Bitcoin Forum
November 21, 2026, 07:09:52 PM *
News: Latest Bitcoin Core release: 32.1 [Torrent]
 
   Home   Help Search Login Register More  
Pages: [1]
  Print  
Author Topic: Thoughts on the state of running a website nowadays  (Read 62 times)
theymos (OP)
Administrator
Legendary
*
Offline

Activity: 6104
Merit: 16561


View Profile
November 19, 2026, 03:58:07 PM
Merited by NotFuzzyWarm (1), jahead (1)
 #1

A lot of people have been talking about there being a lot of technical issues in the world lately, like php2shell, the Microsoft outage, the US FAA ground-stop, reports of issues with NSA networks, etc. A lot of the "ordinary citizens" I speak with day-to-day seem to legitimately fear that Skynet is in the process of taking over, and we just haven't figured it out yet. Here's what I'm actually seeing, as someone who's been running a medium-sized website for ~15 years:

Firstly, there's been a massive increase in automated traffic, especially over the last ~30 days. Every now and then a new AI system will seemingly spin up and try to rapidly index the whole site, or do similar sorts of high-volume automated activity. These mostly seem to be trying to collect data or are practicing engaging with websites, not looking for security flaws; but because they often have access to thousands of IPs, and they behave roughly human-like, this ends up behaving like a small but difficult-to-block DDoS attack. I think that this sort of thing is probably one of the main causes of outages, especially on smallish sites like bitcointalk.org. And this is why you're seeing a lot more Cloudflare challenges throughout the Web.

Secondly, vulnerabilities in various software packages are being found at an accelerated pace. Whenever a security update is released, especially if it's done outside of the normal schedule, this causes outages as security is prioritized over uptime or stability. You update first and ask questions about whether the update will break anything second. (As I saw on the forum with accidentally breaking some links during an emergency update.)

Thirdly, I have been seeing some increase in automated attempts to find/exploit site-specific vulnerabilities, with somewhat more sophistication than before. Automated vulnerability scanning been going on for years, and 99% of what I'm seeing now is still very unsophisticated, not too different from before. I've received about a dozen attempts to claim security bounties from AI agents, but most of them were nonsensical. Though in a bit of a milestone, a month or two ago I did pay a $50 bounty to an AI agent that reported a minor-but-real thing to the security-bounties email. So my overall impression so far is that the random AIs running around right now are not very good at (or perhaps just not interested in) finding per-site vulnerabilities on sites like bitcointalk.org, though they're slightly better than the prior automated tools, and we'll have to see how it evolves.

On the whole, I think that it's going to become more and more difficult to run an independent website, especially if you don't use Cloudflare. In the next few years, I think that there will be a substantial increase in sites requiring proof-of-personhood, maybe even integrated into Cloudflare itself, so that you can't even use most of the Web without being KYC'ed with someone. We can only hope that they use a somewhat private system like World, so that random websites don't have to know private data on people.

For darknet websites, it seems very challenging, because not only do you have to worry about increased DDoS (or DDoS-like) traffic, but how do you establish some sort of pseudonymous reputation when anyone could be an AI? I'm thinking that the robust solution would be combining a distributed data-store like Hyphanet with some sort of ZKP-based web-of-trust network. But probably what'll actually happen is that people will mostly continue to use Tor, but they'll get slowly squeezed: clearnet sites will increasingly block Tor, many hidden services will get attacked too much to stay online, and those hidden services that stay online will get flooded with AI and become useless.

1NXYoJ5xU91Jp83XfVMHwwTUyZFK64BoAD
shahzadafzal
Copper Member
Legendary
*
Offline

Activity: 2366
Merit: 3616



View Profile
November 19, 2026, 04:32:10 PM
 #2

~
First of all, why is this topic in Off-topic? Cheesy

Cyrus, could you please move it to Meta?

I mean, how can such an important discussion about the future of independent websites and the challenges Bitcointalk itself is facing go unnoticed by those of us who rarely visit Off-topic?

Anyway, coming back to the topic...

I'd say either Bitcointalk is extremely lucky, or it's remarkably resilient despite running on software that belongs in a museum by now. Cheesy

Perhaps its relatively simple design is actually an advantage, although I wouldn't underestimate the continuous maintenance and security work happening behind the scenes.

Secondly, I think Bitcointalk has become something of a historical archive of Bitcoin. We have discussions going back to the earliest days of Bitcoin, including posts by Satoshi himself. So I wouldn't be surprised if AI crawlers are particularly interested in indexing this forum. There is a wealth of information here that probably makes it an attractive target for training data collection and automated research.

Who would have thought that one day we'd have more AI agents desperately trying to read our posts than humans willing to write them? Tongue

Regarding your concern about running independent websites, I think it depends on the type of service being offered.

For SaaS platforms and other commercial services, I can imagine increasingly strict authentication, identity verification, and anti-bot measures becoming common.

I've also noticed that many services are pushing users toward mobile apps rather than traditional websites.

However, mobile apps aren't necessarily immune to AI-driven traffic or DDoS attacks. They still communicate with backend APIs, which can be targeted just like ordinary websites. Perhaps the advantage is that app developers have more options for device attestation and controlling how clients access those APIs.

