It's easy to agree to that, but you're disregarding that the resources you spend on the stronger model could be allocated elsewhere. Conversely, you're assuming that spending more on AI will always yield better results and be worth it, compared to spending the money on other things.

This might actually still hold true now, or or at least many actors in the market behave that way. But I'm not so sure there isn't a cliff to that effect. At some point, if SOTA models remain expensive, it'll turn into a market advantage to figure out how to get things done without depending on the most expensive tooling available.

Similar scenario, different phrasing: if your company relies on overqualified workers to deliver 100% quality, the market may still decide that it's fine to go with 90% quality for 50% the price.

@orwin has claimed that SOTA LLMs have already hit that diminishing return where spending more money on a SOTA LLM today does not add more value than a non-SOTA LLM (assuming high value tasks).

I never said there will never be a diminishing return. I'm challenging the statement that we've already hit.

Note: We're still scaling chip nodes. It's still worth it for TSMC and chip design companies to invest hundreds of billions into every new chip node every 2-3 years. This is after decades of scaling already.

he clearly has never ran a firm

the constraint long-term is vision

and vision is really hard - no LLM will help with this.