#this is something genAI can't properly replicate

[fayesemblemring]

Lately because of my job I have to reluctantly use Gemini for the sake of testing it (for context I work as a mobile QA tester. Not gonna say any more than that because I'm a big fan of not doxxing myself) and after having to repeatedly try it for days I started noticing a couple of things.

First things first: When critics say genAI is basically just a fancy autocomplete, that is a huge understatement.

Since my tests consist of giving it various prompts and making sure the answers are correct, I had to look closely at the answers. And aside from bugs like it randomly coughing up words in Russian which I mean, that's what I'm there to help fix, I started noticing quickly that a lot of the responses are basically identical.

Sure, the context may vary depending on what exactly I ask it, but the phrases? The ways of speech? They're basically the same. Hell, sometimes it reuses entire paragraphs.

And to be clear, it's not like tendencies or patterns of an actual human writer. Rather it's like it takes a preexisting module of the response and just adds the appropriate words to make it make sense.

And sure, bugs aside it's technically correct, but at the same time it's eerily uncanny because no human genuinely writes like this all the time. Again, it's a fancy autocomplete, nothing more. It's like writing an essay by picking the suggested words from the phone and have it just happen to make sense somehow.

When you realize what genAI actually is and recognize the patterns... well, that's a hell of a lot less impressive than what all the techbros make it out to be, right?

Which leads to my second point- Ironically, now that I've been forced to use it repeatedly, I'm not afraid of genAI replacing writers and artists. Not anymore.

I've heard a few people say that there is no genuine threat of genAI replacing authors and essays or students passing by submitted works generated with it because genAI... can't actually write.

I was skeptical of this at first- after all, capitalism loves prioritizing cutting costs over quality (nevermind the sheer amount of electricity this thing wastes but I digress), so why would companies care as long as it's passable?

But honestly, I think I get it now.

Remember, I noticed entire paragraphs and sections being practically copypasted, and this is only after a few short responses. And I haven't study a thing about academic writing or linguistics, most I've done is write fanfiction whenever the planets happen to align.

Can you imagine having Gemini, ChatGPT or whatever write a book or heck, even an entire novel and trying to get it published into mainstream? It'd immediately get clocked out and laughed at like it rightfully deserves.

Now, I'm not naive. Again, capitalism. Companies will keep shoving genAI down our throats as long as they can get away with it (especially since they now have employees "correct" the generated works which is more work than having them make it from scratch. Best possible system amirite?)... but at the same time the fact they keep pushing it this aggressively says a lot, doesn't it?

genAI is already starting to cannibalizing itself. Ironically, by now plagiarizing those works it already made by plagiarizing actual artists and writers and now pollute the Internet. And considering how much electricity it costs to keep these things running...

Well. I'm no financial expert, but all these aggressive tactics sure sound like sunken cost fallacy, don't they?

Just saying. If this was a thing people actually wanted or used, every tech company under the sun wouldn't practically force us to use it (always opt-out of course, never opt-in) with increasingly hidden buttons to turn it off. Microsoft wouldn't passive-aggressively announce Edge now having genAI features with "Resistance is futile". Musk wouldn't feel the need to slap a button under every art post on Twitter to plagiarize it with its genAI and then pretend those original works were also generated with it.

It'd be a feature like any other, with no need to push it so aggressively. This reads more like "please, please you HAVE to use our genAI, we wasted millions on this", because just like NFTs, they've always been a pyramid scheme. This one was just better at hiding it and appeal to the average person. Ironically, I think they of all people know they can't push this for long the most.

I still despise genAI for everything it represents- the wastefulness and disregard for the environment, the confusing buzzwords of the techbros who are in love with it and most of all, the utter contempt and lack of respect society has for artists, writers and anyone who has done any kind of creative work- and I truly think it can't die soon enough.

But having to use it so much for work ironically gave me confidence it will die someday. It's not even a hope, more like the logical conclusion.

Because genAI is not the future, is not an exciting new technology, is not revolutionary and most importantly is not a replacement for artists and writers.

It's just a fancy autocomplete.