Six Minutes of Silence

This week, AI wasn't about bigger models or smarter chatbots — it was about trust. Governments opened a new chapter of global AI governance in Geneva. NATO treated AI as a strategic asset while Europe pushed for technological independence. The UN's AI for Good Commission launched, India showcased AI innovations from flood prediction to logistics, and Indian businesses declared trustworthy...

The AI conversation is no longer about what AI can do — it'sabout whether it can be trusted, who is accountable when it decides, and who controls the most powerful systems. Indian enterprises are turning trustworthy AI into a competitive advantage, banks are being urged to bring AI into core strategy, and the world's major AI powers are tightening control over frontier models. Plus — a bra...

AI has entered its institutional era. Today the AI for GoodGlobal Commission holds its inaugural meeting in Geneva — world leaders and AI company CEOs at one table, deciding how AI gets governed and who it answers to. At NATO, European allies accelerate investment in home-grown AI. And in India, IIT Bombay unveils an AI model that predicts flood depths before the water arrives, while IIM Bangalore...

The biggest AI story this week isn't a new model or benchmark — it's that AI has become a matter of national security, diplomacy, and economic sovereignty. AI takes centre stage at the NATO Summit in Ankara, the UN'sGlobal Dialogue on AI Governance concludes in Geneva, and India's national agenda shifts from experimentation to large-scale sovereign deployment. Countries are now tre...

Today the UN opens its Global Dialogue on AI Governance inGeneva — governments, researchers, and industry leaders asking who is accountable when AI makes a consequential decision. In India, a deepening AI partnership with Japan and a growing push for sovereign infrastructure reflect the same question asked differently: not just can AI do it, but who controls it, and on whose terms. Plus — a quick,...

This week on Six Minutes of Silence — four stories, one question underneath all of them. Colorado's AI Act took effect Monday while Washington watched. India's central bank mandated AI kill switches for every bank in the country. AI arrived in India's criminal justice system, and factory workersrecorded their own hands to train the robots that may replace them. Then the United Nations...

A landmark UN report this week warned that AI risksconcentrating all the wealth and power of the next century in a handful of countries and companies. Not the robots-will-take-over risk. The simpler, older risk -- that powerful technologies follow the same pattern they always have,and the people already ahead stay ahead. Meanwhile, India launched two new sovereign AI infrastructure initiatives thi...

India is preparing to bring AI into its criminal justice systemby the end of this year -- faster investigations, better case management, more transparency. But faster isn't the same as fairer. Meanwhile, Indian factory workers are being asked to record their own work to train AI-powered robots -- teaching the machine to eventually do the job, using footage of themselves doing it. When AI learn...

Two weeks ago, this show covered the night the US government forced an AI company offline for the first time -- a surprise, no warning. This week, India's central bank did something related, the opposite way round: a draft framework requiring every bank using AI to build in an immediate kill switch, board-level accountability, and constant human oversight -- not a surprise, a rule built in fro...

Three weeks ago, this show covered the Great American AI Act -- a 269-page federal bill that would freeze every state AI law for three years. That bill still hasn't moved out of committee. Which means on June 30, in two days, Colorado's AI Act takes effect for real -- no federal override coming tosave the timeline. Companies have spent months preparing for federal preemption instead of pre...

A Saturday story. Not an AI story, not a cooking story -- justLinu's story. This one runs a little longer than usual. A microphone that sat untouched for almost a year. A niece's documentary that quietly changed everything. And the realisation that technology, cooking, and poetry weren't three separate interests -- they were three chapters of the same person. Closing with the poem &quo...

Meta is building an app called Arena for AI-powered prediction markets -- and its AI won't just generate the questions people bet on, it will decide who won, in near real-time. Up to now, AI has mostly assisted judgment. This is AI being handed the judgment itself. Meanwhile, a separate report this week found most Americans still don't use AI for daily news -- adoption is real, but still a...

For the first time since it launched, ChatGPT's share of the AIassistant market has dropped below half. It's still the leader -- but Claude, Gemini, and a growing list of specialised tools have, together, caught up. A monopoly becoming a marketplace. Three weeks ago, this show covered Appleletting iPhone users choose their AI assistant for the first time. Choice, once available, gets used....

