Shekhar Gupta: “1985-1995 would have the most stories that dominate our democracy and debates today. At home and around, think terminal decline of the Congress after peaking, the first coalitions, the Bofors scandal, Mandal versus Mandir contestation, insurgencies in Punjab and Kashmir, two Indian military interventions overseas (Sri Lanka and the Maldives), two war-like situations with Pakistan (Brasstacks, 1987; Pakistan’s first nuclear blackmail, 1990), a fraught Sumdorong Chu standoff with China (1986-87) and then a thaw with Deng Xiaoping, assassinations of Zia-ul-Haq and Rajiv Gandhi, the globalisation of Islamist (as distinct from Islamic) jihad and its spread to Kashmir, and the freeing of India’s economy. Although it started with rock-like stability with the Congress at 414 in the Lok Sabha, the 10 years saw four Prime Ministers. Isn’t that enough for a mere decade?“
John Plender: “Today’s euphoria is chiefly around AI and crypto. The curious thing is why people feel a need for a debate on whether they constitute a bubble, given that they so manifestly meet all the usual bubble prerequisites…Yes, there could be a 1929-style crash. But once again, as after the dotcom burst and the 2007-09 financial crisis, the central bankers will be putting a safety net under markets. And the US will continue to act as the ever constant borrower of first and last resort in the global economy. So this rerun of the Roaring Twenties will not be followed by a Great Depression. The prognosis for 2026 is thus for more froth before the crash, more Fomo and growing financial stress. Nor should we forget geopolitical friction, which could well be instrumental in pricking the bubble. And when we reach the post-crash world, prepare to reconcile yourself to protracted stagflation and unremitting populist political pressure.”
WSJ: “Nvidia’s graphic processing units, or GPUs, dominated the initial push to train AI models. But companies like Furiosa are betting that for the next stage—referred to as “inference,” or using AI models after they’re trained—their specialty chips can be competitive. Furiosa makes chips called neural processing units, or NPUs, which are a rising class of chips designed specifically to handle the type of computing calculations underpinning AI and use less energy than GPUs.”
FT: “In his next chapter, [Yann] LeCun believes that setting up a “neolab”, meaning a start-up that does fundamental research, is the new, most fertile ground. He cites OpenAI former chief technology officer Mira Murati’s Thinking Machines (“I hope the investors know what they do”) and OpenAI’s co-founder and chief scientist Ilya Sutskever’s Safe Superintelligence (“There I know the investors have no idea what they do”) as good examples. His new architecture uses videos to give AI models an understanding of the physics of our world, which will allow them to make better predictions of what will happen next. The model also relies on “emotions”, meaning past experiences and evaluations, to guide its predictions…LeCun says we will see “baby” versions of this within 12 months, and on a larger scale within a few years. It’s not quite yet superintelligence, but a path towards it.”
