Conversation

Notices

1. Embed this notice
   David Chisnall (*Now with 50% more sarcasm!*) (david_chisnall@infosec.exchange)'s status on Saturday, 12-Jul-2025 18:07:20 JST David Chisnall (*Now with 50% more sarcasm!*)

   Wow, an alleged human actually did this:

   Someone posts a series of sensible questions to frame a discussion on a working-group mailing list and, in reply, someone pasts them into Gemini and just attaches the output, a huge PDF that is full of completely irrelevant slop. It adds zero value to the discussion. It isn’t even a useful framing for more sensible thinking. It’s just a massive wall of text.

   I’m starting to think that the concerns about LLMs hurting critical thinking ability might have been seriously underrated.

   • RamenCatholic 🐢 🌈 repeated this.
   • Embed this notice
     David Chisnall (*Now with 50% more sarcasm!*) (david_chisnall@infosec.exchange)'s status on Saturday, 12-Jul-2025 18:08:04 JST David Chisnall (*Now with 50% more sarcasm!*)

     in reply to

     • random stranger
     • Stephen Illingworth

     @JetSetIlly @kyonshi

     That’s the confusing thing. I might post a lmgtfy link if:

     • The questions were stupid and obvious and,
     • I wanted to be a dick.

     But, in this case, I think the person posting slop genuinely thought they were helping. The questions were asked by someone who had put a lot of thought into them and were a set of things that we, as a group, should think about. The response was someone who seemed to think that not thinking was a helpful substitute for thinking.

   • Embed this notice
     David Chisnall (*Now with 50% more sarcasm!*) (david_chisnall@infosec.exchange)'s status on Saturday, 12-Jul-2025 18:08:06 JST David Chisnall (*Now with 50% more sarcasm!*)

     in reply to

     • random stranger

     @kyonshi I’m not the WG chair, but if I were they would be gone from the WG instantly, no first warning. I did reply pointing out that posting slop added nothing to the discussion and that it was fine to use stochastic parrots to help frame your own thoughts, as long as the thing you share is then the result of thinking. But I just don’t understand the mindset of someone who thinks that this was a useful contribution to the discussion.

   • Embed this notice
     Stephen Illingworth (jetsetilly@mastodon.gamedev.place)'s status on Saturday, 12-Jul-2025 18:08:06 JST Stephen Illingworth

     in reply to

     • random stranger

     @david_chisnall @kyonshi It's akin to posting a "Let Me Google That For You" link in response to a question. Posting LLM output is even more obnoxious.

     Rich Felker repeated this.
   • Embed this notice
     random stranger (kyonshi@dice.camp)'s status on Saturday, 12-Jul-2025 18:08:07 JST random stranger

     in reply to

     @david_chisnall point out and block participant who did that?

   • Embed this notice
     Tyrone Slothrop (slothrop@chaos.social)'s status on Sunday, 13-Jul-2025 07:20:04 JST Tyrone Slothrop

     in reply to

     @david_chisnall Not too long ago, we had a manager who was tasked with mapping out a new topic.

     He would produce a new white paper every 4-6 weeks, and distribute them to his peers. 40 page beasts, mostly LLM generated.

     Our usual medium of exchange in the company is presentations of 2-10 very dense PPT slides.

     He wasted lots of senior management time, without ever achieving anything. They moved him off to a strategy department elsewhere after 12 months.

   • Embed this notice
     CyberFrog (froge@social.glitched.systems)'s status on Sunday, 13-Jul-2025 07:23:38 JST CyberFrog

     in reply to

     @david_chisnall@infosec.exchange I’m starting to think that the concerns about LLMs hurting critical thinking ability might have been seriously underrated.yes, absolutely

   • Embed this notice
     Robert Kingett (weirdwriter@caneandable.social)'s status on Sunday, 13-Jul-2025 07:24:02 JST Robert Kingett

     in reply to

     Something similar happened to me https://sightlessscribbles.com/posts/20250616/ @david_chisnall

   • Embed this notice
     RamenCatholic 🐢 🌈 (ramencatholic@mindly.social)'s status on Sunday, 13-Jul-2025 07:26:03 JST RamenCatholic 🐢 🌈

     in reply to

     • DNA schedule
     • tuban_muzuru

     @ryanprior ...and it doesn't do anything to keep it from stealing it as "training data." @tuban_muzuru

   • Embed this notice
     DNA schedule (ryanprior@mastodon.social)'s status on Sunday, 13-Jul-2025 07:26:04 JST DNA schedule

     in reply to

     • tuban_muzuru

     @tuban_muzuru why do you think this is a useful observation? What do you picture the summary might contain?

   • Embed this notice
     tuban_muzuru (tuban_muzuru@ohai.social)'s status on Sunday, 13-Jul-2025 07:26:05 JST tuban_muzuru

     in reply to

     @david_chisnall

     Heh. If you want Gemini to summarize, it's very good at that - but my advice is to ask for a limit on the output "-in 500 chars or less"