In a few hours working with a deep-learning system, an AI has designed a new wireless computer chip, a feat that would have taken human engineers weeks to carry out, according to engineers at Princeton University’s engineering school and the Indian Institute of Technology.
“Physical AI” and related automation is coming faster than most realize, and it adds another dimension to the threat of AI-fueled human job Replacism that we have long forecast.
The surveillance needs AI to know and feed off every observable phenomenon, in order to advance and bring in profits, something we have long warned, made news again this past week.
Meta will jettison about 4,000 employees, or roughly 5 percent of its workforce, as the company devotes more focus to AI.
At last week’s two-day Artificial Intelligence Action summit in Paris, something was missing that had been present at past gatherings in Britain and South Korea: extensive discussions of AI’s risks and how to mitigate them.
AI models share another human trait: without constant maintenance, their mental faculties can deteriorate.
China’s DeepSeek AI was trained for $6 million using second-tier chips, the company says.
Alphabet, Amazon, Meta, and Microsoft together are planning to spend $320 billion this year to expand their AI data centers and continue building bigger, more powerful large language model AIs, according to the four companies’ announced plans.
In 2024, Chinese researchers filed 300,510 AI-related patents, while scientists in the world’s other countries 188,454, according to data compiled by Triangle IP, a firm that helps companies manage their intellectual property.
I recently interviewed Gabriel Custodiet of The Watchman Privacy Podcast, and escapethetechnocracy.com.