Bloomberg: “The reality is that Hangzhou, where DeepSeek is based, and other Chinese high-tech centres have been roaring with little AI dragons, as AI startups are often called. Sophisticated chatbots from homegrown startups such as MiniMax and Moonshot AI have rocketed in popularity, including in the US. Alibaba Group Holding Ltd’s Qwen family of large language models (LLMs) consistently ranks near the top of prominent leaderboards among LLMs from Google and Anthropic; Baidu Inc’s chief executive officer, Robin Li, boasted in April that the search giant could develop models that were as good as DeepSeek’s but even cheaper, thanks to its new supercomputer, assembled with in-house chips. Huawei Technologies Co is likewise winning plaudits for the products it’s designed to compete against equipment from Nvidia Corp, whose graphics processing units (GPUs) power the most advanced AI models in the US and Europe.” [via The Edge Malaysia]
FT: “OpenAI, Google, Meta and Microsoft have stepped up their focus on memory in recent months, releasing upgrades that enable their chatbots to store a larger amount of user information to personalise their responses. The move is viewed as an important step to help top AI groups attract customers in a competitive market for chatbots and agents, as well as a means to generate revenues from the cutting-edge technology…“If you have an agent that really knows you, because it has kept this memory of conversations, it makes the whole service more sticky, so that once you’ve signed on to using [one product] you will never go to another one,” said Pattie Maes, a professor at MIT’s media lab and specialist in human interaction with AI. AI chatbots such as Google’s Gemini and OpenAI’s ChatGPT have made huge strides. Improvements include expanding context windows, which determine how much of a conversation a chatbot can remember at once, and using techniques such as retrieval-augmented generation, which identifies relevant context from external data.”
Nikkei Asia: “Unlike e-commerce, ride-hailing or financial services, where revenues kick in from the beginning, startups focusing on deep tech take years to develop and commercialize their products. Venture capitalists, however, often want to see revenues — or at least a clear path to revenue — before they invest. They also like to see an exit path, and a number of consumer and software companies have gone public, with many others waiting in the wings. That has created a funding crunch in the deep tech realm. Indian deep tech startups raised $4.7 billion from 2014 through April 28, about 3.2% of the total funds funneled into Indian startups during that time, according to data company Venture Intelligence. Nearly 500 deep tech startups shut shop, estimates Tracxn, a company that tracks private investments.”
Mint: “Today, the business book has become something of a glorified business card. Many are padded essays stretched to 200 pages, bloated with anecdotes, loaded with buzzwords and dressed in covers screaming for attention. With an estimated 12,000 business books published every year, the genre is suffering a global glut. Few books offer a breakthrough idea. Many are simply checklists. And in a world of digital content overload, thought leadership is shifting to more dynamic platforms. Podcasts, newsletters, X threads, LinkedIn posts and short videos are shaping professional discourse faster than traditional publishing cycles can churn out books. A book can take two years to write and publish. A podcast can go live right away and viral soon after. Publishers are adapting: Simon & Schuster recently announced it will reduce reliance on expert reviews on back covers. This isn’t merely administrative; it’s an acknowledgment that curated praise doesn’t drive sales. Another brutal reality for authors is that a book that takes four years to write can vanish in four months, even before it gets a chance to appear in a cheaper paperback edition.”
