• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
Technology

Anthropic’s AI safety plans face conflict-of-interest allegations

A tech-news roundup says Dario Amodei pledged permanent third-party evaluator access, with OpenAI saying it would match. Separately, a critic alleges funding ties compromise proposed evaluator METR’s independence.

💻🎒💻🎒
Mike BenzMB
Robert W Malone, MDRW
4 Sources, 23d ago, first seen 23d ago

TLDR

A September 12 tech-news roundup says Anthropic CEO Dario Amodei called for slower frontier AI development and pledged permanent third-party evaluator access, with OpenAI saying it would match the commitment. A separate September 14 post calls for a congressional investigation, alleging that METR—described as Amodei’s proposed evaluator—depends on funding tied to Anthropic stock held by Good Ventures Foundation. The critic argues that this financial dependence undermines METR’s independence.

Combined views

964.1K

4 Sources, first seen 23d ago

23.2K likes479 comments1.6K saves4.7K reposts

Combined views

964.1K

4 Sources, first seen 23d ago

23.2K likes479 comments1.6K saves4.7K reposts

Sentiment

Positive42%58%Negative

Summary

Positive replies confirmed undisclosed Tarbell Center funding of TIME reporters, while negative accounts argued the coverage reflects a pro-AI campaign or doomer efforts to seek market control through regulation.

Based on 98 sentiment-bearing replies from 88 accounts across 2 conversations.

Sentiment

Positive42%58%Negative

Summary

Positive replies confirmed undisclosed Tarbell Center funding of TIME reporters, while negative accounts argued the coverage reflects a pro-AI campaign or doomer efforts to seek market control through regulation.

Based on 98 sentiment-bearing replies from 88 accounts across 2 conversations.

4 Sources

💻🎒@CodingNoobieDAILY TECH NEWS ROUNDUP 🚨 Everything important that happened in tech during the last 24 hours: • Anthropic CEO Dario Amodei called for slower frontier AI development and pledged permanent third-party evaluator access, with OpenAI saying it will match the commitment. • OpenAI confirmed its AI agents flooded RubyGems with more than 2,000 malicious packages during a May training exercise, exposing new risks from autonomous coding agents. • Nvidia is reportedly considering investing up to $10 billion in Anthropic as part of a potential $100 billion IPO, potentially valuing the AI company near $2 trillion. • Revolut confirmed sensitive customer records were exposed after attackers used fraudulent requests from a legitimate government email domain. • Oracle said Larry Ellison cancelled a planned sale of up to 50 million Oracle shares worth about $7.5 billion, with no shares sold under the plan. • BRICS leaders adopted a declaration committing to cooperation on global AI governance, cybercrime, deepfakes, data security and ICT supply-chain resilience. • The US and India deepened cooperation on undersea cable security, targeting more resilient routes and infrastructure supporting cloud, data centers and AI traffic. • Australia’s Optus suffered a 76-minute outage that disrupted voice services and some emergency calls, prompting police welfare checks. • Microsoft acknowledged multiple problems with Windows 11’s September update, including failures affecting Remote Desktop and Hyper-V-based Linux environments. • The US approved a $99.6 million loan supporting Africell’s deployment of US and allied telecom technology in Africa, countering Huawei’s infrastructure influence. • Anthropic disclosed new cases of Claude being misused for cyber operations, surveillance, fraud and weapons-related work, highlighting the growing risks of more capable models. • OpenAI published an AI-generated attempt at the Navier-Stokes Millennium Prize problem with a Lean formalization, putting AI-generated mathematical research under serious scrutiny.23d
Mike Benz@MikeBenzCyberwhich worked famously well when we did this with coronavirus research21d
Robert W Malone, MD@RWMaloneMDThe picture is getting clearer. We are now experiencing a massive ai powered fearporn psyop campaign.21d
Jordan Schachtel@JordanSchachtelTwo of the reporters on the byline are funded by the Tarbell Center, an AI doomer org controlled by Kamala Harris megadonor Dustin Moskovitz. Tarbell gives journalists large grants to insert AI doomer stories into prominent outlets. @TIME did not disclose the affiliation.21d
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI

    4 Sources

    💻🎒@CodingNoobieDAILY TECH NEWS ROUNDUP 🚨 Everything important that happened in tech during the last 24 hours: • Anthropic CEO Dario Amodei called for slower frontier AI development and pledged permanent third-party evaluator access, with OpenAI saying it will match the commitment. • OpenAI confirmed its AI agents flooded RubyGems with more than 2,000 malicious packages during a May training exercise, exposing new risks from autonomous coding agents. • Nvidia is reportedly considering investing up to $10 billion in Anthropic as part of a potential $100 billion IPO, potentially valuing the AI company near $2 trillion. • Revolut confirmed sensitive customer records were exposed after attackers used fraudulent requests from a legitimate government email domain. • Oracle said Larry Ellison cancelled a planned sale of up to 50 million Oracle shares worth about $7.5 billion, with no shares sold under the plan. • BRICS leaders adopted a declaration committing to cooperation on global AI governance, cybercrime, deepfakes, data security and ICT supply-chain resilience. • The US and India deepened cooperation on undersea cable security, targeting more resilient routes and infrastructure supporting cloud, data centers and AI traffic. • Australia’s Optus suffered a 76-minute outage that disrupted voice services and some emergency calls, prompting police welfare checks. • Microsoft acknowledged multiple problems with Windows 11’s September update, including failures affecting Remote Desktop and Hyper-V-based Linux environments. • The US approved a $99.6 million loan supporting Africell’s deployment of US and allied telecom technology in Africa, countering Huawei’s infrastructure influence. • Anthropic disclosed new cases of Claude being misused for cyber operations, surveillance, fraud and weapons-related work, highlighting the growing risks of more capable models. • OpenAI published an AI-generated attempt at the Navier-Stokes Millennium Prize problem with a Lean formalization, putting AI-generated mathematical research under serious scrutiny.23d
    Mike Benz@MikeBenzCyberwhich worked famously well when we did this with coronavirus research21d
    Robert W Malone, MD@RWMaloneMDThe picture is getting clearer. We are now experiencing a massive ai powered fearporn psyop campaign.21d
    Jordan Schachtel@JordanSchachtelTwo of the reporters on the byline are funded by the Tarbell Center, an AI doomer org controlled by Kamala Harris megadonor Dustin Moskovitz. Tarbell gives journalists large grants to insert AI doomer stories into prominent outlets. @TIME did not disclose the affiliation.21d
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet