[sublinear]

> LLMs are a way to make software take orders of magnitude more computational power, electricity, and human labor, while delivering a product whose extremely volatile quality is impossible to assure. The work will never be completed; it will only create the need for ever more labor. For investors, all this churn is attractive. It's disruptive. It's why we can't have nice software.

This quote is the foundation of the blog post and unfortunately misses the point entirely.

Yeah, I agree with all these arguments against using LLMs for software development. That's precisely why they aren't going to be used for that.

Nobody wants to increase the cost of real development. Everyone would rather skip having to write code at all to get a result. From a high level perspective there's no difference between an LLM producing junk results and a team of humans writing bugs.

Though when humans get it right they really get it right and the product is worth much more due to its maintainability built off a good release. This is the missing insight from the blog post. Many businesses will still have to pay for real software dev and those days are not numbered.

LLMs really do grow the market legitimately. Low quality software is like store brand canned goods. It's a high upfront cost for the factory, but the product can scale on other merits.

These products are not mutually exclusive. You can get rid of your call center with LLMs including all the software the agents needed to service an account, but you still want humans writing the critical infrastructure and all the high visibility frontend stuff because you care about the data integrity and good UI/UX. This is what most devs already do all day. There's no threat here.