Comment by photochemsyn
4 hours ago
If corporations really thought LLMs were a great cost-savings tool, then the obvious target for replacement are not the lower-paid staff, but the higher-paid staff - the ‘product managers and stakeholders’. That justifies token burn, replacing the 7- and 8-figure people, right?
But that’s not the real goal, is it? The goal is to inflate the stock value, take the cream off the top, and dump the whole business on the pension funds, maybe creating a too-big-to-fail scenario where the government steps in an bails out the industry as with the airlines during Covid.
This is why all the testimonials and narratives are so suspect - nobody knows what fraction of online posts were created simply to sell the narrative that LLMs are this incredible disruptive tool that will change the world, solely in order to create FOMO in the investor class.
In this particular case, I’d like to see links to samples of LLM created codebases for “PCI compliance, double-entry ledgers, escrows, reconciliation, payment lifecycles, bank transfer idempotency”. It should be easy to put an open-source LLM-generated version up on github, right? And if not, why not?
The idea is to start with the largest, easiest lever. The one which will accelerate all _other_ automation. That lever is software development itself.
Say you are Anthropic and want to shake up the world of law or medicine or whatever. What will you need? Product managers? You need tooling, software, infrastructure and a lot of it and quickly and you need to iterate really F fast on it as well.
If you automate the development of software itself you will enter a new era in which automation of All The Things becomes an engineering problem instead of a pipe dream. Besides software engineering there is (AI) research/science and robotics. That is the holy trinity. Crack that and it's over.
BTW: "double-entry ledgers, escrows, reconciliation, payment lifecycles, bank transfer idempotency", these all sound like solved problems and also things that are festering with accidental instead of essential complexity. I won't bet my career on those things. Now if you say something like physics or geology, that's a tougher nut to crack.
I agree, we are still waiting to see the billion dollar valuation startup that fully vibe-coded their product