Did The Absolute Sound use an AI chatbot to review the Soulution 717 monoblock amp?

comments within this post are proud to say "glad I did not renew" a subscription, or I’m not supporting them anymore. And to demonise each other all over one article! These online comments, read by many does not put a positive light on our hobby.

Are we supposed to refrain from criticizing people just because they’re fellow audiophiles, or industry insiders, as Valin and Harley are?

I support honest journalism, not the other kind.

@gbmcleod
 

You wrote,

whatever glyphs anyone is seeing  on the website is not in the article, print or online

And this is the problem with TAS essentially falsifying the record. Without a word of explanation or accountability, they replaced the original article with a new and improved version that was probably not written by AI. Happened earlier this week, see the post by @Passthedutchie.

You can make your little insinuation that I and everyone else who saw that "glyph" (nice little racist touch BTW, Nepali and Hindi ≠ "glyphs") are making it up...but we all saw it. If you missed it, it’s right there in Passthedutchie’s verbatim dump of the original web article. Valin wrote, "SMPS delivering extreme current क्षमता."

Explain it. Explain that Hindi term. Explain how anyone but an LLM (Large Language Model, aka AI chatbot) could end up with a sentence that contains a random word of Hindi. It’s the smoking gun in the web article that TAS posted and left up for almost three months until someone noticed and TAS started to get slammed for it on the Internet.

And there were half a dozen other documented AI giveaways that you have consistently declined to address, hiding behind your "I’ve worked with these people and they’re paragons of virtue, hurr durr."

The evidence doesn’t lie.

@hardyboy 

I want to thank @Hardyboy for making my point, and the section he so kindly printed online proves exactly what I said: that the subscriber print/digital versions do not have that problem and they came out months ago. Not weeks. MONTHS.

Facts forthwith:

The print version, which I received mid-January (as well as the online version subscribers received) don’t have  the error in the online, NON-SUBSCRIBERS article, which appeared 2 months AFTER the print/digital subscriber versions (which we, as subscribers, ALWAYS GET, months before the public. My print version received around January 15th. TAS’  (error-filled) website article date shows: March 25, 2026.

Neither (subscriber print/digital)  version have that Hindi glyph (or whatever it is). By the way I saw it too, when I went to their website last month.  I even acknowledged seeing it in my post, except that I passed it off to whoever TAS pays to post it, screwing up. It is STILL, however, not in the print/digital subscriber version released in January.  Gee, so AI can reach into my home and magically change my January issue? That IS advanced!!!!!!

 I clearly stated that I’d contacted Harley about the supposed AI writing, and his response, already noted above.

The poster is fine with believing what he wants, although he clearly has no understanding of creative writing, and why using AI is death to those of us who have been writing for decades. It takes a lazy mind not to even  be able to imagine the joy that those of us who have published journalistic works (I wrote for The New Haven Register in the early 1970s, and it was one of the joys of my life) get when we see something we wrote in all it’s glory (or, sometimes, not☹️) . Had AI  even been around in 1976, I would have forgone it, because what good is my imagination if not to translate thoughts into a beautiful article. That just doesn’t compute with the poster, who thinks EVERYONE would use AI. I imagine he thinks Dostoevsky and F. Scott Fitzgerald would have used AI if it was there, and that’s his prerogative. A very shortsighted one, but, anyone can see how attention-deficient people are nowadays. Writing a 7,000 word article is beyond him. It’s just not beyond the rest of us.

By the way, Jonathan gives a clue as to what eventually give a clue as to what went wrong in posting it online: He states that he has just spent "7,000 words" describing the amp. Anyone doing a word count could see that the article only contains 4,700 or so words (my text edit is a basic program). Where are the other 3,000 words of the article? Oops, another error by whoever posted, but certainly one that the poster could’ve confirmed, as he is clearly technologically more knowledgeable than me. Except in reaching the correct conclusion. If he had the whole article (or a subscription),  he would have seen that the bullet points he refers to are part of the designer’s  section where the designer explains what he improved. It is clearly laid out in the bullet points, easily visible to anyone who is a subscriber. That’s not part of Jonathan’s writing

He’s right: the facts DON’T LIE. And the fact IS that the - by now - 5 month old print version looks like any other article in TAS. Nothing’s missing, nothing’s out of place. Oh, and it IS 7,000 words long (as Jonathan happened to mention), unlike the abbreviated version (4,700 words)  that TAS’ web designer screwed up.

I almost wish they’d left the errors online, so the more perceptive viewer could have looked at it. I wish I could publish the print version in the magazine I received in January, but I’m sure that that - even if not illegal - would not be cool with TAS.

As for my ’age’ being a factor, I can only point out that I am alive, in my seventies, when Michael Jackson, Karen Capenter and  Whitney Houston,  died before they turned 50. Unless the poster has an expiration date, he might want to muse on what living a long time gives you: perspective and insight. Right now, he has neither.

I’d have responded earlier, but my aging body needed a checkup. But that’s easier than needing to get a checkup for one’s mind.

Robert Harley’s editorial on using AI.

https://cdn.imgpile.com/f/NMOLkZO_md.png

Please pardon my clumsiness at posting. I simply wanted people to see Harley’s editorial on using AI (which is only for subscribers). 

And, as you can see from Robert's editorial, the AI chatbot gets things wrong. PLENTY of the time. I've seen it myself in dithering it expounds on, which I, as a DJ from 50 years ago, KNOW to be wrong.

I'm sure there are scenarios where AI is correct, but more often than not, the information I've seen is incorrect - ESPECIALLY regarding - but not limited to - music history.

Wondering if AI picked up the TAS review vs TAS using AI.

Is copying AI plagiarism?