Today's AI Daily Brief covers major funding developments, chip supply constraints, and a significant debate about optimal formats for human-AI interaction. The episode features analysis of Anthropic's potential trillion-dollar fundraising round, Cerebrus's oversubscribed IPO, and supply chain shifts affecting TSMC, Apple, and Intel.

The main discussion centers on The Unreasonable Effectiveness of HTML by Tariq Shahippar from Anthropic, which argues for switching from Markdown to HTML when working with AI agents. This technical essay gained 10 million views and sparked industry-wide conversation about how the nature of knowledge work is fundamentally changing in the AI agent era.

Trillion-Dollar AI Valuations and Market Dynamics

Anthropic is weighing another funding round at almost $1 trillion valuation, potentially raising $50 billion and valuing the company at $900 billion pre-money, which would surpass OpenAI's $852 billion March valuation.

A pre-IPO instrument trading on Jupiter blockchain was valued at $1.2 trillion implied valuation about a week ago, with one investor noting 'people are ready to throw any dollar amount at Anthropic.'

Cerebrus is considering boosting IPO price from $115-125 per share to $150-160 per share after drawing orders for 20 times the number of available shares, potentially increasing valuation from $26 billion to over $34 billion.

Semiconductor Supply Chain Constraints and Shifts

TSMC reported 17.5% annualized sales growth for April, the slowest pace since October and roughly half of analyst forecasts, despite frenetic AI buildout pace.

Intel CEO noted 'TSMC is sold out. There's no choice' when discussing diversification, highlighting capacity constraints at advanced AI chip manufacturing facilities.

Apple signed preliminary chip-making agreement with Intel after over a year of negotiations, with the White House pressuring Apple to contract with Intel following government investment.

Major housing developers including Pulte Group are testing micro-data centers on newly built homes with NVIDIA and California startup Span to address compute shortage through distributed computing clusters.

The HTML vs Markdown Debate for AI Agent Interaction

The Unreasonable Effectiveness of HTML by Tariq Shahippar argues HTML provides superior information density, visual clarity, and ease of sharing compared to Markdown for AI agent workflows.

Shahippar writes 'I find it difficult to read a Markdown file of more than 100 lines. I want richer visualizations, color, and diagrams, and I want to be able to share them easily.'

Critics argue HTML will cost significantly more tokens than Markdown, with Josh Dawes noting 'the cynic in me can't help but note that HTML will cost way more tokens than markdown.'

The debate reflects a fundamental shift where 'the atomic unit of knowledge work has shifted' from producing final outputs to staging conditions for agents to do the work.

The Changing Nature of Knowledge Work in the Agent Era

Knowledge workers increasingly operate in 'liminal space' where the job shifts from going 'blank page to finished goal' to establishing and maintaining conditions for AI agents to finish work well.

Projects now exist in states of 'mixed doneness' where some parts are locked and non-negotiable, others are totally open for agent exploration, and some are semi-decided but still provisional.

HTML's native features like tabs, progressive disclosure, color-coded status, and interactive elements can encode different states of project completion without requiring meta-commentary within documents.

The new skill involves 'calibrating how much structure to impose so that the unstructured remainder is something the agent can actually productively resolve.'