Author HubMatthias Bastian

Online journalist Matthias is the co-founder and publisher of THE DECODER. He believes that artificial intelligence will fundamentally change the relationship between humans and computers.
AI research

Deepseek's $5.6M Chinese LLM wonder shakes up the AI elite

AI research

Apple AI researchers question OpenAI's claims about o1's reasoning capabilities

AI research

Study reveals major reasoning flaws in smaller AI language models

Ad
AI in practice
Short

TSMC CEO C.C. Wei says OpenAI CEO Sam Altman's plans to build new fabs for AI chips were "too aggressive" for his company. Altman reportedly proposed last year to build about three dozen fabs to meet the growing demand for AI services. TSMC declined, saying Altman's proposed number of fabs was too high and that it feared it would not be able to operate them at the required utilization rate of at least 80 percent. TSMC's projections at the time did not see enough demand for more than 30 new fabs. In January, the Financial Times and Bloomberg reported that Altman was planning a global chip network and was talking to several partners, including TSMC.

AI in practice
Short

Elon Musk rerouted Nvidia AI chips from Tesla to his AI company, xAI. Internal Nvidia emails obtained by CNBC show that Musk ordered the delivery of 12,000 GPUs with priority to X and xAI instead of Tesla. According to CNBC, this delays the delivery of more than $500 million worth of chips to Tesla by months, potentially hindering the development of the infrastructure needed for autonomous driving and robotics. In a post on X, Musk explained that "Tesla had no place to send the Nvidia chips to turn them on, so they would have just sat in a warehouse." Redirecting the scarce chips illustrates Musk's conflict of interest between Tesla and his AI ambitions. Musk has demanded more Tesla stock in the past, at least 25%, to invest more in AI within the company. It's all about controlling a potential key technology: Musk is the sole ruler of xAI.

Ad
Ad
Ad
Ad
AI in practice
Short

Microsoft is investing $3.2 billion over two years in cloud and AI infrastructure in Sweden. The company will deploy up to 20,000 Nvidia graphics processors, likely AMD processors, and possibly its own chips in its data centers to meet the growing demand for generative AI, Reuters reports. Microsoft also plans to train 250,000 Swedes in AI skills and invest in local renewable energy. President Brad Smith announced additional investments for this fall. The investment joins multi-billion dollar investments in Germany, Spain, Japan, Southeast Asia and the U.S., totaling more than $15 billion.

Ad
Ad
Google News

Bank details

IBAN: DE87 1203 0000 1086 0070 75
Account holder: DEEP CONTENT GbR
Purpose: Support THE DECODER
AI in practice

Meta's LibGen controversy reveals how desperate AI companies are for quality training data

AI in practice

The great AI scaling debate continues into 2025

AI research

Deepseek's $5.6M Chinese LLM wonder shakes up the AI elite

Google News