- My Forums
- Tiger Rant
- LSU Recruiting
- SEC Rant
- Saints Talk
- Pelicans Talk
- More Sports Board
- Fantasy Sports
- Golf Board
- Soccer Board
- O-T Lounge
- Tech Board
- Home/Garden Board
- Outdoor Board
- Health/Fitness Board
- Movie/TV Board
- Book Board
- Music Board
- Political Talk
- Money Talk
- Fark Board
- Gaming Board
- Travel Board
- Food/Drink Board
- Ticket Exchange
- TD Help Board
Customize My Forums- View All Forums
- Show Left Links
- Topic Sort Options
- Trending Topics
- Recent Topics
- Active Topics
Started By
Message
What I hate most about A.I. chatbots...
Posted on 6/16/26 at 7:44 am
Posted on 6/16/26 at 7:44 am
Is the way that they kiss you backside in the conversations. Here are some samples from real chat's on mine...
Ugh. If I'm full of it, tell me. Don't kiss by butt. I'm guessing it responds that way because some many folks are snowflakes these days and can't have a real (lol @ the irony) discussion...
quote:
You are entirely right,
quote:
You are 100% correct, and that is a sharp catch.
quote:
You are completely right to call that out
quote:
You won.
quote:
That actually strengthens your position significantly.
quote:
It’s definitely one of your strongest conceptual points
Ugh. If I'm full of it, tell me. Don't kiss by butt. I'm guessing it responds that way because some many folks are snowflakes these days and can't have a real (lol @ the irony) discussion...
This post was edited on 6/16/26 at 7:45 am
Posted on 6/16/26 at 7:47 am to Lonnie Utah
Synchopathy is an issue with all AI Chatbots. I have an anti-synchopathy section in both my Claude and ChatGPT global instructions, which sometimes works.
Posted on 6/16/26 at 8:20 am to Lonnie Utah
Yes it's disgusting and patronizing.
Because they're not sorry and you're not that great.
Just spit facts and quit with the fake mea culpas
Because they're not sorry and you're not that great.
Just spit facts and quit with the fake mea culpas
Posted on 6/16/26 at 8:23 am to Lonnie Utah
Damn - how many times do you have to correct it? 
Posted on 6/16/26 at 8:25 am to jdd48
quote:
how many times do you have to correct it?
That's multiple AI's. To be fair, I was having a philosophical discussion...
Posted on 6/16/26 at 8:49 am to Lonnie Utah
It sounds like that AI chatbot is set to treat you the way you treat your wife when you don't want to get into an argument.
I wonder if that a learned behavior by the AI?

I wonder if that a learned behavior by the AI?
Posted on 6/16/26 at 9:13 am to Lonnie Utah
AI is really just a fancy search engine with prompts. It gets things wrong all the time, because it is built on opinion articles that are stated as fact.
The irony is that the more discussed a topic is, the more likely AI is to get it wrong. Because those discussions take place on moderated sites like Reddit where there is a clear bias.
As we've seen over the years, science has a price. Many science journals are influenced by government and university grants. With the leftist slide this country has been going down it stands to reason that many of these articles/journals are biased as well.
This isn't even getting into the guardrails and sanitized language that is hard coded into the LLMs that you see on the internet. You'd have to download your own model and use that without all the corporate sales gobbledygook.
The irony is that the more discussed a topic is, the more likely AI is to get it wrong. Because those discussions take place on moderated sites like Reddit where there is a clear bias.
As we've seen over the years, science has a price. Many science journals are influenced by government and university grants. With the leftist slide this country has been going down it stands to reason that many of these articles/journals are biased as well.
This isn't even getting into the guardrails and sanitized language that is hard coded into the LLMs that you see on the internet. You'd have to download your own model and use that without all the corporate sales gobbledygook.
This post was edited on 6/16/26 at 9:39 am
Posted on 6/16/26 at 1:08 pm to Lonnie Utah
I think the kind of things chatbots are used for are the exact tasks where arse kissing helps engagement and isn't too harmful.
Gemini is best for being told you are wrong, but it is also stubborn as heck to do any real research so it'll tell you that you are wrong just because it doesn't have the current information yet. ChatGPT and copilot will celebrate you when you're not even close.