Google just lost the man who helped invent the technology behind almost every AI you use. Noam Shazeer, co-creator of Gemini and a Transformer architecture pioneer, left Google for OpenAI this week. Nobel laureate John Jumper, the scientist behind AlphaFold, joined a rival too. Alphabet's stockdropped enough to wipe out $225 billion in market value in days. For a hundred years, companies were...

The US government just admitted the power grid can't keep up with AI. The Federal Energy Regulatory Commission issued emergency orders to six regional grid operators this week, bypassing the years-long normal regulatory process. Microsoft alone added 4 gigawatts of new data centrecapacity in the past 18 months — roughly the electricity use of a small country. "The grid cannot handle curre...

Episode 20. Four weeks of Six Minutes of Silence. This week gave us three threads worth keeping: a kill switchthat was used for the first time (the US government took two AI models offline — that power is now real). A country finding its place across three days — India losing IT jobs, building its own AI, and showing up at global tables to negotiate the rules. And a rocket company that bet its fut...

This week, India showed up on three different world stages for AI — in three different countries. In Slovakia, PM Modi announced an "India Chair on Artificial Intelligence" at a local university. At the G7 Tech Security Summit, India sat at the table discussing global AI and social media risks. And at VivaTech in Paris, India and France launched a Year of Innovation to bridge startups ac...

India minted a new AI unicorn yesterday. Sarvam — aBengaluru-based startup building voice agents in Indian languages — raised $234 million (about ₹1,950 crore) led by HCLTech, valuing the company at $1.5 billion. Their platform handles more than 20 million daily interactions, including Government of India farmer outreach programs and insurance campaigns. They built an AI that speaks Indian, to Ind...

Two weeks ago, the IMF warned that AI would eat the bottom rungs of the career ladder. This week, that warning landed in India. TCS announced 12,261 layoffs. Cognizant is cutting up to 15,000 under Project Leap, with India expected to absorb the largest share. And last week, TCS chairman N. Chandrasekaran said the day is not far when TCS will have half a million AI agents in its workforce. For thi...

Friday evening, at 5:21 PM Eastern Time, the US CommerceDepartment ordered Anthropic — the company behind Claude — to immediately disable its two most powerful AI models for any foreign national, anywhere on earth. The practical effect was simple: they had to disable Fable 5 and Mythos 5 for every customer, every country, within hours. Both models had launched just three days earlier. Whether the...

End of week three. Three threads to look back on. The Freeze:Congress unveiled the Great American AI Act with a three-year pause on every state-level AI law. The Eating: the IMF warned that entry-level hiring is down 15-25% as AI eats the bottom rungs of the career ladder. The Healing: researchers at the University of Michigan built an AI that reads brain MRIscans in seconds — flagging strokes and...

After a heavy week of news, a story that goes the other way. Researchers at the University of Michigan built an AI called Prima that reads brain MRI scans and produces a diagnosis in seconds — across 50+ neurological conditions, with 97.5% accuracy. When you walk into a hospital with a suspectedstroke, every minute matters. Prima can flag a stroke or a brain bleed in the time it takes to slide off...

The head of the IMF gave a warning yesterday that didn't make the front pages — but should have. AI is creating a shock to entry-level jobs. The first rungs of the career ladder are being eaten from the bottom up: drafting legal documents, writing first-draft copy, sorting tickets, assembling research. Entry-level hiring at large US firms is already down 15-25%. India is showing the same patte...

Yesterday at WWDC, Tim Cook walked on stage for the last time as Apple's CEO — and announced three things that quietly change how your phone works. Siri is being rebuilt by Google. You get to choose your AI assistant — Gemini, ChatGPT, or Claude. And Cook said goodbye. For thirty years your phone has been made by one company that picked everything for you. Starting this fall, the assistant bec...

Three weeks ago, California started writing the world's AIrules. This week, Congress is trying to stop them. A bipartisan 269-page draft — the Great American AI Act — includes a three-year freeze on every state-level AI law in America. And meanwhile, something else is happening that almost no one is talking about: President Trump, Senator Bernie Sanders, and OpenAI's Sam Altman are converg...