It's not just content/info/data, it's a performance (in the creative sense).

Brands spend a lot of time honing their appearance - not just fonts and colours but the whole composition and visual pacing - their entire "say something without saying anything at all" aspect etc. Just walk through any place with physical shops and really look at how the stores have worked on their appearance and how they present themselves to customers. They're not just selling a product, they're selling a lifestyle/feeling/etc/etc. They're not just going to give that creative control away to some LLM.

Another way to think of it is instead of people watching a movie or play when they go to the cinema or theater, they're just given the script to read. Same information but the entire artistry of both the performers and the directors is totally absent, leaving it up to each reader to imagine the delivery of lines or the scene's setting etc.

I think on HN and in tech in general people seem to forget that "the first bite is with the eye", and that is why "normal people" never liked or used RSS. The desire to leave our mark and to create (and view!) visually appealing things seems to be pretty innate in humans - we've been doing it since cave paintings. I struggle to think of a world where we just hand that over to AIs and humans have zero creative control.

1. Branding. Companies want to control their interfaces for all sorts of reasons. Branding is a big one. Clarity and comms are another.

2. LLMs in the hot path. LLMs are expensive, a hell of a lot more expensive than executing some Javascript locally. Hell, you'd still probably need to do that under this model anyway. We're likely to see LLM usage filter into the right places, use-cases with higher leverage, LLMs to create a UI that is shipped to all users over LLMs creating UI on the fly every time. Costs and time will dictate this just like they have dictated how every other technology is used.

And it's not just semantic information, presenting any kind of information in a way which enable the user to seamlessly interpret and use it is not an easy task.

AI, definitely lowered the bar for making some UI, but it doesn't help with the fundamentals challenges of making a UI, at least not more so then it helps with the fundamental challenges of any other job in our industry.

Put that in a SaaS for an office and the outcome is the true representation of work being done in that office, plus clear signals about edge cases (aka “the user is not using his custom built flow, why?)

In a sense, related to https://danieldelaney.net/normal/

Stability etc can be handled post-hoc: once a customized ui proves some benefits (via user adoption, or whatever you think efficiently measures productivity gains), it can be formalized by a human coder, who gets the full picture and has all the domain knowledge baked in, as long as you don’t capture UIs only but also the reasoning that built it.

Back to article: smart to think this in terms of browser, since that crosses the boundary between SaaS

Having a traditional web page with styles and assets AND the spec allow LLM's to be a bit more guided by the original site's design and intent. More of a remix or like Arc's boosts/skills feature.

There's also the reality the a lot of the things you'd want to be promptable (sorting, functionality, enrichment) couldn't be done on just the front end. You need some mix of UI and API logic to be promptable...

Why’d you make the prototype a separate browser instead of implementing with a chrome extension? Something like greasemonkey but with an LLM generating the scripts on the fly..

Going further, AI internet browser could be an entirely new app to break from the legacy.

I feel this with coding agents, so often where it fetches web data and interprets it, html in that loop is only occasionally additive. Feels quite futuristic

If it's personalized clients, that's what we already had for most web services before the iPhone and app-ification of everything. It failed because making things compatible is a hard problem and a highly political/bureaucratic tarpit.

> most SaaS products still ship hand-crafted React apps, each building its own UI, its own accessibility layer, its own theme system, its own responsive breakpoints

Contrary to popular belief on HN, building these React apps are not "bullshit jobs" in the broader corporate world, nor going to be replaced by AI. They're the backbone of all ecommerce today and the ground floor for business operations because they keep us out of the tarpit. The implementation details are irrelevant here anyway. The actual problem was always how a business retains full control of its brand and UX.

Smells like default Claude voice. I like the ideas, but if someone can't be bothered to proof read their own article, then I don't know why we should trust that any of it was human generated.

I am convinced neither client side nor backend side AI solutions will solve this.

Fully on topic: It would be naive to believe that serious web offerings would allow you to do this. Reality is moving in a different direction: Try applying custom css and js to reddit, for example: The website is a nightmarish matryoshka of shadow dom components and that‘s only the beginning of the flashification and silverlightification of the web.

[0] https://www.w3.org/TR/xforms/