3 ms·
Dead links are the worst. Can someone make a browser extension that uses gpt to hallucinate websites for dead links?
by ibash 2y ago
Dead links are the worst. Can someone make a browser extension that uses gpt to hallucinate websites for dead links?
- breck 2y agoYou might find this interesting: https://news.ycombinator.com/item?id=40948221 https://news.ycombinator.com/item?id=40948221
- jdougan 2y agoThere are extensions that talk to the Wayback Machine
- euroderf 2y agoWould it be so difficult to ping all links in a page being loaded, and clearly mark the dead ones ? Sure it would eat some bandwidth, but so do prefetch schemes and other stuff intended to speed up page renders.
- 6510 2y agoYou could probably do some fun compression scheme where you provide just enough information for a fixed llm version to generate a page that satisfies its authors goals. Your idea seems fun if people can one by one describe what the page use to look like and/or crawl the web for clues. We can have our under reconstruction banners back.
- GaggiX 2y agoI don't think there is a browser extension but there is a site called websim AI that creates fake pages in real-time using LLMs and it honestly works surprisingly well.
- sva_ 2y agohttps://websim.ai/ https://websim.ai/
- xyst 2y agoWe used to call this “imagination”. I guess in the digital age this is something that is lost :3
- quectophoton 2y agoAnd there are probably a few whole generations already that don't even know that link rot affects all links, not only links to "internal" URLs (like, say, a Discord image). So there's artists that just link to some social media website, not considering that accounts can be suspended, usernames can change, etc. Similarly there's also developers that "link to"[1] dependencies without considering that repositories might disappear (together with the source code for that dependency's version if nobody backed it up), a package's version might be removed from registries, online documentation for a dependency could disappear (ugh), etc. [1]: Just adding name+version to whatever manifest file and forgetting about it forever. Maybe adding a cache (not even a proper mirror, much less any self-sufficient way to build the dependency in case of disaster).
- bofh23 2y agoTo mitigate link rot I always include a title along with the URL. This is especially important for URLs with opaque ids like on YouTube instead of slugs. If you visit a pulled YouTube video link you’re left not knowing the title to search for it elsewhere. This doesn’t always work because many websites neglect page titles. I’ve always wanted to ask a wide range of web developers why they neglect titles. Why?
- cxr 2y ago> I always include a title along with the URL The fact that people use the HTML title attribute for everything except for including the proper title of the thing pointed to is sort of perturbing. Even for Google SERPs and here on HN, it would be useful to have access to the full thing instead of truncated titles when they appear, but neither site is an exception to the tradition of non-use.
- mikewarot 2y agoAt least if you have a dead link, you can paste it into the Internet Archive's Wayback Machine and have a non-zero chance of seeing what was there.