New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Hello i would like to report a bug @oloke @onidel
A while back I won a LET giveaway ($50 onidel credit for pming oloke 'polishfemboy') and PM'd Oloke the keyword but still havent recieved anything several months later. I think there must be a sivir bug in the system !
I would also like to claim the $500 bug bounty for reporting this issue so more giveaways dont go unclaimed. Could someone please check on both of these for me?
Thanks and regards
The real bug is that I haven't gotten a VPS with them yet.
If an actual vulnerability was discovered by a person, and a way to reproduce was provided, the report would not be rejected.
I think we should clarify that we meant purely AI generated vulnerability reports here (without any evidence of exploitability or impact). We had a few of those in the past...
"Without any evidence of exploitability" should be the key.
And I can only imagine how frustrating it must be to get loquacious slop reports containing nothing but fluff confidently explaining to you that you're vulnerable because someone with root privileges could make bash setuid lmao
Right, much of AI slop report is horrible (like many those MJJs badmouthing hosts). However, I personally read the bug report that claude wrote and I think it's not sloppily written. Though, don't know how @crunchbits takes it
Deep, right up to the balls
@vitobotta
@rpqu already knew that.
AI-organized reports are fine. Especially if you're making sure it's harnessed with "brevity" or "be concise". Actually preferable. Often when I write replies I will ramble a bit or be out-of-order. Still send it that way anyways (proof of life) but for documentation/reports/etc organizing it is totally fine. What @oloke is talking about makes perfect sense.
I can't imagine directly giving an AI's output as a report. If I ever had to use an AI, I'd only use it as assistance but I'd write the report completely.
I’m on the server provision screen and for some reason Canberra isn’t showing up as an available location.
Thanks for the heads up. 500 bucks max is not worth my time though
WADR forest, LLM AI is a tool. It's usefulness depends on your ability.
Every token you input determines its state and the output. So, if you feed it low quality content, it will produce crappie content (read the report of AI getting brainrot). Also read claude's j-space
As for myself, I don't micromanage it like what's being thought by prompt engineering. I give it concise guideline, opcode, and evaluation. Perhaps, the motivation to keep the token cost low increase the accuracy because there's less noise on the matrix.
Oh no, would have been some easy money.