AI tool catches pancreatic cancer in routine scans before symptoms appear
According to physician Zhu Kelei, AI has definitively saved the lives of patients whose scans were only flagged by PANDA, an AI tool developed by Alibaba researchers. The system analyzes non-contrast CT images – scans where even experienced radiologists can easily miss tumors.
"The exponential continues until it doesn't," says Anthropic President Daniela Amodei, quoting her colleagues. At Anthropic, the team believed every year that this pace couldn't possibly keep up, and yet it did, Amodei says in an interview with CNBC TV. But that's not guaranteed, she adds. Anthropic doesn't know the future either and could be wrong about this assumption.
Economically, things get more complicated, Amodei says (from 15:56). Even if the models keep improving, rolling them out in companies can stall for "human reasons": change management takes time, procurement processes move slowly, and specific use cases often remain unclear. The key question for whether AI is in a bubble comes down to whether the economy can absorb the technology as fast as it's advancing, she suggests.
Google engineer says Claude Code built in one hour what her team spent a year on
A senior Google engineer publicly praises Anthropic’s Claude Code: the tool built in one hour what her team spent a year developing. The quality and efficiency gains exceed anything anyone could have imagined, she says. Plus: Claude Code’s creator shares his best workflow tips.
Baidu's AI chip division, Kunlunxin, has confidentially filed for an IPO in Hong Kong. The company submitted its application on January 1, setting the stage for a spin-off while keeping Kunlunxin under Baidu's umbrella. According to Reuters, a recent financing round valued the company at around $3 billion. The final size of the offering hasn't been determined yet.
Kunlunxin started as an internal unit back in 2012 and has primarily supplied chips to Baidu. Over the past two years, though, the company has been expanding its customer base beyond its parent company.
OpenAI has reportedly shifted production of its first AI hardware device from Luxshare to Foxconn. The company wants to avoid manufacturing in China, according to Taiwanese business newspaper Economic Daily News, citing supply chain sources. Instead, production will happen in Vietnam or the US.
The device is still in the design phase and could turn out to be a smart pen or portable audio gadget. It's expected to have a microphone and camera, letting users transfer handwritten notes directly to ChatGPT. The company is aiming for a launch in 2026 or 2027, and according to the newspaper, Foxconn would handle all OpenAI orders, from cloud infrastructure to consumer devices. OpenAI is calling the project "Gumdrop" internally.
OpenAI is building its planned ChatGPT hardware around conversation. To make that work, the company is pouring resources into improving its audio AI models, according to The Information. Over the past two months, OpenAI has combined several internal teams to focus on audio.
Right now, OpenAI's audio models can't match the accuracy and response speed of their text-based counterparts, according to current and former employees. A new audio model architecture in development aims to sound more natural and emotional, deliver more accurate answers, and handle real-time back-and-forth conversation. OpenAI is targeting a release in the first quarter of 2026. Kundan Kumar, a researcher the company recruited from Character.AI, is leading the effort.
Chinese AI startup Moonshot AI, the company behind the Kimi chatbot, has closed a $500 million Series C funding round. The deal values the company at $4.3 billion, according to an exclusive report from LatePost. IDG led the round with $150 million, while existing investors including Alibaba, Tencent, and Wang Huiwen also participated.
CEO Yang Zhilin said in an internal letter that the company now holds more than $1.4 billion in cash. The funds will go toward expanding computing capacity and developing the K3 model. An insider says this cushion means Moonshot AI isn't rushing to go public, unlike Zhipu and MiniMax, two other Chinese AI startups pushing for IPOs. In September, Moonshot AI launched its "OK Computer" agent feature and a subscription model. According to Yang, paying users are growing 170 percent month over month.
This year, Moonshot AI made waves with its Kimi-K2-Thinking model, an open-source reasoning model that held its own against proprietary competitors.
Alibaba has released Qwen-Image-2512, an update to its text-to-image model. The company says the new version does a better job generating realistic images of people, with finer facial detail and less of the artificial look that plagued earlier versions.
The updated image model aims to eliminate the "plastic" look of its predecessor. | Image: Qwen
The display of text in images, for example, in infographics or presentations, has also been improved. Landscapes, animal fur, and other natural elements also come out with finer detail.
The new Qwen model handles text in images better than before. | Image: Qwen
In over 10,000 blind tests on Alibaba's AI Arena platform, Qwen-Image-2512 came in fourth overall, making it the top-ranked open-source model, according to Alibaba. It's up against other open models like HunyuanImage-3.0, Z-image, and Flux.2.
