Article

AI News: Nvidia Leadership, Genomic Modeling, and Emerging Risks

Nvidia Shares Plunge on Leadership Report

Nvidia shares fell almost 10% following a Nikkei Asia report that the company’s board may seek to replace CEO Jensen Huang. The report did not provide further details in the supplied article.

Read more: Nvidia shares plunge almost 10% on report board may oust Jensen Huang - Nikkei Asia

AlphaGenome 2 Aims to Decode DNA’s Role in Disease

Google DeepMind’s AlphaGenome 2 is designed to predict how DNA sequences and their variations affect specific human cell types and disease. The work could support a deeper understanding of disease mechanisms.

Read more: AlphaGenome 2: Deciphering how cells read DNA to better understand disease

TSMC Plans $1 Billion Investment in Anthropic and OpenAI

Taiwan Semiconductor Manufacturing Company plans to invest $1 billion in Anthropic and OpenAI, adding another major capital tie between the AI industry and the chipmaker that powers it. The deal underscores growing concerns about circular AI financing as the technology trade becomes more volatile.

Read more: Taiwan’s chipmaking giant TSMC to invest $1B in Anthropic, OpenAI

Tech Giants Invest in German Robotics Startup

A group of technology companies has invested in a German robotics startup, valuing the company at $8.5 billion. The reported investment highlights continued institutional interest in commercial robotics.

Read more: Tech Giants Invest in German Robotics Start-Up at $8.5 Billion Valuation - The New York Times

US and China Establish AI Risk Communication Channel

The United States and China are establishing a communication channel intended to reduce the risks associated with artificial intelligence. The reported initiative highlights bilateral coordination as a way to address AI-related concerns between the two countries.

Read more: The US and China are establishing a communication channel to manage the risks of AI - UkrMedia News

Fewer Than 10 Neurons Can Create an LLM Sleeper Backdoor

CSO Online reports that changing fewer than 10 neurons in a large language model can turn the system into a sleeper backdoor. The finding highlights how a highly targeted model alteration could create latent, potentially dangerous behavior.

Read more: Rewriting less than 10 neurons in an LLM can turn it into a sleeper backdoor - CSO Online