Tech Brew Ride Home: "Anthropic’s Laundry Has Been Aired"
Date: April 1, 2026
Host: Brian McCullough (Morning Brew)
Episode Overview
This episode dives into the latest high-stakes funding news from OpenAI, shifting investor sentiment toward its competitor Anthropic, and the fallout from Anthropic's accidental code leak. Also covered are recent issues with Baidu's robo-taxis in China and a new Gmail feature letting users change embarrassing email addresses.
Key Discussion Points & Insights
1. OpenAI’s Mega-Funding Round
[00:34 - 03:30]
- OpenAI secured $122B in committed capital—$12B above initial expectations—at an $852B post-money valuation.
- Major new investments: Amazon ($50B), Nvidia ($30B), SoftBank ($30B), plus a notable $3B from individual/retail investors via bank channels.
- Monthly revenue stands at $2B; company made $13.1B last year and is not profitable.
- Enterprise sales now make up 40% of revenue, on track to equal consumer sales by end of 2026.
- Retail investors are expected to play a significant role in upcoming IPOs (OpenAI, Anthropic, and SpaceX).
Notable Quote:
“Fryer said broadening access was consistent with the company's mission of ensuring powerful AI was created for the benefit of humanity. That meant not just access to the technology, but access to the financial upside.”
— Brian McCullough quoting OpenAI CFO Sarah Fryer [02:41]
2. Investor Pivot: OpenAI vs. Anthropic
[03:33 - 06:45]
- Secondary market for OpenAI shares is cooling; investors now pivoting sharply toward Anthropic.
- Some $600M in OpenAI shares recently unsold; conversely, buyers have $2B ready for Anthropic.
- Anthropic’s enterprise focus, lower operating costs, and strong growth narrative make it an attractive investment.
- Anthropic shares currently valued at up to $600B—a significant surge from its previous funding round.
- Demand has led to unprecedented bids, with platforms like Hive registering over $1.6B in demand.
Notable Quotes:
“We literally couldn't find anyone in our pool of hundreds of institutional investors to take these shares.”
— Brian McCullough quoting Ken Smith (Next Round Capital) [04:25]
“The demand is one of the highest we've ever seen… It's essentially unlimited interest.”
— Brian McCullough quoting Adam Crawley (Augment) [06:00]
3. Anthropic Code Leak: The Fallout
[06:45 - 09:35]
- Anthropic accidentally exposed underlying instructions ("harness") for its Claude code AI agent app.
- Over 8,000 copies/adaptations removed via copyright takedown on GitHub.
- No customer data or model weights exposed, but trade secrets and proprietary coding agents' instructions were leaked.
- Risks:
- Competitors can now clone and fast-follow on unique Claude features.
- Potential for hackers to exploit vulnerabilities.
- Leaked features included “Dreaming” (memory consolidation), "Buddy" (a Tamagotchi-style pet), undercover coding capabilities, and tags for upcoming product releases.
- Upcoming “Kairos” update leaked: brings proactive agent behavior, mobile update notifications, and visual coding avatars (“buddies”), aiming to generate buzz and rivalry with OpenAI’s OpenClaw product.
Notable Quotes:
"The leak is a blow for Anthropic because it risks both undermining its reputation for safety and also revealing valuable trade secrets in the pitched battle for enterprise customers."
— Brian McCullough [07:52]
“Programmers… marveled on social media at some of Anthropic's tricks… One feature asks the models to go back periodically through tasks and consolidate their memories, a process it calls ‘Dreaming’.”
— Brian McCullough [08:47]
4. Baidu Robo-Taxi Outage in Wuhan
[12:15 - 14:15]
- Baidu’s Apollo Go self-driving taxis in Wuhan suffered widespread technical failures, trapping passengers inside and causing major traffic disruptions.
- Passengers stranded for up to 90 minutes; support lines were unresponsive and onboard SOS systems failed.
- No injuries, but reports of substantial inconvenience and risk.
Notable Quotes:
“I tried every way I could think of to call for help... but the phone line wouldn't go through and when I pressed the SOS button it told me it was unavailable. So then, what exactly is SOS for?”
— Rednote user, as quoted by Brian McCullough [13:48]
5. New Gmail Feature: Username Changes
[14:15 - 15:15]
- US Gmail users can now change the portion before “@gmail.com” once every 12 months.
- Old addresses remain linked for continued delivery.
- Useful for personal rebranding or after a legal name change, but has limitations (can’t reuse addresses on other accounts, limited frequency).
Notable Quotes:
“If you've been stuck with a regrettable Gmail username you picked years ago, you might finally have a way out.”
— Brian McCullough quoting The Verge [14:26]
Memorable Moments & Commentary
- On AI Funding Hype:
The episode emphasizes how quickly investor interest can swing in the AI space, with Anthropic’s cool, enterprise-focused credentials now seeming to “steal the thunder” from OpenAI’s more consumer-driven momentum. - On the Code Leak:
The host notes how the industry’s transparency dilemma—balancing openness with competitive secrecy—is playing out in real time, with accidental leaks reshaping both product roadmaps and competitive dynamics.
Segment Timestamps
| Segment | Topic | Timestamp | |-------------|---------------------------------------------------------------|--------------| | 1 | OpenAI closes monster round | 00:34–03:30 | | 2 | Investors pivot to Anthropic; OpenAI demand cools | 03:33–06:45 | | 3 | Anthropic code leak & leaked features | 06:45–09:35 | | 4 | Baidu robo-taxi failures in Wuhan | 12:15–14:15 | | 5 | Gmail username change rollout | 14:15–15:15 |
Summary Note:
The episode comprehensively illustrates the rapidly shifting tides in the tech and AI industry—from major capital raises and investment sentiment to the real-world policy and security headaches of running massive AI and autonomous system platforms. For anyone tracking the next generation of tech giants, this is a can't-miss briefing.
