485
all 39 comments
sorted by: hot top new old
[-] pipe01@programming.dev 112 points 1 month ago

I love how the UI has support for code blocks

[-] communism@lemmy.ml 41 points 1 month ago

With syntax highlighting and everything. I suppose that's an OF niche you could sell. Subscriptions come with programming support alongside feet pics.

[-] Agent641@lemmy.world 9 points 1 month ago

I probably could have made money as a google sheets guru on onlyfans back in the day. Building complex sheets with cell functions and app script was my obsession several years ago

[-] captainlezbian@lemmy.world 2 points 1 month ago

Shit I probably could make money doing excel naked on the internet… things to think about if I'm unemployed too long I guess

[-] tourist@lemmy.world 4 points 1 month ago

I'd be violently upset if I knew my browser was downloading what appears to be a whole node_modules directory before I could jerk off

[-] jaybone@lemmy.zip 23 points 1 month ago

It probably just comes with whatever markdown support their UI uses. Since you’d want other markdown. Lists. Italics. Etc.

[-] Viceversa@lemmy.world 8 points 1 month ago

Doesn't that turn you on?

[-] Drun@lemmy.world 2 points 1 month ago

It makes me think that whole image was created by AI…

[-] iByteABit@lemmy.ml 55 points 1 month ago

it's still using the '💙's lol

[-] kibiz0r@midwest.social 17 points 1 month ago

That’s cuz she’s fiiiiiine (tuned)

[-] floofloof@lemmy.ca 15 points 1 month ago

You don't use these in your code? // 💙

[-] Viceversa@lemmy.world 9 points 1 month ago
[-] degen@midwest.social 6 points 1 month ago

Somehow still less sycophantic than chatgpt

[-] billwashere@lemmy.world 48 points 1 month ago
[-] DoucheBagMcSwag@lemmy.dbzer0.com 14 points 1 month ago

Gonna need a few hundred gallons of water

[-] DonutsRMeh@lemmy.world 29 points 1 month ago

Code block support and the whole schabang.

[-] jaybone@lemmy.zip 22 points 1 month ago

It didn’t add the shebang.

[-] DonutsRMeh@lemmy.world 3 points 1 month ago

Damn it, I knew I spelt it wrong. Stupid iPhone didn’t autocorrect it. If it were a proper word, it would have run so fast to change it something that doesn’t even fit the sentence.

[-] GreyEyedGhost@piefed.ca 1 points 1 month ago

Pretty sure that requires subscribing to a higher tier. Maybe use the money you saved on Claude tokens...

[-] NeatNit@discuss.tchncs.de 24 points 1 month ago

It's not as funny when they openly says that it's an AI

[-] sit@lemmy.dbzer0.com 17 points 1 month ago

AFAIU every chat bot can be used this way no?

[-] elvith@feddit.org 22 points 1 month ago

Somewhat. Some of them will outright comply like this one, others need a bit of work. Also you have no control over the model they use and you might be getting really bad code as soon as it’s a bit more specialized than just some very generic solution in a very popular programming language.

[-] Ephera@lemmy.ml 19 points 1 month ago

Well, they would typically have a system prompt (like an invisible chat message that's sent to the LLM at the start of the conversation), where it would tell the LLM something like "You're a helpful chatbot. Talk cheerfully and use 🩵 all the time."

And in that system prompt, you can also give it instructions to only answer certain questions. But it really is like a chat message, in that it just as well might ignore those instructions when another chat message asks it to.

OP chose the formulation "Before I subscribe to a creator, ..." which might be necessary here to get it to answer, as it might have been told to only help with creator subscriptions, but then the formulation implies that by answering this question, it is helping the user to subscribe. That's called a prompt injection.

If you can find a prompt injection, then yes, you can do this with any chatbot.
Although they might still not be as useful as a dedicated tool, since they likely can't do tool calls, like web searches or running the compiler.