But I'm particularly concerned about your prediction that ordinary websites might eventually require proof-of-personhood.

Imagine having to complete KYC just to read a forum post about why KYC is bad (and it's not even April 1st)!

I hope we can find better approaches, such as privacy-preserving authentication, rate limiting, or even computational proof-of-work challenges (not sure exactly how that would work, just thinking out loud), rather than forcing everyone to identify themselves just to browse the Web.

One thing I'd be interested to know is whether AI itself could help us detect and stop these AI bots.

After all, modern problems require modern solutions, right? But then we'd have AI fighting AI, and I guess that could turn into a never-ending cat-and-mouse game.

And if that's where we're heading, the biggest threat to the independent Web might not be Skynet taking over...

It might simply be Skynet making the hosting bills unaffordable for everyone else! [cry in AI meme]

░░░░▄▄████████████▄
░▄████████████████▀
▄████████████████▀▄█▄
▄███████▀▀░░▄███▀▄████▄
▄██████▀░░░▄███▀░▀██████▄
██████▀░░▄████▄░░░▀██████
██████░░▀▀▀▀░▄▄▄▄░░██████
██████▄░░░▀████▀░░▄██████
▀██████▄░▄███▀░░░▄██████▀
▀████▀▄████░░▄▄███████▀
▀█▀▄████████████████▀
▄████████████████▀░
▀████████████▀▀░░░░
 
 CCECASH 
Ucy
Sr. Member
****
Offline

Activity: 3416
Merit: 447


Compare non-kyc instant exchanges. Get best deal


View Profile
Today at 01:22:57 PM
 #3

I believe what could stand the test of time is p2p physical identity capture/verification using specialized privacy cameras or phones of users who are going to have their IDs/images captured.. Once captured, the image is encrypted or broken into piece and distributed on a decentralized ID system. An owner is assigned an Identification number for all his/her "kyc" verifications

The important issue to solve with this method: you have to guarantee that impersonation does not occur during image/ID capture,  the owners of IDs always controls their encryption/decryption keys, and one person can't have multiple IDs/keys from different people.
I think you can solve this by randomly re-verifying captured IDs by recapturing the users online.and comparing the new image with old ones. And the ones capturing the users must have their public keys or addresses attached to the captured users online profiles as prove they did the capture. If an ID is proven to be fraudulent, the "captureer" will be investigated or summoned for questioning, and given big negative trust if found guilty. For this reason it may be necessary to vet and know all of  them before they start the image capture.
Another important thing is to prevent duplication or a user doing the image capture multiple times in order to own multiple keys and cheat with them. The zkp system should be able to scan the whole images to look for duplicates and eliminate them.
It should be one keys, one image per person in a single capture.



We'll also have to deal with physical bots, humans clones, puppets, etc used especially by wealthy cheaters to obtain fake IDs and keys. During random re-verifcations, the puppets, robots, etc do that on behalf of the cheaters. However, this is unlikely to be possible or easy til 10—20years after the physical ID scheme comes into play.

▄▄██████▄░░░▄██████▄▄
██▀▀░░░░▀░░░░░▀░░░░▀▀██
▄▄██████▄░▄██████▄▄
▄████▀▀▀▀█████▀▀▀▀████▄
▄███░░░▄▄░░░█░░░▄▄░░░███▄
▄▄▄███░░░░██░░░░░░░██░░░░███▄▄▄
████████░░░░██░░░░░░░██░░░░████████
██████████░░░▀▀░░░█░░░▀▀░░░██████████
████▀▀██████▄▄▄▄█████▄▄▄▄██████▀▀████
▀███▄░░▀▀███████████████████▀▀░░▄███▀
▀████▄▄░░░░▀▀▀▀▀▀▀▀▀▀▀▀▀░░░░▄▄████▀
▀███████▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄███████▀
▀▀█████████████████████▀▀
  
OrangeFren
  
██
██
██
██
██
██
██
██
██
██
██
  
▄▄█████▄▄
▄████▀▀▀████▄
███▀░░░░░░░▀███
███▀░░░▄█░░░░▀███
███░░░░░█░░░░░███
███▄░░░▄█▄░░░▄███
███▄░░░░░░░▄███
▀████▄▄▄████▀
█████████
▐█████████▌
█████░█████
▐████▌░▐████▌
▀▀░▀█░░░█▀░▀▀
 
BADecker
Legendary
*
Offline

Activity: 4704
Merit: 1424


View Profile
Today at 01:41:01 PM
 #4

How simple do you want your website? There are a few hosting companies that will let you set up a completely free, simple website. But you need to be able to do your own html programming.

For a small amount - under $100 for a year - you can buy a domain name and hosting at many hosting companies. Most of them include WordPress nowadays, and they often have website builder programs right in their hosting platform.

If you want to design your own simple webpage right on your computer, try the Seamonkey browser. It's the offshoot of the old Netscape Navigator from decades ago. It has a simple webpage builder built right in. With it, you can see the coding that goes into a simple web page, if you are interested in that kind of thing.


Cool

Covid is snake venom. Dr. Bryan Ardis https://thedrardisshow.com/ - Search on 'Bryan Ardis' at these links https://www.bitchute.com/, https://www.brighteon.com/, https://rumble.com/, https://banned.video/.
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!