Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit

Overview of the Video

Title: Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit

Channel: Bloomberg Originals

Video Link: https://www.youtube.com/watch?v=v1wZwxY3CMg

This 47-minute episode of The Circuit by Bloomberg Originals offers an in-depth, behind-the-scenes look at Anthropic, its co-founders (siblings Dario Amodei and Daniela Amodei), its rapid rise to a sub-trillion-dollar evaluation, and the profound safety, political, and economic dilemmas created by its AI technology (Claude) [00:23], [01:09].

Key Sections & Narrative Outline

1. Origins & The OpenAI Split [03:07]

Background: Dario and Daniela Amodei grew up in San Francisco; D

ario studied neuroscience and math [03:07], while Daniela focused on arts/reading and later joined Stripe [03:53].

OpenAI Era: Dario joined OpenAI in 2016, leading research into scaling laws—the idea that scaling computing power and data predictably improves model intelligence [04:08], [04:34].

Departure: In 2021, Dario, Daniela, and five co-founders left OpenAI due to fundamental disagreements over values, trust, and direction with Sam Altman, founding Anthropic as a safety-focused counterweight [05:07], [06:08].

2. Model Design, Culture, and Values [06:54]

Constitutional AI: Claude is trained according to a set of principles ("constitution") deri

ved from universal human rights declarations and multi-faith principles to maintain "professional warmth" without being sycophantic or cold [07:09], [07:28], [08:25].

Enterprise Focus: Instead of consumer video/social media tools, Anthropic focused heavily on B2B, developer tools, and enterprise applications (e.g., healthcare, energy) [10:04].

3. Enterprise Impact: Claude Code & Claude Co-work [09:40]

Product Breakthroughs: Engineer Boris Cherny shares how products like Claude Code transformed software development by moving from line auto-completion to fully autonomous coding agents [12:04], [12:46].

Internal Impact: Anthropic engineers report usin

g Claude to write up to 100% of their code [13:29].

Market Disruption: Software automation triggered widespread fears in white-collar sectors (dubbed the "SAS Apocalypse" by traders) [11:09], [11:15].

[ Traditional Coding ] ----> Tab Completion (Lines)

[ Claude Code Era ] ----> Autonomous Agents (Full Feature Build)

4. The White-Collar Labor Dilemma [16:50]

Job Disruption: Dario Amodei addresses his controversial prediction that AI could automate a vast percentage of white-collar entry-level jobs [17:22].

Economic Paradox: Discussions center around rapid GDP growth coinciding with structural unemployment or low-wage displacement [17:05].

Future H

uman Roles: Transition toward human-centric, relationship-driven, and physical-world roles (e.g., doctors shifting from diagnosis to bedside care) [21:23], [22:08].

5. Defense, Geopolitics, and Pentagon Clashes [23:24]

Geopolitical Stance: Dario supports strong export controls on AI hardware/chips to China [23:58].

Department of Defense Conflict: Anthropic contracted with the U.S. Pentagon/Palantir for national defense applications [25:02], [25:43]. However, tensions flared when the DoD requested unrestricted guardrails [26:21].

Red Lines: Anthropic drew strict boundaries against using Claude for mass surveillance or autonomous weapons, leading to

friction and blacklisting threats from military/government officials [26:34], [26:48].

6. "Mythos" & The Cyber Security Superweapon [32:54]

Model Discovery: Anthropic quietly developed a hyper-capable model, code-named Mythos, which demonstrated unprecedented abilities in discovering zero-day cybersecurity vulnerabilities across major OS frameworks [33:01], [33:17].

Project Glasswing: Early testers called Mythos a "superweapon" [33:41]. Anthropic withheld a broad commercial release, opting to restrict access to defensive partners to prevent malicious misuse—costing the company commercial revenue in favor of security [33:56], [35:10].

7. Regulation,

Public Backlash, and Existential Risk [36:06]

Policy Shift: Critique of extreme Silicon Valley views fluctuating between total deregulation and sudden calls for nationalization during panic [37:52].

Public Anxiety: Protests outside AI offices and rising public fear over rapid, unchecked technological displacement [38:51], [39:14].

Existential Risk Estimate: Dario reiterates his assessment of a 10–25% existential risk/civilizational challenge posed by advanced AI and argues for strict safety protocols, checks, and balances to minimize it [45:21], [46:24].

Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit

Bloomberg Originals · 2.5M view