This is a slightly modified repost of this reply, if you're wondering why the hell I'd post it. There's a happy ending here too.

If you've used Reddit for a while, you'd know they do april fools' things. One was a pixel placing game. After that was over, everyone rushed to clone it with modified properties: faster placement, more space, pyramid schemes and "if your clan is big enough we'll intervene and protect your stuff".

For me it first started with a site called pixelz dot io. (I'm not linking to any of these sites, don't go on them. This one died, anyway.) This was when the Brazilian factions were invading I think -- they had a fuckton of people who were probably doing it to give back to their country or some other baloney. We had a nice base, then it was smeared with green. We were very outnumbered, but I had copied out the requests made by the client and there was no authentication unlike PixelCanvas. I had a Python-based bot working by the end of the month and used it to reclaim the base. This kept us safe for a very long time, and we expanded very slowly but surely.

4chan came around and nearly fucked stuff up with more proxies and a better bot. I rewrote the bot in Common Lisp hoping for a speedup but unfortunately for a few weeks we were being taken over. I tried to trick the owner of the new bot into revealing important information, but I couldn't for a week. The operator of the website tried to put up CloudFlare to stop us. Ironically enough, the website was half broken cause requests would fail after a few minutes on the real client.

I read up on how to defeat CloudFlare, and strangely enough, there was an ftp dot pixelz dot io that wasn't protected at all. After some reading on how CloudFlare proxied connections, I had a bot which could push out thousands of pixels without any proxies needed. (I was about to improve the bot by making a fake HTTP "proxy" which lies about where the requests came from, which the server would believe -- but pwning CloudFlare was funnier.) We crashed the server that day, and took it back the next.

The enemy bot user also liked pony porn, who would have thought. A few minutes in the GIMP covered them up, and the operator intervened blanking out the particularly lewd parts. The bot owner would have to divvy out his bot time as well, whereas our program was only bottlenecked by the network. The player count easily climbed to and beyond 60,000 players. I tried to grow his defense thin, but that didn't quite work out. We won nonetheless, but the operator added some JS token obsfucated stuff I couldn't figure out and the base was slowly consumed. (I have a copy of the client JS sitting around. I think it's transpiled Scala or something.)

Later on, I wrote a fancy bot called cl-pixelcanvas based on the cl-pixelz intrinsics for some people to manage a hell of a lot of proxies for PixelCanvas dot io. (This one is still alive.) I'd wager almost 2,000 proxies were managed on one $3 VPS near the end. We had it working well for a month or two, but they kept asking me for more features (more analytics, captcha solving) and I couldn't cope with all the new requests, maintaining the server and defending it. There was even one scumbag who fucking tried to DDoS the server using leaked information from our captcha solving tool. In a breakdown, I destroyed their server and left.

Pixel* games are a mental drain and kept me from hacking the things I wanted to. I didn't want to work on much else after, realizing that if I got big again I'd have too many requests to deal with. I abandoned some projects I figured weren't going to make it through this burn out phase.

The moral is that you shouldn't waste your time being destructive if it's not a large problem. Internet swastikas are stupid and are made by stupid people, but there's better ways to procrastinate and hack. Go work on something constructive, it's way more fun. Last month I started work on a reboot of an old project, cl-decentralise. Writing it was fun to do. Do fun things.