[diffeomorphism]

Being a "failure" means it fails to meet its success criteria.

If you claim 10x and deliver 3x, that is a failure. The 3x may still be impressive or a gamechanger or ..., but it still falls short of its promises.

[diegolas]

10x is a marketing gimmick in development, AI and pretty much everywhere

[KptMarchewa]

So what? Who cares about such a binary claims?

[diffeomorphism]

People making decisions.

Pay me 8x to get 10x, great. Pay me 8x to get 3x, nope.

[KptMarchewa]

I have not seen reasonable claims that people are paying 8x the salary of a SWE to get those 2-3x results.

[richiebful1]

Directionally, that's why companies are getting rid of token leaderboards and imposing limits on LLM costs. There's a diminishing marginal return to tokens

[KptMarchewa]

That's true and it's also very easy to game.

There is increase in (not only perceived) value _somewhere_ - IMO depends on the organizational culture. At my work I found going over 100$ on OpenAI/Anthropic's API pricing does not produce any meaningful additional output. It might be much different in different companies.