The article starts out interesting: > But DeepSeek didn’t just build another AI model. They wrote their own parallel file system (3FS) to optimize costs—someth

The article starts out interesting: > But DeepSeek didn’t just build another AI model. They wrote their own parallel file system (3FS) to optimize costs—something that would have been unthinkable for a company of their size. This was possible because AI helped write the file system. I thought, hmm, now that’s interesting. But then the breathless leaps of faith comes quickly > Now, imagine what will happen in a couple of years—AI will be writing code, optimizing infrastructure, and even debugging itself. An engineer with AI tool can now outbuild a 100-person engineering team. > This kind of disruption was previously limited to narrow consumer products like WhatsApp, where a 20-member team built a product that led to a $19 billion exit. But now, the same thing is happening in business applications that requires breadth Umm, not so fast, AI will not enable a couple of guys to write a robust production grade kernel in a couple of weeks or months, unless that kernel is simply a fork of Linux with some cosmetic tweaks. I always wonder, do the people making these sorts of pronouncements actually understand software? Have they built software of any significant size and depth? And then: > First, the internet killed the need for sales teams (distribution moved online). Sales teams have not been killed simply because distribution has moved online, although kind of sales people you need, and what they do, has certainly changed.