Max Child contends that the predictions of 2025 being 'the year of agents' and 'the year of voice' did not materialize.

The release of AI video models VO3 in May and Sora a month prior to the podcast were described as 'minor nuclear explosions' in the consumer technology scene.

The legal AI startup Harvey experienced a significant improvement in performance after integrating the latest generation of 'thinking models' in the last 9-12 months.

Max Child argues that OpenAI's rapid development of agent-related tools like SDKs and its Atlas browser is primarily a competitive defense strategy.

Andrej Karpathy predicted a timeline of approximately 10 years for AI agents to reach fully human-capable levels.

The term 'vibe coding' was coined by Andrej Karpathy in 2025.

Max Child believes that while AI is far from AGI in visual, audio, and multi-step tasks, text-in/text-out models could reach AGI-level thinking within the next two or three model generations.

Synthesia reportedly turned down a $3 billion acquisition offer.

Smart glasses startup Sesame raised $250 million from Sequoia Capital.

Max Child estimates that combined commitments for AI infrastructure projects involving companies like NVIDIA, AMD, and Broadcom have reached approximately $1.5 trillion.

The consumer breakout of OpenAI's Sora model significantly decreased after the company removed the ability to generate content using copyrighted material.

Andrej Karpathy compared the current development state of AI agents to the state of self-driving cars in 2016 or 2017.

In early 2025, some industry prognosticators, including individuals from OpenAI, predicted that AI agents and voice AI would be the two biggest themes of the year.

Max Child identifies the emergence of 'thinking models' with reasoning capabilities, such as O3 and GPT-5, as the most thematically important AI development of 2025.

A rumor circulated approximately 18 months ago that Sam Altman was attempting to raise $7 trillion to scale AI compute infrastructure.