Alphabet
Soup, man.
The fuckers enticed me to “trial” Gemini for free. A chance to see how Google’s latest and greatest stacks up against the competition.
It turns out it’s rubbish compared to GPT for what I do. And slow, to boot.
No problem, I thought. I’ll just cancel the trial. That’s when I discovered it wasn’t a free trial anymore. I’m paying for this shit.
My simple task had become a mission: get my old plan back and stop paying for a tool I don’t want.
And that’s where the ethical line got crossed. I went to “downgrade” to my previous plan, and there’s no option anywhere. None. Just a comprehensive list of more expensive plans that I can upgrade to.
Good job, Google; we’ll turn a blind eye to the monopoly so long as you pretend to be a good corporate citizen. Walling off the exit isn’t that.
So I asked Gemini to help me cancel Gemini. It couldn’t. To its credit, it did acknowledge the absurdity of the situation, noting it was a “sharp, insightful, and frankly damning summary” when I pointed out the deep irony of the situation.
Gemini’s final suggestion? Email customer support.
I did. An email came back from “JC” at the Google One team. It was a masterpiece of AI-generated empathy, a word-perfect template that managed to say nothing while asking me to do the very thing I had already told them was impossible.
“Thank you for contacting Google One support and for being a Google One member!
This is JC and I’ll be the one who will assist you today. I hope this email finds you well!
I do apologize for the delayed response.
I understand that you want to downgrade from the Google One 2TB Premium plan to the original 2TB plan because you do not want to pay for the Gemini AI feature. I know how important this is for you to resolve as soon as possible. No worries, let’s work on this together.
To better assist you and help resolve the issue as quickly as possible, could you please provide a few more details? This will help us investigate and give you accurate guidance:
When you go downgrading the subscription, can you please send me a screenshot of the option you have when you try to downgrade your subscription?
I’ll be looking forward to your response.”
So I went back to my unwanted AI assistant and complained that its advice had led me to another AI. Probably the same one through some internal API. And, no shit, this is what it suggested I do:
‘Reply directly to the email and write: “This automated response has not resolved my issue. The self-service options have failed. I demand this ticket be escalated to a human support agent for a manual plan downgrade.’”
Let that sink in. My unwanted and expensive Google AI assistant suggested that I get rid of it by zig-zagging around Google’s AI-generated customer support with Google’s AI-generated prompt engineering.
This snake has gone so far up its own arse that there’s only a black hole where it used to be.
The one thing I will say for Gemini is that you can copy any part of the LLM reply without having to hit the “select text” menu option in Android. It’s a small thing, but it goes to show that even with a groundbreaking technology like an LLM, good UI is 90% of the competitive differentiator.
On that subject, whoever first figures out how to make their LLM remember and inject user preferences into every single chat, they win. Down with emdashes!
Also I wouldn’t mind a “post chat to blog” option. Imagine all the great content that would be published and then used as the source of future LLM-generated answers?
offshore westerly 