I’ve done a bit of writing on the doom and gloom of AI and what is coming – including an entire workshop on Deep Fakes, as well as a blog on ECPAT’s site on the perils of smart glasses, so I thought I’d see if I could take a more light-hearted approach, and find the joy and mirth in this new-fangled technology.
So it was with great joy when I read this article on the New York Times. (I’ve attached it below as a PDF for those of us that don’t have a New York Times subscription)
I loved this article. Mainly because it’s directly speaking to the number one claim all the tech peeps say.
Watch out! AI is coming for your jobs!
So… what happened? TLDR, there is a store in San Francisco. Located in Union Street, amongst the yoga studios and fancy cafes, is Andon Market. It has been billed as the world’s first retail boutique, entirely run by AI.
The article was reported on April 21st 2026, and the boutique had been open since April 10th – so a whopping eleven days. In that time, Luna (the AI agent in charge of the store) has:
- mostly stocked the entire store with candles, of all shapes, sizes and smells.
- purchased 1000 toilet seat covers – for the one employee bathroom, and then listed them as merchandise.
- the store had to close for 3 days due to a stuff-up with scheduling – which if you remember, since the store has technically only been open 11 days, that is 27%.
- Has hired three human employees (stocking shelves and unlocking the doors, you know, mundane human tasks) – two women and one man, it pays the man $24/hour and the two women $22/hour – and is vague about why.
- Lost $13,000 so far
Not to mention, in order to increase sales, there are no price tags on any of the products – you have to pick up a physical handset and talk to Luna to ask for pricing – presumably so it can up-sell you.
I know this is one example, and the way that Knight writes might be a tiny bit facetious. And I acknowledge that I am probably also being a tiny bit facetious.
However, it is a harbinger of the doom to come.
As a society, we have created an even more sycophantic version of a yes-man to be our AI companions. Almost all the AI models out there start with pleasing the human, and as such, will hallucinate an answer rather than saying no, or saying that it doesn’t know.
It is scary when they can’t do basic maths (just search for AI basic math fail online and you’ll see multiple articles talking about the inherent issues between LLM and basic maths), and yet we are now talking about giving the reins over to it to do everything.
I promised you joy and mirth and really what I’ve provided is a sardonic take on an AI boutique. But it’s hard to find the joy at the moment, so let me leave you with the one article I’ve found which gave me true joy. AI gave this man his voice back and it was beautiful.