That doesn’t answer the question at all, and I wonder if it’s actually true? A cache is not magic; it is, itself, just a static file server in the end. If I self host a static page website on an nginx box, do I actually need cache to serve today’s crawlers?
The screenshot in the image says 3k req/day. That’s 2 requests per minute (amortized). At that rate, you can serve it with cgi and Perl.
Cache is only relevant if you have a lot of traffic AND dynamic pages, or if you care about latency (which is only relevant for humans).
Everyone can be lying. But I’ve been around human beings long enough to know that there are two very different types of self delusion: valiant assumptions about what you will do in a never before seen situation, and observations about what you have done. GP’s was an objective statement:
> I've worked with people who were super productive with high quality work, and my reaction was to... gravitate toward working more with them.
Neither type of statement is perfectly trustable (nothing is) but IME there is a categorical difference. Your paper (and first comment, “don’t be so quick to judge”, which imo was ironically prescient) are about the former type.
Of course if you disagree with me on this fundamental distinction then we have found our contention :) which would be a nice end to this debate. Don’t you think?
You use research as an argument, which is valid in a conversation where nobody has any information about specifics. E.g. in the pre-life, before a soul is about to be incarnated, you can point to that research and say: you are more likely than not to behave this way. Were the soul to reply, “no I am not, I know myself”, you could call them delusional.
But you’re talking to a person who can point at their actual life and say: I have been in that exact situation and I can confirm that I did not behave that way.
That’s a new observation, and afai understand Bayesian statistics, this is the moment where we must update our priors: how likely is someone who has observed themselves in the past not to behave that way, to behave that way?
Your argument is now incomplete.
Maybe someone with real understanding of Bayesian statistics can frame this better, or tell me why I’m wrong XD
FYI I recently discovered a 1p browser extension feature named “Password Generator History”. It has a record of all generated passwords, whether their respective items ever ended up saved or not. Live saver.
I used to see far more references to Postel’s law in the 00s and early 10s. In the last decade, that has noticeably shifted towards hyrum’s law. I think it’s a shift in zeitgeist.
pianu.com used to be a website where you could learn piano by connecting your piano through usb with the browser. It seems defunct now but I found a video demonstrating it : https://www.youtube.com/watch?v=kTBmRV02NgI
I used something similar in the past. It was a legitimate use case for web usb which changed my mind on it quite a bit.
https://www.charachorder.com/ sells ergo keyboards and allows you to update their firmware directly in the website, through web usb. No local apps at all. Also an improvement in overall security from having to download some .exe / .dmg and running it locally.
There’s discussions in a (thankfully banned) sub thread pulling this into question so I just wanted to add sources:
> The short-form video-hosting service TikTok was under a de jure nationwide ban in the United States from January 19, 2025, until January 22, 2026, due to the US government's concerns over potential user data collection and influence operations by the government of the People's Republic of China. However, the ban was not enforced. The ban took effect after ByteDance, the China-based parent company of TikTok, refused to sell the service before the deadline of the Protecting Americans from Foreign Adversary Controlled Applications Act (PAFACA). Prior to the ban, individual states, cities, universities, and government-affiliated devices had restricted TikTok.
The executive branch explicitly defied an act of Congress which was upheld by the Supreme Court. First Biden for one day, then Trump II as he took office and continued not to enforce the ban.
I’m guessing the app works but their prod servers don’t? If they can point the app during review at a “self hosted” GitHub Enterprise server on a test domain with AAAA that would pass the requirement as stated by gp , without requiring GitHub.com actually support ipv6.