Skip to content
Go back

What Folding Laundry Taught Me About Working With AI

Published:

What Folding Laundry Taught Me About Working With AI

Yesterday evening I was folding laundry. It was one of those pesky loads, the basket was filled with socks. We’re a four person household, in theory, that should make it easier to distinguish all the socks.

I made some space on the table to accommodate the individual socks. Laying them out flat helps find pairs. After folding one-third of the container, I realized that the space I had assigned was way too small and was already overflowing. Since the rest of the table was full, there was no more space to allocate for more socks. This was a seemingly simple and mundane task, that suddenly induced stress in me. Where would I put all these socks now?

Granted, the solution was quite easy in this case. I created some space by stowing some folded laundry, and I had enough room for the socks. What’s the connection to working with AI, you ask?

When AI became publicly available with the launch of ChatGPT, many people immediately recognized this technology’s potential. Recognizing that it’s a new technology with many unknowns, they created companies and planned generously, allowing the companies ample time to find product-market fit and generate revenue.

Stress occurs when plans and reality diverge. It’s the same mechanism, whether you have enough space for your socks or how much runway your company has. Right now, we see many companies entering a stressful phase, especially the big ones. OpenAI, for example, issued a Code Red in an internal memo. Apple abruptly fired their AI chief, John Giannandrea.

Delivering value with AI is a lot harder than everyone thought, we underestimated the complexity of AI. This has led investors to attempt crazy things, this TechCrunch article provides an absurd example: Pumping $90 million dollars into a business with an annual recurring revenue of around $400,000, valuing it at $415 million dollars sounds absurd. This strategy is called king making: declaring a winner in a market and hoping to convince customers to choose the “market leader.” It’s another symptom of the stress we’re seeing in the system right now.

This great article by Paul Ford brings it all together. He wishes for the bubble to burst, because the the frenzy for return on invest ends and we can focus on letting nerds do their best work.

Happy hacking!


Share this post on:

Next Post
Why You Should Buy an AMD machine for Local LLM Inference in 2025