Ai

GPT-5 launched yesterday. 94.6% on AIME 2025. 74.9% on SWE-bench.

As we approach the upper bounds of these benchmarks, they die.

What makes GPT-5 and the next generation of models revolutionary isn’t their knowledge. It’s knowing how to act. For GPT-5 this happens at two levels. First, deciding which model to use. But second, and more importantly, through tool calling.

We’ve been living in an era where LLMs mastered knowledge retrieval & reassembly. Consumer search & coding, the initial killer applications, are fundamentally knowledge retrieval challenges. Both organize existing information in new ways.

The most successful AI startups of 2024 shared an unlikely secret: they didn't rely on proprietary datasets. Instead, they leveraged synthetic data to outmaneuver competitors who were still chasing exclusive data partnerships and expensive labeling operations.

The numbers tell a compelling story. Synthesis AI grew 410.6% last year while Datagen raised $72M – the largest funding round in the synthetic data space. Meanwhile, companies burning millions on human data labeling watched their unit economics deteriorate as synthetic alternatives delivered 500-1000x cost reductions.