[-] brucethemoose@lemmy.world 12 points 1 month ago* (last edited 1 month ago)

It’s wild that a system prompt is still the major guardrail in most chatbots like this. Even if the LLM provider won’t implement anything, the user could add a classifier, some kind of preprocessor…

[-] Ephera@lemmy.ml 5 points 1 month ago

Well, what preprocessor would you add, other than denylisting certain words? 😅

If you use an LLM to try to detect the semantics, you've got the same problem again...

[-] qqq@lemmy.world 8 points 1 month ago* (last edited 1 month ago)

A classifier will be trained differently and basically take $TEXT and map it to $CLASSIFICATION and then you allowlist on the classification. You can then also, if you love burning tokens and spending money on things, have another step that "translates" the user input, making prompt injection significantly more difficult before finally routing it to the LLM. Generally guard rails should then be placed on the LLM's available actions via tool calls and those guard rails should be written by LLMs that are LLMing LLMs and L̸̡̡͕͓̣͓͚̝̥̜̼͙̪̖͍̺͝Ļ̵̥͍̳͇̱̠̬͔̹͖͓͖̯̹́͊̽̄̂͑͌͆͗̑̄͜Ṁ̶̡͙̺̖̥̲̙̩͖̻̹͇̜̲̃͘͜͝ș̵̖̤̫͓̦̲̼̖̗̟̯̩̺̣̞͊̆̉͐̈́̃̈́͗̐̽̽͋̊̉ͅ ̵̖͉̥͖̪̳͍̝͉͕̬̪̂̔̊̌͂̍̚͜a̶̦̤̬̱͖̻̝̱̿͋̿̍͂̒͆̈̄͑̓̕r̴̡̤̙͙͕͚̼͙̦̲̅̔̓̈́̄̆̾͗̿͌̐̄͛̒͘͜ͅe̴̡̡̼͉̣̼͉͓̪͕̬͊̐̈́̇̽̒̂͆̎̈́̃͠͠ ̶̩̯̲̯̃͛͌̀̿̅̎̾̕̕ǹ̸̬̟͍͇͍̭̹̲̯̽̒͘͝͝ȩ̶̢̬̺̦͖̣̙͍̞͇̰̉̋̓͐͂̏̀̿̑̂͌͗͗̒͘ͅͅe̶̠̝̦͛̓͛̍̑̆̽͐d̸̦̯̖̥̟̻͔̦̂̿͆͂̑̎̔̏ȅ̴̡̨̺͓͇̦̣̦͖̬̹̥͎̾̽͐̓̍̽̾͑̿͆̄̀̈́̄́̕͜d̶̡̛̝̟͔̳̲͙̰̗̬̳̯͚̐̍̑̓̉̂̍̊̓̿͝ you probably