stop obsessing over LLM benchmark scores.
a model can crush the bar exam, solve math olympiads, and write haikus.
put that exact same model inside an enterprise to design a chip, automate logistics, or run SRE, and it completely freezes on day 1.
benchmarks test closed answers
Today we're launching @arcfront with @Naman_Bhalla and deep support from @asxna.
In under a year, four of the five frontier labs became customers.
In 2025 an AI solved all 12 problems at the ICPC World Finals. The best human team solved 11. I competed in two of those finals,