When the Fitbit Ionic had a recall, I applied for a refund 2 years later, sent in the watch (you have to mail it), and it took months to get my cash. I wouldn't hold my breath, it might take a while, which I think is fine, so long as they get it done.

I applied for a ticket refund for a heavily delayed train in the UK. A month after I applied they agreed and said they would deposit the money into my PayPal account. Three months later I remembered to check, and they hadn't. When I contacted them they were all "lol, whoops, how did that happen?". And finally paid me. Big companies really do take the piss.

In my case I waited patiently because it was outsourced, so I assumed its a company with maybe 10 people running through thousands and thousands of recalls for various companies. They might be slow, but as long as they get to it, doesn't bother me.

Why is it okay for a company worth tens of billions of dollars to drag their feet issuing a refund that can be done promptly?

Interesting, I also just used the support link, talked to an agent via text chat, asked for a refund, they asked for a screenshot, I shared a screenshot, and they immediately refunded me. They did not suggest an alternate approach (the 3 months of premium you mentioned). I didn't actually want a full refund because I got some use out of my Car Thing while I had it, I was banking on the 3 months of premium. Oh well. :)

Wow, they end the chat by insulting you.

The slightly clumsy wording would indicate to me that it's a real person rather than a chatbot.

We live in a time where AI chatbots speak way better english than outsourced support workers in 3rd world countries.

We live in a time where AI chatbots speak way better English than most native English speakers.

I'm wondering whether soon we'll see the outsourced CS chats run through a 'small' efficient LLM to 'clean up' (or in some cases, just re-localize) the English of such workers to match the locale of the caller.

unless the bots have been trained on outsourced support worker chat threads?

They were trained on both and can do both, it's a matter of asking for either in the system prompt.

I dont like this timeline

Have we passed the turing test?