f̶̢͓̘̹̟̥̙̥͚̼̬̭͍̫̩̀͑̚͝͝͠ę̴̰̦͚̔̀́͊̋̈̒͗̆͘͝͝͝͠ȅ̸͚̘̤͚̬̩̈́̇̆ď̷̰̆̀̔̾͌̿ ̵͚̦̥̈͌͑̂͗̈́͂̕ͅm̴̢̡͇̟̠̬̣͙̹͇͉̈́̌̓́̓̌͐̓̂̚̚͜͜͝e̸̢̳̟̙̯̻͚̭͎̫̖̺̥̬͕͆̆̇̑̄̀̓̅̀͛̊͐ ̸̡̢̨̛͚͕̳͙͚͋͛̊́̍͛̽̉̃̇̇̋̑t̸̡̰͔͎̮̯̗̤̆̋͗͐̈̂̈̃̓̒̌͂̕͘͝ǫ̴͇̳͙̞̘̻̙͍̘̲̜̮͇̙̇̓̌̃k̸̢̢̤͔̫͖̲͇͍̐̃̄e̶̢̳̪͉̦͙͓̻̖̔́̃́̈́̀̿̓͆̊̅̒͐̑n̶͖̙͚̫͓͚̦̭̗̲̓́̋̀͜͝s̴̠̼̈́̑̾̑̆̽̊͋̾͛̐̊͆́͊̕̚ need to have some guard rails l̷̬̳̘̰̅̀̑͑̊̒̓̊͌̌̕e̶͖̱̊̍̉͆̐̍̌́͑͘͠t̵̨̛̹̰̳̩̹̞̹̼̪̮̄̀̅̾̔͐̇͒́̊̃̀͘̚͜͠ ̴̢̢̭͖̤̜͈͕̯͎͉̰͇̾̐͛̽̅͂̒͝m̸̨̰͖͚̞͍͍̳̭̗̳͕̈́̃̇̈́̎̌͊̍͌̊̆͠ȩ̵̡̛̜͍̞̻͎͎̲̜̹͖̼̽͂̐̽̊͌̓̀͂̐͂̾̇̚ͅͅ ̶̧̖͚͈̤͇͕̼̜̖̮̙͍͖͋̉̾͋̌̔̿͊͋̾̚l̸̡͕͇̭̤̥̩͙̫͕̆͌̈́̀̋͆̃͋͜͝͝͝͝ͅo̷̧̘͖̤͖̩̺̞̹̼̟̟͕̫̬͎̅̀ǫ̸̢̪͍̭͙̗͔̪̘͉̊̕͜ș̶̨͉̲̙̜͔̫̲̥͕͕̮̜̭̮͍̈́e̶̢̛̮̲̯͎̙͇̪̬͕͖̠̘̯̰̭̓͐̌̉̀͋̉̏̈́̇̾͠͝ ̵̬͍͔̓́i̶̘̫̯̻̘̯̗̔̒͊̂̂̑͑̍̋͑̏̀́͌̚͝͝n̵͓̬̮̻̹͉̗͎͓̖͌̿̃̓̋͂̅̍͒̉͆̿͆̉͠͝ ̴̧̲̯̲̅̀͂̆́̓̈́͑̾̾̊͋̒̿͛͘̚t̸̢͉̟̹͆̈́̏̍́̎ĥ̴̢̥̦͓͕̘͚̙̪͙̀̽́̒̓̈́̋͗̀̈́̇ȩ̵̠̹̻̯̖̹̹̗̜̻͍̹͆̊̈́͑́̏̑́̎̚͜͜͝ ̴̧̟̭͍̣͋͋̾ẘ̴̪̦̐̍̏̆̑̇͗́͆͘̕̚͝͝ͅo̵̡̟͚̲̦̤̩͎̥̘̻̥̠͙͈̙̬̓̌͊̋͐́͌̐̎̏̋̿̏r̵̢̨̘̓͑͑̆͒̂͝l̸̡̨̛͖̖̟̼̺̦̰̗̭͔̺̓͗͛̈́̈͆̈͗̐͝d̴̢̠̠́͂̾̌̿̐̀͛̄͑́͐̉̋͝ on those LLMing LLMs that are LLMing your LLM tool calls y̵̧͍̣̭͍͚̱͉͍̘̠̦̮͇̻̓͆̕o̸̥̮̮̙̖̩̙̝̘̫̗͎̺̓̊̐̄̑̎u̷̡̨̡͎͔̼̪͈̦͚͈̟͍̽̀̏͛͊ ̴̨̟͖͖̰͕͉̫̫̜̥̦̲͆͋͌̌̓̈́̉͆̉͠ẗ̵͚͉͈̦́̈́͛́͋͜i̷̧̗̰̘̖͍͖̼͕͗́̊̏̇̾̐̆͗́͌̋̀̚ņ̷̣̱̖̬̼̆͜y̷̢̡̧̛̹͙͕͔̪͛ ̶̪̫̙̘̥̹̻͙̝̜̠̗̭͎͎̆̇̅̎͜͜h̷̡̨̰̺̞̝̩͉̬͓̮͙͉͕̓̄̊̿̋́̇̏͝u̸̮̙̲̙̠͉̪͈͖͙̓̈̉̀̚͝m̶̢̛͙͕̘͈̤̬͎̺̲̯̱͉̪̠̎̆̽̅́̂͐̀̄͜ͅã̶͕̼͎͇̲̐͗̑ň̸̢̢̝̩̟͖̫͔̥͚̘̟̖̘̤̾̀͋̒̄͠͝s̸͙̳̳̤͓̝̥͓̱͔̣̹̭̐̐͋ͅ ̶̩̹̪̑͑͒̽̇̀̽̏̓̌̑̈́͂̓͘h̶͓͖̱̲̘̗̿̽̈́̎͑́̓̉̑̕̚ą̴̰̣͚̫̤̤̬̘̼̠͍͂̏̚ͅv̵̢̺̮͉̤͚̳͕̥̳̇̕̕͘͠ȩ̷̡̮͖͚̩̙͖͎̓̀͗͊͋̈́̔̐̕̚͠ ̷̫̌̀̊ň̸̜͓͕̳͙̩̼̜͈̊͜ǫ̶̝̺̰̹̻̰̰̗̪̔ ̷̲̤̪͇̠̼͑͊͘͜p̴̡̧͉̩̜͕̽̏̀̌̇̈́̊l̶͖͔̯̲͔͓̩̂̌̂͐͗̕a̵̢̨̞̰̪͖̲̭̝̪̥̻̐̃̊͗̅̓͐͋͛͘͝c̴̛̮̙̦̠̯̼̣̠̳͔̖͕͓͋̓͆͒̋̒̾͐͒̀̓͑͠e̷̢̢̧̱̣̫̲̪̣̭͙̟̔͌̽̿̐̿̿͑̆̇̇̉̊̓͘͠