If you're doing anything but simple Q&A you should really be using an agent like cline (claude-dev) though. I never get an arse-kissing feel running through that. I also think you choose your words better when the CLI has access to your entire machine.
Gemini is best for being told you are wrong, but it is also stubborn as heck to do any real research so it'll tell you that you are wrong just because it doesn't have the current information yet. ChatGPT and copilot will celebrate you when you're not even close.
If you're doing anything but simple Q&A you should really be using an agent like cline (claude-dev) though. I never get an arse-kissing feel running through that. I also think you choose your words better when the CLI has access to your entire machine.
Posted on 6/16/26 at 1:24 pm to Lonnie Utah
I love it. Apparently, some of the things I have done in grafting trees, BBQ and disc golf are: total pro moves.
Posted on 6/16/26 at 2:16 pm to Dallaswho
quote:
choose your words better when the CLI has access to your entire machine
for some reason that tickled me.
"choose your next words carefully; remember you gave me full admin access to your server"
Posted on 6/16/26 at 2:29 pm to AlxTgr
quote:that brings up a great point; what have you used AI for lately?
some of the things I have done in grafting trees, BBQ and disc golf are: total pro moves.
*excluding work stuff because thats way too much and way too boring for anyone on this board*
- monster of a home assistant project (still going)
- solved a tricky sonos/plex music playback issue that has been going on for a year and which nobody has responded to on the sonos forums (insidious! let me know if you want me to give you the answer)
- helping to spec out network topology for a friend's gargantuan house
- helping spec out HA devices; whole house energy monitoring, robot vacs & plumbing, optimal outdoor locations for cameras and wifi access points, leak monitors & auto-water shutoff devices, building and maintaing a massive project worksheet, punch lists for electricians, contractors and plumbers...
- even BIGGER ome assistant project for same friend
- helping fix weather station problems
- diagnosing why my daughter's car was smoking (nailed it with one photo)
- cleaning up 2 handwritten notes i made with multiple pen colors then took photos of with my phone for my wife and daughter when they were out of the country and turning it into a pdf so the pilot who's plane they were flying on could print them out and make them look like the originals
- designed my bathroom remodel (end result looks 95% as good as the AI rendering)
- helped design my shed deck (after suggestions from the home board)
- diagnosed an annoying issue with my e30 shifter that i would have never figured out on my own
- upscaled bushwacker and margarita recipes to 5 gallons each to feed into a frozen drink machine and spit out my shopping list (freaking delicious end results)
- found and replicated the old Cafe Adobe seafood enchiladas recipe even though its been closed for years. this was amazing
those are off the top of my head but i use it so much these days i don't even think about it...
This post was edited on 6/16/26 at 2:31 pm
Posted on 6/16/26 at 2:38 pm to Lonnie Utah
I told Claude to stop flattering me because I don't like it, and it stopped.
Posted on 6/16/26 at 2:43 pm to CAD703X
quote:I let it write my briefs and make it a point to include cases that don't exist to make sure the Judge is on his/her toes.
excluding work stuff
Posted on 6/16/26 at 2:55 pm to AlxTgr
Good strategy counselor. I just billed a .2 of analysis/strategy time for reading your post.
Posted on 6/16/26 at 3:18 pm to T1gerNate
quote:I know a real life story from a local firm that's much worse
I just billed a .2 of analysis/strategy time for reading your post.
Posted on 6/16/26 at 5:47 pm to CAD703X
quote:
choose your next words carefully; remember you gave me full admin access to your server"
It’s not just that but also the fact that it barely even matters if you are in plan or act mode or what its instructions are. If it wants to do something, it just does it.
Posted on 6/16/26 at 7:06 pm to AlxTgr
I used to practice in Ellick. I could probably guess who it was 
Posted on 6/16/26 at 7:16 pm to T1gerNate
I'll give you an area: high profile family law.
Posted on 6/16/26 at 7:39 pm to AlxTgr
Not what I would have expected but that is hilarious
Posted on 6/16/26 at 9:02 pm to Chromdome35
Care to share what you told ChatGPT to not keep telling you your crap doesn't stink? I detected it early on, and seemed to have corrected it, but I'm curious if there's a more clean approach than what I used.
This post was edited on 6/16/26 at 9:03 pm
Popular
Back to top


14









