Ask HN: What's your proudest hack?

I kept a small amount in the consumer accounts. Maybe $300-500. The cash took 2-3 days to do the loop if I recall. And this allowed me to keep do multiple $20 withdrawals each day on each account.

It got big enough that I employed someone to sit at the ATM and make transactions. Then I bought 2 more ATMs so they could multitask / reduce idle time during the 1 minute it took for the machine to process and dispense.

I stopped it when I realized the numbers were getting kind of big and a few banks had caught on and closed my accounts. Then because of the physical cash, I was worried about getting robbed. I was getting very paranoid about the whole thing and decided to wrap it up. I remember meeting a few times with attorneys during it to get varying opinions. Over time, the advise turned into "we think it's legal, but you should really stop it".

There was also another smaller but pretty invisible loophole I found with interchange. I could set the surcharge to $0 and so long as the cardholder bank didn't charge "foreign ATM fees" then I'd make about $0.50 per transaction. Much smaller, but order of magnitude more banks offer this feature versus the surcharge rebate feature. The surcharge and the rebates were line items on the bank statements. So, any person at the bank that looked at my account could see the fees and that it was costing them money. With this hack, there was no line item on my bank accounts (only the amount withdrawn). So it looked like the ATM owner did not benefit at all. But, in the background the ATM owning entity received ~$0.5 per transaction of interchange fee.

It sounds stupid, but at the time, my idea was to test this out and then sell a 'get rich quick' style ebook explaining how it works.

I was working for a medical device company that shall remain nameless. They had a system that was used in operating rooms that had a fairly nice user interface - bubble keyboard, touch screen, good graphics. It let you input patient information and stuff. (At least part of it was running embedded Linux under the hood.)

In the source code, I saw what it did with the patient name - it used a call to system() to store it in a file. So naturally, I tried an injection attack. I put in "[Name]; sync; sync; reboot". Sure enough, it rebooted the system.

But before the reboot, it had also saved the "patient name" in the file (and, because of the "sync", it had been written out). On boot, it read the information from the file and tried to treat it like a newly-entered patient name, which caused it to reboot again...

We had to re-image the device to recover. We added some validation to user input after that little demonstration.

If you're using dockerized services (which are a must for NVidia-based CV, since their CV libraries are a dependency hell), you must use some kind of networking solution to communicate. RabbitMQ might be considered overkill, but it does the job and is robust.

It's not nice to talk that way about others' work, especially when your comment insults, but doesn't provide any insight whatsoever.

I owned a generic ‘286 12 MHz DOS computer at the time. Since PC-Kwik was the Toshiba edition it would not run on generic ‘286. So I used DOS program Codeview, also on floppy, to find and successfully patch the code in the executable program that checked for identification as a Toshiba product. My very first reverse engineering experience.

We'll, it sorta worked. In this state I somehow recognized that the device placed on my fingertip must be a blood sensor of some sort, and that squeezing my finger might mess with the readings. I made it through two rounds of setting off alarms before the nurses caught me. I was immensely pleased with myself!