- cross-posted to:
- [email protected]
- [email protected]
- cross-posted to:
- [email protected]
- [email protected]
The web is fucked and there’s nothing we can do about it. Kev Quirk looks back fondly at Web 1.0.
The web is fucked and there’s nothing we can do about it. Kev Quirk looks back fondly at Web 1.0.
I think it’s time for a new standard to replace HTML and Javascript for the web that formalizes all the most used functions and gives developers way less freedom to make a crappy website.
Alternatively train AI to recognize crappy websites and severely punish them in search.
Or use AI to reformat websites into something user friendly. Considering the coding skills of GPT-4, I don’t think that’s too far away.
Between GDPR prompts, auto-generated articles, banner ads, normal ads, filler content, related articles, the web has become unusable.
Really it’s Google’s fault for not cracking down on these practices. And their competitiors for not doing so either.
Search engines in general have become beyond useless. I barely find anything anymore and it’s not just the fault of bad web design. Even if the search results followed human friendly design, they don’t even contain anything related to my search.
My only hope is that retrieval-augmented LLMs can fix this mess. Basically they read all these crappy websites for you and extract the actually useful information.
Replacing HTML and JavaScript does nothing to stop people from creating bad websites. People would still post auto-generated content or ads, filler content, related articles, etc… And having LLMs summarize bad content will only give you a shortened version of bad content.
This is exactly what I want. A simple text-based protocol. Sure, throw in support for images, too. Provide basic layout options so you can do proper wrapping and scaling for different size screens. Nothing else.
The web is too bloated for the basic use cases at this point. With HTML5 and JavaScript and CSS you can do anything, and honestly, that’s great. It’s great that I can run an entire OS emulator in a web browser. It’s great that I can run games and paint apps and everything you can imagine. But why the ever-loving hell is the same platform used for all of that and plain-text news? Madness.
AI-powered reformatting and extraction is bound to come, which is probably one of the reasons Google is pushing for their web DRM bullshit.
well there’s gopher or gemini, but I was thinking it should have some more modern features.
Like if you took all the features of all the best modern websites and apps, and condense them down into a fully integrated stack replacing everything from http up. Only allowing elements that don’t get in the way of UX.
Ideally it should completely preserve privacy and anonymity, so perhaps bolt something on like I2P. Make it pretty much impossible to track people beyond them voluntarily giving their information or doxxing themselves.
But then also have 99% of the conveniences of the best of modern web/app design. But beyond those fixed functions, you have zero freedom as a webdeveloper.
Fixing the content of the websites is then another problem entirely seperately. This is just to fix UI/UX.
Like if a zoomer designed gopher.
Together with more effective search engines, whose actual goal it is to bring quality content to users, you could fix the web forever. I think if you handed control over page rankings over to users, you could fix search engines too. You have to create incentive structures to align the interests of the search provider with that of the users. Currently Google has little incentive to actually provide you good search results, that doesn’t neacessarily make them money.