of AI.

One of the effects of capitalism in the US used to be that companies were trying to make higher-quality things for less, because of competition. The strategy now, by the big players, is to just sort of swallow everything up, and then as quality gets lower and prices get higher, consumers feel compelled to just roll with it, because the shift is happening everywhere all at once. In corporations’ ideal world, there is no social contract, no real market forces, the consumer has no leverage, the corporation puts in as little effort as possible, and the consumer pays as much as possible, and still buys the thing, because what else are they going to do.

You can put ice cream in front of someone and they’ll eat it, so let’s just all eat ice cream for breakfast, lunch, and dinner.

I found this article a bit disappointing.

The link at the bottom doesn’t work.

The grids shown do not follow the well-known rules of (American) crosswords: every square is part of two words of three or more letters each.

Coming up with a pattern of black squares, and writing good clues, are two parts of making a crossword puzzle that are IMO fun and benefit from a human touch, and are not overly difficult. There are also databases of past clues used in crossword puzzles (eg every NY Times clue ever, and various crossword dictionaries) for reference and possible training. If you don’t care about originality (or copyright) and want quality clues, you can just pull clues from these. If you do care about all those things, you can surface the list of clues used in the past to the human constructor and let them write the final clue. Or you can try to perfect LLM clue-writing. In my experience, LLMs are terrible at clues. Like sometimes if I try to give it feedback about a clue, it will just work the feedback into the clue… it’s a little hard to describe without an example, but basically it doesn’t seem to understand the requirements of a clue and the process of a solver looking at a clue and trying to come up with an answer.

Coming up with an interlocking set of fun, high-quality words and phrases is the hard part. I agree that LLM wordlist curation is a great idea, and I started playing around with that once.

Beyond that, I don’t think LLMs can help with grid construction, which is a more classic combinatorial problem.