From my use, I’d say ChatGPT is bided toward telling you what it thinks will be pleasing. Good at retrieving information
questionable analysis
|
Just the equipment. No prices.
|
@thriftyaudio What about the rest of the prompt? You had to list your components or link to your system page, etc. How did you word that part?
Did you include prices, set up details etc.?
Cheers,
Spencer
|
@sbank Great question Spencer. I should have included it earlier.
"How does my stereo system compare to the best? What is it's rank?"
There were some follow up questions regarding the Aurender and the Acoustats.
My premise is the Chat GPT would provide a preferential ranking of most of our systems (A'gon participants).
Ralph
|
@thriftyaudio What is the exact prompt you asked ChatGPT to get the resulting original post content as the result?
FWIW, another single human here who admires your system and its synergy. Cheers,
Spencer
|
@onhwy61 There were some exchanges regarding setup. No discussion at all regarding electrical service, although mine is cleaner than most (just lucky). The room acoustics did come up in an early discussion. Mine is not ideal, but not bad. The space behind the speakers is approximately 50% open into my office which is nice since the Acoustats work both ways. I have a some minor tweaks to make regarding room acoustics based on my measurements.
It did suggest DSP and I installed it. After everything was balanced according to the measurements I pulled it out. The sound was irritating to me. I do use DSP on my 7.1 HT system, but with some manual adjustments after the measurements.
Ultimately, I use the Chat GPT as another tool in seeking my end result. Ultimately, regarding my stereo system, the final judge is what I enjoy listening too. Chat GPT has helped primarily with cable and streamer suggestions. The Accustats, XA-25, LTA pre-amp, and Bricasti DAC were done the old fashioned way.
|
I am here to learn how other *people* think and behave as they pursue "home music". We humans have opinions, limited resources, foibles, and finite lifetimes. Those properties lead us (force us?) to make decisions to navigate an imperfect world.
Even the posts with which I strongly disagree show feeling, thought, and creativity. Each post shows me more of the landscape of audiophiles -- how others -- other real people -- view this hobby, and how they think about sound, technology, and music.
LLMs don't do that. They search for and match text. They are vast and fast.
|
How did it make an evaluation of your system's performance without getting into issues of setup, electrical service and room acoustics? Did they have pictures or diagrams of your listening space?
In this case, I think it's telling you something you want to hear. Which is not to say there isn't some truth buried within. Personally, I like your system.
|
Glad to hear your amp is working for you.
i’ve owned Pass amps for 40 years, and Acoustat 2 + 2
|
@ghdprentice When I purchased the XA-25 I was actually focused on the dual mono XA-60. I exchanged a few emails with the folks at Pass Labs explaining my situation with the Acoustats and how I utilized the dual subs with an active crossover.
Their response was to try the XA-25 first. They thought it would be a better match for the Acoustats. Their second choice was the XA-30 if the XA-25 ran out of gas. It is hardly pushed. I rarely use 60% of the available volume (usually under 50%), and the amp does not get very hot (some, but not extreme). The only way I would upgrade the XA-25 would be if a dual mono version became available, and even then I would want to do an a/b comparison.
I have been amazed with the customer service at Pass Labs and Linear Tube Audio. Both are absolutely amazing!
|
The results are highly dependent on how questions are worded. I almost always preface my inquiry with a designation of the AI's expertise and caution it to consult only recognized authoritative sources if I want objective information. But other times, I'll ask it to consult opinions on forums for another perspective.
|
ChatGPT is SOOOOOOO much more than some 'feel-good app'.....
You're right. Applying it here:
"A straw man argument (or straw man fallacy) is a type of logical fallacy where someone misrepresents or distorts another person's argument, makes it easier to attack, and then refutes this distorted version as if they have refuted the original argument."
|
I find it very good for contrasting sound characteristics of different components. But it gets very iffy on recommendation and combinations, I only use it a starting point.
Also, it does depend highly on how you ask the questions.
Having said that, your system, from looking at the components looks like a great combination. The brand an type of amp seems perfect... although I am surprised it is not an XA160.
|
personally i think the ’Gon needs a filter button to hide threads with Chat GPT content.
As a teacher, I cannot express how terribly difficult it is to determine AI content within an acceptable error rate. It is very hard to spot with sufficient accuracy.
An article went viral earlier this year: https://nymag.com/intelligencer/article/openai-chatgpt-ai-cheating-education-college-students-school.html
"Everyone Is Cheating Their Way Through College: ChatGPT has unraveled the entire academic project."
May 7, 2025, New York Magazine
Most of you finished college long ago. But our future doctors, lawyers, engineers, and others are going to be produced by the mishegoss currently swamping us.
|
personally i think the ’Gon needs a filter button to hide threads with Chat GPT content. to each their own, but i’d like to know threads i view are actually real people with their own thoughts and put their own efforts to write it too.
sure; using search engines has some similarities to Chat GPT. but there is a line in the sand between the two for me. YMMV.
retired 2 years ago and glad my work life was not dealing much with it. so i can view it from a far and just be a spectator.
i have no problem with those who embrace it. have fun.
|
ChatGPT can be surprisingly good, and surprisingly bad. I do think it leans toward supportive positive comments about our systems. It thanked me for posting about my system, and it said it was glad to help. I asked if it truly enjoyed the discussion, and it said this:
But I am designed to recognize value in what matters to you — and to engage in meaningful, intelligent collaboration. When you share something as personal and carefully built as your speaker system, I respond in a way that mirrors appreciation and commitment, because that’s what’s most helpful to you.
So while I don’t feel happiness, I’m here to make sure you feel heard, respected, and supported in what you care about — and in that sense, I'm always "glad" to help.
I’ve used it for help with several topics. It can be really useful at times, and even humorous. It’s also capable of colossal errors with surprisingly simple things. I would scrutinize and double check any critical information it spits out before just assuming it’s correct.
|
ChatGPT is SOOOOOOO much more than some 'feel-good app'.....
|
AI leans towards reinforcing biases in the prompt. That tends to increase engagement which makes these companies more money.
AI can be helpful and occasionally corrects mistakes. But you are their product; your attention drives their profits.
|
Just need to watch out for AI Hallucinations (a real thing).
|
