A Better Newspaper

## Overview Dario Amodei is CEO of Anthropic. On Saturday, September 12, 2026, he called on AI firms to slow the development of powerful AI technology amid mounting worries over the risks of "superintelligent" computer systems (SCMP, September 12). The call came days after AI researcher Jacob Coxon — who had left OpenAI to join Anthropic — quit the industry entirely, accusing both companies of "gambling with our lives" (SCMP, September 12). Amodei's slowdown push coincides with Anthropic's advanced plans for a Nasdaq IPO, creating what commentators describe as a tension between the company's safety rhetoric and its commercial ambitions (CNBC, September 14). Amodei has since called for AI labs to more deeply embed third-party safety testers, and Anthropic has brought in Accenture for AI safety testing (FT, September 18). Separately, Amodei and Sam Altman have called for global risk evaluation standards for AI, a plea the US government has rejected (BBC, September 23). By late September, Amodei's warnings had entered mainstream popular culture, becoming the subject of a satirical Weekend Update segment on Saturday Night Live (SNL/YouTube, September 27), and market-watchers and commentators began openly questioning whether his safety rhetoric was undermining his own IPO (Jason Calacanis/YouTube, September 27). In early October, Sam Altman publicly distinguished his own risk posture from Amodei's, saying people need to accept that "some bad things" will happen if society wants to enjoy AI's benefits, while maintaining that AI security risks should be taken seriously (SiliconANGLE, October 5). Amodei's warnings were reportedly set out in a 38-page essay, which has prompted outside commentators to debate whether his stance reflects genuine concern or "virtue signaling" (Ethan Mollick, via YouTube interview, October 8). ## Recent Statements - Amodei said: "We must slow the pace at which we improve" AI capabilities (SCMP, September 12). - Amodei reportedly set forth a three-step plan to pace AI development, in what one commentary called "a week of doomsday concerns" (Substack, September 13). - In a new essay posted to his personal blog, Amodei argued for slowing frontier model development; OpenAI CEO Sam Altman and Elon Musk reportedly backed the call (SiliconANGLE, September 14; Arab News, September 12). - Musk reportedly agreed with aspects of Amodei's plan (Substack, September 13). - Amodei says AI labs should more deeply embed third-party testers as concerns grow around risks associated with the technology (FT, September 18). - Amodei, along with OpenAI's Sam Altman and Hugging Face's Clement Delangue, has called for global AI risk evaluation standards (BBC, September 23). - Sam Altman said people need to accept that "some bad things" are going to happen if they want to enjoy the benefits of AI, positioning this view as distinct from Amodei's approach, while saying AI security risks need to be taken seriously (SiliconANGLE, October 5). - Amodei's essay is described as 38 pages long and delivers an "ominous warning" about AI's threat to society, per an interview with analyst Ethan Mollick, who was asked whether Amodei's stance is sincere or "virtue signaling" (YouTube interview, October 8). Mollick's own response was not fully captured in available material. ## Safety Testing Partnership - Anthropic has brought in Accenture for AI safety testing (FT, September 18). - The move follows Amodei's public calls to slow AI development and reflects a broader push for third-party validation of frontier model risks (FT, September 18). ## Global Standards Push - OpenAI's Sam Altman, Anthropic's Dario Amodei, and Hugging Face's Clement Delangue have called for global AI risk evaluation standards (BBC, September 23). - The US government has rejected these pleas from OpenAI and Anthropic for global AI standards (BBC, September 23). ## IPO Context - Anthropic was valued at $965 billion earlier in 2026 and confidentially filed its IPO prospectus in June; it has been widely expected to list shares as soon as October 2026 (CNBC, September 14). - Anthropic is reportedly pursuing a $2 trillion valuation for its Nasdaq listing (CNBC, September 14). - Nvidia is reportedly in talks to invest in Anthropic's IPO, according to sources (Reuters, September 11). - CNBC frames Amodei's slowdown advocacy as creating tension ("walking a tightrope") with the company's simultaneous pursuit