Google’s new AI training method helps small models tackle complex reasoning

Google's SRL framework provides a step-by-step "curriculum" that makes LLMs more reliable for complex reasoning tasks.
Nature

‘Malicious use is already happening’: machine-learning pioneer on making AI safer

Yoshua Bengio, considered by many to be one of the godfathers of AI, has long been at the forefront of machine-learning ...
Nature

A Chinese AI model taught itself basic physics — what discoveries could it make?

A tool called AI-Newton can derive scientific laws from raw data, but is some way from developing human-like reasoning.