Chris and Rohan talk about the major areas of progress and improvement for LLMs in the future and the landscape of LLMs in 5 years. The multi-billion dollar question - what will the landscape of LLMs look like in 5 years? Do people expect a handful of dominant LLM Models from OpenAI, Anthropic, Meta, Google, etc., or will there be an ecosystem of LLMs created by every enterprise? They also talk about AI Safety: what are people's concerns from an AI safety standpoint? Do people subscribe to the more general-purpose Alignment?