LLMs are becoming commoditized. Like gas from different stations, models can be swapped based on price or marginal performance. This means competitive advantage doesn't come from the model itself, but how you use it with proprietary data.

On financial analyst benchmarks, top models from Anthropic, Google, and OpenAI are now almost indistinguishable in capability. This convergence suggests the frontier is commoditizing, questioning the return on investment for massive training runs and shifting value up the application stack.

The dramatic improvements from GPT-2 to GPT-4 were driven by a simple law: bigger models and more training data yielded better results. This trend has stopped. Recent attempts to scale even larger models have produced only marginal gains, forcing the industry into more complex, narrow optimizations instead of giant leaps.

For the first time in years, the perceived leap in LLM capabilities has slowed. While models have improved, the cost increase (from $20 to $200/month for top-tier access) is not matched by a proportional increase in practical utility, suggesting a potential plateau or diminishing returns.

Analysis of model performance reveals a distinct shift with GPT-4 and subsequent models. These newer models are much more correlated with each other in the tasks they succeed or fail on compared to the pre-GPT-4 era. This could suggest a convergence in training data, architectures, or agent scaffolding methodologies across different labs.

AI model improvements have shifted from revolutionary leaps (e.g., high-school to college-level intelligence) to marginal gains. The current difference between top models is akin to a PhD student getting an 'A' versus a 'B'—an improvement that is irrelevant for the majority of everyday tasks that a 'college kid' model can handle perfectly well.

For sophisticated users, the performance difference between top AI models is negligible. They are all "so good" that the limiting factor for value creation is no longer the tool's capability, but the user's creativity, time, and attention.

Leading AI models are becoming increasingly similar in capability. This rapid convergence suggests the underlying technology is becoming a commodity, and competitive advantage will likely shift to user interface, distribution, and specific applications rather than the core model itself.

The rapid, step-change improvements in LLMs are likely slowing down. This is because models have already been trained on most of the available internet, and the compute budget required for each incremental improvement is increasing exponentially to an unsustainable degree. A new architectural breakthrough, not just more data and compute, is needed for the next leap.