STORY: From when Sam Altman thinks AI will make a leap, to how Meta plans to automate advertising, this is AI Weekly.
:: AI Weekly
Sam Altman is expecting big things in 2026.
Speaking in San Francisco, he said that’s when AI agents could start functioning at a higher level.
“I would bet next year that in some limited cases, at least in some small ways, we start to see agents that can help us discover new knowledge or can figure out solutions to business problems that are kind of very nontrivial. Right now, it’s very much in the category of, okay, if you got something like repetitive cognitive work, we can automate it at a kind of a low level on a short time horizon. And as that expands to longer time horizons and higher and higher levels, you know, at some point you’re going to add a scientist, an AI agent that can go discover new science, and that will be kind of a significant moment in the world.”
DeepSeek has updated its bot for the first time since January, when the debut of the low-cost, high-power model rocked the tech sector.
The Chinese startup said the new version of its R1 reasoning system offered significant improvements.
Benchmark tests showed it was closing the gap with products from OpenAI and Google.
Meta aims to fully automate advertising with AI by the end of next year.
The Wall Street Journal says firms will be able to provide a product image and a budget – and leave bots to make and target ads.
Companies will reportedly be able to personalize marketing, so that viewers see different versions of ads in real time based on factors including their location.
:: ETH Zurich
And a Swiss team has developed a bot that can play badminton against humans.
The system uses so-called reinforcement learning – a type of AI that improves by repeating tasks.
Researchers say it paves the way for bots that can do other roles like search and rescue.
