Literally used an LLM to answer the latest one 20 minutes ago, which was "what can climb a tree".
Cloudflare is a blight on the internet, but HN loves it.
Literally used an LLM to answer the latest one 20 minutes ago, which was "what can climb a tree".
Cloudflare is a blight on the internet, but HN loves it.
Cloudflare just started blocking Googlebot, so maybe we'll see an exodus away from Cloudflare by people who want their sites indexed? Hopefully?
I thought Google Search is dead, everyone is using ChatGPT now to find information.
but how does chatGPT find information?
With bots that break through cloudflare captchas. As a human, I immediately close the browser tab when I see one. If you make your site inaccessible for humans, only the bots remain.
I was more so asking how does it know where to look to find information, if things are no longer being indexed.
I assume they have their own index of the web. Or perhaps there are companies where you can buy such indices. And then mix and match those, perhaps with your own too.
Data is basically their business.
Cloudflare spreads everywhere and kills the drive to do nice things. It is akin to a disease.
The question is, what is the alternative to cloudflare? We've tried (to various degrees) user id, micropayments, device attestation... I'm not hopeful we can actually solve this. Maybe if there was a neutral, "NGO-like" cloudflare of some sort?
> Cloudflare spreads everywhere and kills the drive to do nice things.
How about an internet where we don't need CloudFlare?
One without zillions of AI crawlers and script kiddies with Terabit DDoS capability… (the latter being the harder problem)
Bittorent grinds on undisturbed.
Imagine archives with static content and user friendly file selection. Like those old world printed product catalogs mixed with content more like books and magazine. One can still have interactive headless websites/api and checkout pages.(Updated) torrents can share folder and file names but you have to force re-check. Could automate that too.
They can be many TBs large, no need to download the entite thing.
Not any kind of original work.
But
The llm thing started to change me in a sense that I no longer really feel obligated to use everybody else's solutions. It may be harder and harder to get onto the 'old web', but my new small project put a website from cellphone and is available for as long as I need it to be available.
The short answer to what we can do is 'adjust accordingly'.
It's like the good old joke about a bear-proof trash can.
There is now a considerable overlap between the capabilities of the dumbest users and the smartest AIs.
> but HN loves it
There is popular sentiment on HN that CF is making internet worse by being a core infrastructure element most websites use.
It’s not the core infrastructure part that makes the internet worse.