Scaling up LLMs is only going to go so far, and it will yield diminishing marginal returns on all of that money, compute, and data. It’s a regime of exponential

Scaling up LLMs is only going to go so far, and it will yield diminishing marginal returns on all of that money, compute, and data. It’s a regime of exponential increases in inputs for linear gains in the outputs - barring some technological breakthroughs which could come from anywhere, not just from OpenAI.