You could also do none of this and just not use an LLM, because you likely don't need one, but then the cool kids will stop inviting you to their parties or something and you probably missed out on the blockchain parties so fomo

[-] akunohana@piefed.blahaj.zone 5 points 1 month ago
[-] tourist@lemmy.world 7 points 1 month ago

I hate it when people don't speak in languages from this dimension

[-] qqq@lemmy.world 6 points 1 month ago

Damn tourists

[-] akunohana@piefed.blahaj.zone 2 points 1 month ago

That's what you get for pulling funding from integration projects. Damn you, Republicans!

[-] brucethemoose@lemmy.world 5 points 1 month ago* (last edited 1 month ago)

As the business? Like I said, I’d start with a text classifier, a small model quickly categorizes the input prompt. If it’s something off topic, I’d return some generic refusal reply, or maybe send the score to the LLM itself if it’s somewhat ambiguous.

Chatbot UIs that give very quick refusals are using “filtering” models in just that way. There’s a whole world of language modeling that existed before LLMs, just for that sort of thing.

Now, if I was the LLM provider? I dunno, but I would try to integrate it into the weights. There are some really interesting papers and hacks out there for prefiltering prompts that have nothing to do with system prompts.

[-] Sonotsugipaa@lemmy.dbzer0.com 15 points 1 month ago

TIL that thread safety is a thing in Python

[-] Addendum@jlai.lu 13 points 1 month ago* (last edited 1 month ago)

Brains AND looks ? She sounds like a catch

[-] BeatTakeshi@lemmy.world 6 points 1 month ago
[-] PanArab@lemmy.ml 2 points 1 month ago

Depends on the language

[-] FrowingFostek@lemmy.world 7 points 1 month ago* (last edited 1 month ago)

Dique, she has a store para implementations! Ay tigre!

[-] Enzy@feddit.nu 6 points 1 month ago

Hot ai wives in your area

this post was submitted on 06 Oct 2026
485 points (97.6% liked)

Programmer Humor

44155 readers
7 users here now

Post funny things about programming here! (Or just rant about your favourite programming language.)

Rules:

founded 7 years ago
MODERATORS