TheSilkLake, Self Learning Factories, Sinking City & More!
Under the cold water of Lake Issyk-Kul in Kyrgyzstan, divers are exploring the remains of an entire medieval settlement that vanished beneath the lake roughly 800 years ago. The city once sat directly on the Silk Road. The enormous trading network connecting China, Central Asia, the Middle East and Europe. Researchers think it may have covered as much as five square kilometres. Weirdly, we don't even know what the city was called or how many people lived there.
The latest dives have revealed pottery, animal bones, two cemeteries and the remains of a decorated public building that may have been a mosque, bathhouse or religious school. They even recovered a child's chamber pot, which I love because it takes something that sounds like a lost civilisation and suddenly makes it feel very human. These were people trading across continents, raising families and going about everyday life, and then at some point, their city disappeared underwater.
Scientists think an earthquake or landslide may have altered the lake and submerged the settlement, but the exact cause is still uncertain. One of the expedition leaders has compared the site to Pompeii: a functioning community suddenly destroyed by nature and preserved for centuries afterwards. Except this time, archaeologists have to put on scuba gear to reach it.
Today's Nobel Prize in Physics went to Francis Halzen for an idea that sounds slightly ridiculous until you realise it actually worked: use the ice beneath the South Pole as an enormous telescope.
The IceCube Neutrino Observatory contains more than 5,000 sensors buried as deep as 2.5 kilometres beneath Antarctica, spread through roughly a cubic kilometre of ice. They're not looking for ordinary light. They're waiting for neutrinos. Tiny particles sometimes nicknamed “ghost particles” because trillions can pass through matter without interacting with much of anything at all.
That makes neutrinos incredibly difficult to catch, but also incredibly useful. Unlike light, they can escape from some of the most violent places in the universe and travel enormous distances without being significantly deflected. IceCube has already detected high-energy neutrinos arriving from beyond our Milky Way, potentially giving astronomers a completely different way of studying exploding stars, black holes, neutron stars and events we may not even understand yet.
And they're making it bigger. IceCube-Gen2, planned for around 2033, would expand the instrument to roughly eight cubic kilometres and could detect around ten times as many neutrinos. Basically, humans drilled thousands of sensors into Antarctic ice and accidentally created a new way of looking at the universe.
Traditional factory robots are incredible at doing the same carefully programmed movement thousands of times. Move this panel. Weld here. Pick that box up. Repeat forever.
A new generation is trying to move beyond that. German robotics company RobCo, which has just hit a valuation of $1 billion, is preparing to launch a two-armed robot called Alfie designed specifically for industrial jobs that are too varied or complicated for traditional automation. The interesting bit is that RobCo describes Alfie as self-learning, rather than something that has to be painstakingly programmed for every single movement.
And this isn't just a CGI robot they've announced for five years from now. RobCo says prototype Alfies are already operating with customers, with the wider launch planned for March 2027. The company has raised nearly $200 million, and its growth is part of a much bigger race among European and American companies to build robots capable of operating in messy, human-designed environments rather than perfectly organised assembly lines.
That's where robotics gets properly interesting to me. The real shift isn't a robot that can perform one impressive trick on stage. It's a machine you can put inside a normal factory and effectively tell: “Here's the job. Work it out.”
AI Is Starting to Change the Hacker
AI becoming better at coding always had an uncomfortable flip side: if it can help people build software faster, it can potentially help someone attack it faster too.
South Korea's president said today that investigators have found signs that AI models appear to have been used in recent cyberattacks against commercial banks. Police are now investigating breaches affecting major lenders including Shinhan Bank, KB Kookmin Bank, Hana Bank and Woori Bank, some of which resulted in customers' personal information being exposed. Authorities haven't yet revealed exactly which AI systems were involved or precisely what they did, so it's important not to jump ahead of the investigation.
But the wider trend is difficult to ignore. Reuters notes that Australia said an AI agent gained unauthorised access to a government health-data portal earlier this year, while researchers also recently reported an AI-driven attempt against a Canadian government site that apparently failed. South Korean authorities are now openly saying cybersecurity needs to adapt to an “AI era.”
We've spent the last few years asking what AI can write, draw and create. A much more serious question is beginning to emerge:
What happens when AI becomes good at breaking things too?
Which country has the most volcanoes?
A) Japan
B) Indonesia
C) United States
D) Iceland
Answer: C — United States.
If you count potentially active volcanoes across places like Alaska, Hawaii and the Pacific Northwest, the United States has more than any other country.
Alaska alone contains a huge chain of volcanoes because it sits along the Pacific Ring of Fire.
So while Indonesia is famous for volcanic activity, the US actually takes the crown by total number.
Which animal can go the longest without drinking water?
A) Camel
B) Kangaroo rat
C) Desert tortoise
D) Fennec fox
Answer: B — Kangaroo rat.
Kangaroo rats can survive without ever drinking liquid water.
They get the moisture they need from seeds and from water produced inside their bodies during metabolism.
Their kidneys are also incredibly efficient, meaning they lose very little water.
Basically, evolution built an animal perfectly suited for a place where finding a drink is not guaranteed.
Which city is sinking faster than almost any other major city on Earth?
A) Venice
B) Jakarta
C) Amsterdam
D) Bangkok
Answer: B — Jakarta.
Parts of Jakarta have been sinking by several centimetres per year, largely because groundwater has been pumped out from beneath the city for decades.
At the same time, sea levels are rising.
That combination is one reason Indonesia is developing Nusantara, a new capital city on Borneo, rather than relying entirely on Jakarta in the future.
So yes, one of the world's biggest cities is literally sinking beneath its own weight and water use.