We started working with OpenAI on language models in 2019 and with the Department of Defense on AI applications in 2020, well before the recent AI craze.

At OpenAI, we pursued robotics not for commercial motives, but to demonstrate the power of machine learning, which later influenced large language models.

Many startup founders are already using AI coding tools, with a significant increase in adoption from 0% to about 25-50% in just a few months, indicating a shift in how software is developed.

OpenAI's latest model, o1, is gaining traction among developers, with 15% of the summer batch already using it despite its recent release.

Elon Musk mentioned that we're seeing a 100x increase in computational and algorithmic efficiencies per year, which aligns with the progress we've made in large language models over the past two years.

The development of large language models (LLMs) like GPT-2 marked a significant turning point for OpenAI, showcasing the potential of AI.

The biggest change in AI research over the past year is that reinforcement learning (RL) in language models has finally proven effective, demonstrating expert human reliability and performance with the right feedback loop.

A software development company expanded its AI feature rollout from 100,000 developers to 25 million in three months, highlighting the fascinating speed of scaling in AI.

Large language models are becoming significantly larger and smarter, with AI labs adopting a strategy of scaling more parameters, data, and compute to improve model performance.