of an ambitious IPO valuation (CNBC, September 14). - Market expectations for an Anthropic IPO occurring in 2026 reportedly peaked at 96% earlier in September but had fallen to 76% as of late September, according to commentator Jason Calacanis (Jason Calacanis/YouTube, September 27). Note: the underlying prediction-market source for these figures is not specified in the segment. - Calacanis argued that Amodei founded Anthropic after telling the public there was roughly a 10% chance of AI causing human extinction, and questioned whether Amodei's continued safety rhetoric amounts to "sabotaging" his own IPO and a dereliction of duty to shareholders (Jason Calacanis/YouTube, September 27). This is Calacanis's characterization/opinion, not a confirmed fact about Amodei's fiduciary conduct. ## Cultural Reception - Amodei's warnings about AI's threat to humanity were parodied on NBC's Saturday Night Live in a Weekend Update segment (SNL/YouTube, September 27). - The segment indicates Amodei's public profile and AI-risk messaging have reached mainstream entertainment and late-night comedy, beyond trade press and policy circles. - Commentator Jason Calacanis publicly criticized Amodei on YouTube, suggesting his safety warnings are commercially self-defeating and possibly a breach of duty to Anthropic's shareholders (Jason Calacanis/YouTube, September 27), reflecting growing business-press and commentator scrutiny of the safety-vs-commerce tension first flagged by CNBC (September 14). - In an October 8 YouTube interview, host and guest Ethan Mollick were asked to assess whether Amodei's pessimistic essay reflects "virtue signaling" versus genuine concern about building safer AI, indicating the debate over Amodei's motives has extended into AI-focused commentary and analyst circles (YouTube interview, October 8). ## Altman-Amodei Divergence - Despite earlier reported alignment — Altman reportedly backed Amodei's slowdown essay and joined him in calling for global AI risk evaluation standards (SiliconANGLE, September 14; BBC, September 23) — Altman has since articulated a notably different risk tolerance, telling audiences that people need to accept "some bad things" will happen as a tradeoff for AI's benefits (SiliconANGLE, October 5). - Altman reportedly still maintains that AI security risks need to be taken seriously, suggesting a difference in emphasis/tolerance rather than an outright rejection of safety concerns (SiliconANGLE, October 5). - This marks a public point of divergence between the two CEOs after a period of apparent cross-industry alignment on safety messaging in September. ## Why It Matters As CEO of one of the leading frontier AI labs, Amodei's public statements on slowing AI development carry significant weight in shaping industry norms, regulatory expectations, and public perception of AI risk. The apparent contradiction — advocating a slowdown while pursuing a historic $2 trillion IPO with backing reportedly sought from Nvidia — has drawn scrutiny, especially as rivals Sam Altman and Elon Musk have voiced support for aspects of Amodei's plan, suggesting rare cross-industry alignment on AI safety messaging even amid competitive tensions. Anthropic's move to engage Accenture as a third-party safety tester marks a concrete follow-through on Amodei's rhetoric, potentially setting a precedent other labs may be pressured to follow. The US government's rejection of the industry-led push for global risk evaluation standards (BBC, September 23) signals a gap between lab-level safety advocacy and governmental appetite for binding international coordination. Amodei's appearance as fodder for SNL satire (September 27) and pointed criticism from tech commentator Jason Calacanis (September 27) suggest his warnings have permeated both popular culture and business commentary, with some now framing the safety-vs-commerce tension as a potential threat to the IPO itself. Altman's explicit framing that "some bad things" are an acceptable cost of AI progress (October 5) signals that the earlier appearance of cross-lab consensus on slowing development may be fraying. By October 8, debate over Amodei's motives had reached AI-focused analyst commentary, with questions about whether his lengthy essay reflects sincere concern or strategic positioning ("virtue signaling") — underscoring that both the substance and sincerity of his warnings remain contested as Anthropic's IPO approaches.