You have commented 358 times on Rantburg.

Your Name
Your e-mail (optional)
Website (optional)
My Original Nic        Pic-a-Nic        Sorry. Comments have been closed on this article.
Bold Italic Underline Strike Bullet Blockquote Small Big Link Squish Foto Photo
Cyber
There's a Whole Hobby Industry I Was Unaware of - Poisoning AIs
2025-01-31
[ArsTechnica] AI haters build tarpits to trap and trick AI scrapers that ignore robots.txt. Attackers explain how an anti-spam defense became an AI weapon.
Posted by:Mercutio

#6  If you've ever touched somebody else's code, you'd know.
Posted by: Skidmark   2025-01-31 18:05  

#5  Unfortunately I fear people will 'cite' AI the same way they would, or used to, 'cite' Wikipedia

I did an experiement with the Jebrains AI assistant which 'helps' you write code. I asked it to 'refactor' a good chunk of python. First time was pretty good, Second... ok, by the fifth time the code was just mush. Looked like Joe Biden had written it while drunk... That was just code which it should have been able to handle easily...
Posted by: CrazyFool   2025-01-31 16:29  

#4  I have absolutely no idea what that is, mossomo, but congratulations on accomplishing something that sounds terribly impressive! :-)
Posted by: trailing wife   2025-01-31 16:04  

#3  Like any tool - AI proficiency will be in the hands of the holder.

I hit a level with my certifications that book learning, MS Learn, Tutorial Dojo, etc - just dont get me over the hump to pass the tests. AI has changed that.

If AI has all these bad facts and hallucinations, I'm not sure I would have passed my AZ-104 exam b/c I relied solely on ChatGPT to drill down the higher level concepts I was not getting from the exam review material.

Posted by: mossomo   2025-01-31 13:01  

#2  How an AI-written book shows why the tech 'terrifies' creatives
Posted by: Skidmark   2025-01-31 09:35  

#1  I can only recommend that anyone planning to rely on AI processing data with factual accuracy have a conversation about a topic where he has intricate knowledge.

So far I've always found AI to be prone to make subtle and also glaring mistakes like hallucinating fictional historical figures.

AI is good with esthetics and bad with facts.
Posted by: Elmerert Hupens2660   2025-01-31 00:41  

00:00