China’s AI is catching cancers early

AI finds pancreatic cancer early

A routine CT scan caught a cancer no one was looking for.

At a hospital in Ningbo, China, doctors are using an AI system called PANDA to scan ordinary, low-radiation CTs for pancreatic cancer. Since late 2024, it has reviewed more than 180,000 scans and flagged about two dozen cancers, including 14 early-stage cases that likely would have been missed.

This is one of the clearest examples yet of AI doing quiet, unglamorous medical work that actually changes outcomes. A 93 percent detection rate in trials is impressive, but the real story is scale, since China already runs millions of CT scans a year at about $25 each.

The trade-off is fear and follow-ups: about 1,400 alerts led to only 300 real concerns, stretching staff and patient trust. Still, in hospitals without enough specialists, this kind of system starts to look less like hype and more like basic infrastructure.

If software can quietly save lives like this, who decides where it gets deployed first?

AI moves into event photo booths

Event photo booths are turning into content machines.

A Los Angeles company called TikTok Photo Booth is rolling out AI-powered photo and video booths that turn guest captures into sketches, caricatures, and character-style visuals on the spot. The system is built for vertical video and short clips, aimed squarely at TikTok, Instagram Reels, and similar platforms.

This isn’t a tech breakthrough so much as a market signal about expectations. People now assume events will double as content factories, and vendors are responding with AI tools that feel fast, familiar, and low-friction rather than novel.

For brands, this becomes a cheap way to generate hundreds of branded posts in a single night. For guests, it raises the bar, since a plain photo booth already feels dated when AI versions show up instantly on a phone screen.

Once every event produces shareable content by default, what actually counts as memorable anymore?

Data centers might leave Earth

Some AI leaders think land and power are running out.

Google says it will test a space data center project in 2027, and Elon Musk claims training advanced AI will be cheaper in orbit within five years. The pitch is simple: nonstop solar energy, fewer regulations, and no local pushback, even as OpenAI alone commits $1.4 trillion to data center buildouts on Earth.

This reads less like a plan and more like pressure venting from an industry that keeps promising scale without limits. The physics and costs still look brutal, with launches at about $2,000 per kilogram today and servers weighing over 1,000 kilograms each.

If launch costs really drop toward $200 per kilogram in the 2030s, some version of this could move from talk to pilot projects. Until then, space data centers mostly signal how boxed in big AI players feel by energy grids, water shortages, and local politics back home.

When the answer to growth is orbit, it’s worth asking what got ignored on Earth.

Prohuman team

Free Guides

Explore our free guides and products to get into AI and master it.

All of them are free to access and would stay free for you.

Feeling generous?

You know someone who loves breakthroughs as much as you do.

Share The Prohuman it’s how smart people stay one update ahead.