- Migrate code and back it up
- Set up local AI
- Have local AI "patch" your github code by converting the entire program into brainfuck (https://en.wikipedia.org/wiki/Brainfuck?wprov=sfla1)
- Merge patched version to GitHub
- Profit
this post was submitted on 26 Mar 2026
441 points (98.9% liked)
Programmer Humor
30616 readers
1512 users here now
Welcome to Programmer Humor!
This is a place where you can post jokes, memes, humor, etc. related to programming!
For sharing awful code theres also Programming Horror.
Rules
- Keep content in english
- No advertisements
- Posts must be related to programming or programmer topics
founded 2 years ago
MODERATORS
I'm torn between wanting to opt-out because it's morally correct, or remaining opted-in so I can poison AI models with my terrible code.
so I can poison AI models with my terrible code.
Don't forget to teach it obscenities and yell at it whenever it fucks something up!
Nah, guarantee the models have rules built in to deal with obvious stuff like that.
You need to be more subtle. Give them information that is slightly wrong.
Artisanal crap code.
Just need to use less obvious insults, a la, "your mother was a hamster, and your father smelt of elderberries"
Still poisons the model with something an end user won't like, but isn't easy enough to train out
Perhaps by generating a bunch of complex copilot code to upload. It's easy to mass produce and would look plausibly functional.
Training AI models on AI content is the fastest route to model collapse.
... and tell it things, that are slightly obscene
I love the idea of giving CoPilot Torrettes.1
I signed up to github purely to opt in and upload terrible python code.
If they desperately want to train the idiot machine on my awful self-taught code, that's on them.
Chaotic good
Step one: Download a C or CPP repository.
Step two: Replace all semicolons with a greek comma.
Step three: ??
Step four: Poison Copilot, so that it randomly insert greek comas that the compilers totally choke on.
Por qué no los dos?
Opt out on one account, use another as poison. If you're gonna do this, I'd say move all your code to a new account and use the older account to poison - that way they can't filter the bad out by account age.
Don't worry, the models already spit out poor code quality.
Name all your variables poorly and with swear words
load more comments
(4 replies)
Link for opting out: https://github.com/settings/copilot/features
In the "Privacy" section, set "Allow GitHub to use my data for AI model training" to "Disabled".
Illusion of choice. That setting will, totally unbeknownst to them, enable itself.
I turned it off ages ago, and if you already had it off, it's staying off. (for now at least)
load more comments
(2 replies)
It seems to be off by default if you've already opted out of Copilot entirely. Definitely still a reminder I should set up my own git though
I run mine on a raspberry pi zero - wasn’t terrible to set up.
As soon as Microslop got involved I pulled all my repos and left.
Opted Out and moved all to codeberg
How is codeberg?
Has everything I need, but not more
My god, this is such a positive review these days.
It's great. I also self-host my own Forgejo (that's the software Codeberg runs on) instance for private repos, to avoid using up space on Codeberg's servers.
Main problem is the lack of federation, leading to splintering across Codeberg/GitLab/sourcehut/self-hosted forges. I know there's Radicle, and Forgejo is working on ActivityPub integration, but it's slow-moving to get what should be inherently federated by design (git) to actually be federated. In practice you need accounts on a dozen different websites if you want to regularly contribute to foss.
i love codeberg, though i haven’t had a chance to test the collaboration features all that much
Microsoft:
Fully automating supply chain attacks since (at least) 2026.
Not to be too snarky, but was there ever an assumption that stuff you put in wasn't being used to train it? Safe to assume that any online service you're using is making use of the data you're giving it.
If you’re a business with a contract with them it should state that they won’t use your data to train their models.
If you’re using the free service then you’re right that it’s safe to assume that your data was already being used.
business with a contract
I always wonder at this and have cautioned my managers repeatedly. Yes, we have a contract, but they have a literal army of lawyers and we have less (one lawyer one retainer for hourly work or a small grouping focused on taxes and employment law). As if our ownership won't bend over backwards to avoid suing a large company like Google, AWS, Microsoft, or Oracle. (Maybe OpenAI and Anthropic are sue-able by a $100 million corp?)
As proof I offer the lawsuits between businesses that have proceeded far enough the general public has heard about them. Not a specific one, just all of them.
load more comments
(2 replies)
I mean if it wants some absolutely abysmal code then look no further.
makes me want to fill my repos with absolute garbage code.
Way ahead of you
Hell yeah, I hope I contributed to some bot somewhere absolutely flailing to provide a good python snippet.
fun fact, if you've ever accidentally clicked the "enable" button on copilot because you're a dumbass who can't read, you get a shitton of more settings, most of which are locked to "enabled".
Even more fun fact, if you never clicked the "enable" button on Copilot, most of those settings are locked to "enabled" anyway.
yeah you just can't see them. fun!
Yes I just found that this morning. Time to seriously look at the GitHub alternatives.
Also another setting under CoPilot>Coding Agent - turn off for All Repositories - mine was set to On.
load more comments
(2 replies)
load more comments
(1 replies)
load more comments
(2 replies)
If you're still om github, you're kinda doing for it.
Got this email last night and felt validated for never uploading any code to GitHub because I don’t trust Microsoft. lol I don’t have any big coding projects, but I self-host a ForgeJo server in my mini rack at home behind a Twingate VPN.
FYI: it is not “ForgeJo”
Forgejo is derived from Esperanto where the “ejo” suffix means “place”. The J is pronounced like y is in English.
It’s “forge-ejo” not “forge-joe”
view more: next ›