Andrew Ng Criticizes AI Extinction Theory: More Like Sci-Fi, the New Wave of Hype May Aim to Influence Regulation

According to Bloomberg, renowned expert in the field of artificial intelligence, Andrew Ng, criticized the warnings about "AI existential risks" raised by researchers at some top model developers. He believes that concerns about AI potentially leading to human extinction are closer to science fiction than rigorous scientific conclusions, and may hinder ensuring that the technology brings maximum benefits to society.

As co-founder of Google Brain and the education platform Coursera, Ng pointed out that in the early days of the AI boom, the tech industry had exaggerated the potential catastrophic harms of AI to increase publicity, with the obvious purpose of influencing regulatory direction. In an interview with Bloomberg Television, he said he had felt frustrated over the past two weeks, believing that this wave of public relations campaign is resurging again, possibly for similar purposes.

Coxon's Resignation Sparks New Debate

Recently, the high-profile resignation of Jacob Coxon, a former researcher at Anthropic and OpenAI, has drawn widespread public attention to AI safety issues. Coxon accused his two former employers of "betting human lives," as they raced to develop superintelligent AI; he also said that some AI developers believe the technology could pose a lethal threat to humans within a decade.

In this context, several AI leaders, including Dario Amodei, CEO of Anthropic, and Sam Altman, head of OpenAI, have called on the industry to slow down the development of cutting-edge models and take more measures to ensure future safety. Ng previously worked with Amodei and Altman.

Although Ng acknowledged that AI does indeed carry risks, including cybersecurity threats, he emphasized that concerns about AI causing human extinction are "more like science fiction than science."

Unlike the "slowing down" approach, Ng believes the industry should continue advancing AI development while addressing verifiable safety issues through controlled testing, engineering safeguards, and iterative improvements, and applying AI more widely in scenarios that generate social value. He suggests the industry focus on solving practical engineering problems to continuously improve and protect the technology.

Anthropic Uses AI to Build AI: Claude Has Taken Over a Quarter of the Company's R&D Work

Anthropic revealed that as of August this year, Claude has taken over 26% of the company's AI development tasks, while this proportion was less than 1% in February, showing an astonishing growth rate. Its sixth-level automation framework shows that over 90% of tasks are human-AI collaborations, but none are fully autonomous; level AL4 is AI-led, with humans only setting high-level goals and supervising and reviewing.

Anthropic First Quantifies the Progress of AI Development Automation: Claude Leads 26% of Development, Shifting Pressure to Competitors Like OpenAI

Anthropic first quantitatively disclosed AI development automation: the proportion of internal AI development led by Claude increased from less than 1% in February to 26% in August; the internal Agent platform has about 30,000 active agents, generating over 1 billion decisions in August. This is seen as a key signal that recursive self-improvement is moving from concept to engineering reality.

Andrew Ng responds to the 'AI apocalypse' theory: Don't be swayed by science fiction narratives, focus on real engineering issues

On September 17, Andrew Ng told Bloomberg that warnings from top model developers about AI posing a risk to human survival are more like science fiction than science, and may distract from real AI safety concerns. He pointed out that the tech industry once exaggerated the catastrophic risks of AI to gain attention and influence regulation, and the recent surge in discussions may have similar motives. Andrew Ng co-founded Google Brain and Coursera.

Anthropic Launches Life Sciences Certification Program, Integrates Three Main Models into Unified Licensing Terms

Anthropic launched a Life Sciences Certification Program, incorporating three main models, Mythos 5.1, Opus 5, and Sonnet 5, into unified licensing terms, clearly defining their usage boundaries and responsibilities in life sciences scenarios such as drug development, literature review, and experimental assistance, providing pharmaceutical companies, biotechnology firms, and research institutions with a clear compliance pathway and reducing regulatory uncertainty.

Anthropic Reconstructs the Claude Code Project Function: Multi-threaded Parallel Task Execution

Anthropic reconstructs Claude Code and introduces a multi-threaded parallel collaboration mechanism. Users only need to describe the project goal, and the coordinator will split the task into multiple independent threads, which run in separate cloud sessions. Users can track progress in real-time through the main chat room or individual threads, and mobile support is also available; threads can independently submit pull requests and run tests. During execution, Claude continuously builds shared memory between threads, and the project library centrally collects user-uploaded files.