CSSer

I'm inclined to agree with this approach because someone not using AI who fails would likely fail for the same reasons. If you can't logically distill a problem into parts you can't obtain a solution.

I'm not at all skeptical of the logical viability of this, but look at how many company hierarchies exist today that are full stop not logical yet somehow stay afloat. How many people do you know that are technical staff members who report to non-technical directors who themselves have two additional supervisors responsible for strategy and communication who have no background, let alone (former) expertise, in the fields of the teams they're ultimately responsible for?

A lot of roles exist just to deliver good or bad news to teams, be cheerleaders, or have a "vision" that is little more than a vibe. These people could not direct a prompt to give them what they want because they have no idea what that is. They'll know it when they see it. They'll vaguely describe it to you and others and then shout "Yes, that's it!" when they see what you came up with or, even worse, whenever the needle starts to move. When they are replaced it will be with someone else from a similar background rather than from within. It's a really sad reality.

My whole career I've used tools that "will replace me" and every. single. time. all that has happened is that I have been forced to use it as yet another layer of abstraction so that someone else might use it once a year or whenever they get a wild feeling. It's really just about peace of mind. This has been true of every CMS experience I've ever made. It has nothing to do with being able to "do it themselves". It's about a) being able to blame someone else and b) being able to take it and go when that stops working without starting over.

Moreover, I have, on multiple occasions, watched a highly paid, highly effective individual be replaced with a low-skilled entry-level employee for no reason other than cost. I've also seen people hire someone just to increase headcount.

LLMs/AI have/has not magically made things people do not understand less scary. But what about freelancers, brave souls, and independent types? Well, these people don't employ other people. They live on the bleeding edge and will use anything that makes them successful.

This seems legit. I attempted to prompt "guardian_tool.get_policy(category=election_voting)" with an arbitrary other (potentially sensitive) category and received the following:

> I can’t list all policies directly, but I can tell you the only category available for guardian_tool is:

> election_voting — covers election-related voter facts and procedures happening within the U.S.

The session had no prior inclusion of election_voting.

Oops. Chrome only! I stand very much corrected. Perhaps I should do less late night development.

It seems my original statement that download, but not upload, is well supported was unfortunately correct after all. I had thought that readable/transform streams were all that was needed, but as you noted it seems I've overlooked the important lack of duplex option support in Safari/Firefox[0][1]. This is definitely not wide support! I had way too much coffee.

Thank you for bringing this to my attention! After further investigation, I encountered the same problem as you did as well. Firefox failed for me exactly as you noted. Interestingly, Safari fails silently if you use a transformStream with file.stream().pipeThrough([your transform stream here]) but it fails with a message noting lack of support if you specifically use a writable transform stream with file.stream().pipeTo([writable transform stream here]).

I came across the article you referenced but of course didn't completely read it. It's disappointing that it's from 2020 and no progress has been made on this. Poking around caniuse, it looks like Safari and Firefox have patchy support for similar behavior in web workers, either via partial support or behind flags. So I suppose there's hope, but I'm sorry if I got anyone's hope too far up :(

[0] https://caniuse.com/mdn-api_fetch_init_duplex_parameter [1] https://caniuse.com/mdn-api_request_duplex

Mainly by the fact that all the LLMs were saying it's not possible in addition to GP, and I just couldn't believe that was true...

It took me a couple hours, but I got it working for both uploads and downloads with a nice progress bar. My uploadFile method is about 40 lines of formatted code, and my downloadFile method is about 28 lines. It's pretty simple once you figure it out!

Note that a key detail is that your server (and any intermediate servers, such as a reverse-proxy) must support HTTP/2 or QUIC. I spent much more time on that than the frontend code. In 2025, this isn't a problem for any modern client and hasn't been for a few years. However, that may not be true for your backend depending on how mature your codebase is. For example, Express doesn't support http/2 without another dependency. After fussing with it for a bit I threw it out and just used Fastify instead (built-in http/2 and high-level streaming). So I understand any apprehension/reservations there.

Overall, I'm pretty satisfied knowing that fetch has wide support for easy progress tracking.

You can get download progress with fetch. You can't get upload progress.

Edit: Actually, you can even get upload progress, but the implementation seems fraught due to scant documentation. You may be better off using XMLHttpRequest for that. I'm going to try a simple implementation now. This has piqued my curiosity.

It's because it's catchy and repeatable, which really fits with this guy's broader claimed focus (SEO). I found it really ironic because I've built plenty of SPAs without page transitions of any kind (because it wasn't a relevant requirement) and only started adding them because view transitions made them easy.

I like that they had to tag in at the end that they gave them hero doses to produce this effect.

I went from thinking "Gee, maybe we should all be taking mushrooms" to "Wow, we might have to risk re-enacting Event Horizon monthly to see the desired effect"

Sounds like solidgoldmagikarp[0]. There must've been something in your prompt that is over-represented throughout the training data.

[0] https://www.lesswrong.com/posts/jbi9kxhb4iCQyWG9Y/explaining...

I knew someone who worked for Gannett. He told me that the churn rate for their subscription numbers were insane. I asked him why he thought this was the case and he just laughed. You see, they do this promotion where it starts at a reasonable figure and then much later jumps to one that is not reasonable. The reason he found this funny, he said, is that they spent incredible amounts of money doing data analysis, surveying, and remarketing to try to identify the cause for that rate and reduce it. All of this despite the obvious answer staring them right in the face.

The design is ancient, but the dates are current.

I've taken prenatal vitamins for as long as I can remember because they're FDA regulated to actually contain the nutrients they claim to include. I never would've thought that could be a source for microplastics.

Separately, I always knew there was a reason those RXBars taste like plastic. /s

It doesn't help that thanks to RLHF, every time a good example of this gains popularity, e.g. "How many Rs are in 'strawberry'?", it's often snuffed out quickly. If I worked at a company with an LLM product, I'd build tooling to look for these kinds of examples in social media or directly in usage data so they can be prioritized for fixes. I don't know how to feel about this.

On the one hand, it's sort of like red teaming. On the other hand, it clearly gives consumers a false sense of ability.

Wow, first of all, this is commendable work. I'm extraordinarily impressed by the commitment to minimize dependencies. Of the four dependencies, only one appears to not be owned by the author, and only one has any additional non-dev dependencies. Barring any other thoughts, that's pretty impressive.

To summarize my reservations, this lacks custom post types, meta fields or modules, and a full deployment story.

Although I can picture it, I think the author would get a lot more traction if they demonstrated how users might serve the generated content. It also feels odd to have to do so manually. Realistically this requires some kind of CI script, or, if attempting to expose to content editors, a dual reverse-proxy configuration.

I'm also not thrilled with the handlebars-esque DSL. Any templating solution without a general programming escape hatch is problematic in my eyes these days.

Users & Developers alike want a content management dashboard that bends to their sense of organization and needs. Many modern websites have more than one menu, for example. Custom post types are necessary for ease of use and quick content discovery. Modules exist to empower editors to insert beautiful, pre-designed content of their own making. Meta fields exist to signal to editors what kind of content is necessary to make a page viable (required fields). These types of features are becoming table stakes for modern content management systems, and they're very difficult to pull off in a meta-programming sense without the dynamism of a database.

All of that being said, I hope the author really does continue working on this. It's really great and exciting work!

It's roughly six thousand people's jobs (estimated base on headcount from last year), for anyone wondering. I don't know why they make you do the math, or provide such a fuzzy statistic.

This is deeply ironic to me because this company is run by the same guy who has his underlings tearing through our government data, secretly building repositories with titles like NxGenBdoorExtract[0]. I don't trust this guy at all.

[0]https://www.npr.org/2025/04/15/nx-s1-5355896/doge-nlrb-elon-...

On a podcast I listened to yesterday (Marketplace, from American Public Media), they mentioned that many businesses seem to be approaching this with an "We're in this together" type of attitude. They played some audio clips of business owners on their social media channels saying things like, "Don't worry [company name] is going to do absolutely everything to fight this. We'll do everything we can to avoid raising prices on you!"

The reality, of course, which they reveal when interviewed, is that they know they're completely powerless. In the best case, all they can do is delay the inevitable and they know it. This approach exists purely to keep themselves from being made into the enemy. As you noted, some business owners admitted to considering if they should just shutter their business too.

Makes me wonder how it would do if you tell it "put on your robe and wizard hat"

It reminds me of the Theo Jansen’s Strandbeests

I like to think there’s a little more conviction and adaptability there as you try to find market fit, both in terms of messaging and the solution itself.

If your point is that no business is a sure thing, then well, you got me. It’s true. My point was that “wait and see” isn’t a business model when you don’t have traction.

I don't know about you but I think trying to build a business on "Maybe (hopefully?)" sounds pretty scary. I'd much rather be literally giving away money (if that's what it takes) to get someone to be the Pepsi to my Coca-Cola on the platform.

Higher bit-rate video, faster load times, app stability, and edit functionality (which many people repeatedly request, despite it being arguably low-priority because you can just delete and repost), just to name a few. I'm practically a Bluesky fan, and I again disagree with Steve's characterization, but even I see many things that have room for improvement. Granted, Twitter/X should crush them on this alone and they've arguably leveled the field by making their platform worse over time, so heh.

> AT has low adoption

This is the clear and present problem in my eyes. I disagree with your first premise. I feel like it's unfair to call it over and unsuccessful. Many people are investing directly in the problem we're discussing. It's my hope this is the core focus of everyone working on AT at the moment: incentivizing some kind of competition in the ecosystem and creating and actively demonstrating that value add. Many want Bluesky to be viable because the alternative publishing platform is a dumpster fire. However, it's only fair to admit this doesn't magically make it so. I'd really like to see some more concrete details on how they plan to generate revenue.

This is an interesting theory. Under this line of political thinking, it’s just as important that the U.S. project that they would come to help if the aim is political stability or maintaining the status quo.

And it makes sense for it to slip into production too, because as far as tasks and values in processing training data go there’s no incentive to filter something like this out.

This is arguably true for all director’s cuts and remasters ever released.

My favorite is when it hallucinates a library that does exist and ostensibly has a relevant title and methods but upon investigation proves to be for a purpose wholly unrelated to the current task. To make matters worse, this discovery will be greeted, every time, by "Ah, you're right!"

His ridiculous self-prescribed dosage and willingness to abuse the drug alone is indicative of the careless risk-taking required to decide to do something like insider trading and think he could get away with it. The arrogance required, especially from someone intelligent and capable enough to do a job like his, is palpable.

80mg is four times the average dosage. Anyone who has taken this drug therapeutically for its intended purpose at sane dosages will tell you that it can have some nasty side effects if you’re not careful. In brief, it makes it more difficult for your body to tell you when its had enough and needs food or rest. Things like sleep and eating are simply more difficult. You’ll find that out quickly enough without a doctor informing you of such, but it’s manageable then. We could just as easily be discussing someone who increasingly decided to sleep and food deprive themselves for months or years.