Arnold Kling: “The key characteristic of the new AI is not encyclopedic knowledge. It is its personality. For the first time in the history of computing, we can relate to a computer using human language. We don’t have to learn how to write code or navigate a menu. I think that the economic effect of AI will depend on how far this can be taken.”
Ruchir Sharma: “With signs of a slowdown starting to surface in the US, India and other recent economic stars, many people are looking for the next promising growth stories. The “circle of life” is one guide: it is the cycle of politics in which crisis forces reform, which generates recovery but then breeds complacency and decay. The cycle then starts again, crisis to crisis, ashes to ashes. At any point in time, a nation’s prospects can depend on where it stands in this circle…Let’s not forget that when a nation has its back to the wall, it is often forced to carry out serious reform. And recovery follows. The cycle turns constantly, and the circle of life is always creating new stars somewhere.”
WSJ: “The builders of an app called World—including Chief Executive Alex Blania and his co-founder Sam Altman of OpenAI fame—envision a time in the not-too-distant future when you can’t do much without an ocular check-in. AI agents will be so prevalent, and so humanlike, that we’ll need to repeatedly prove we’re real to prevent those AIs from masquerading as humans on everything from payment platforms to social networks…World says its network is anonymous and secure, and only proves you’re a person—but it could be made part of a broader identity system. One of World’s distinguishing features is that it is an “anonymous proof of human” system. That is, its eyeball scan can verify that you’re a human being and not an AI, but without additional software and systems, it can’t identify who you are. What could drive people to adopt a system like World’s will be the rise of ever-more-sophisticated AIs that will make doing business on the internet almost impossible without something like it, says Tiago Sada, chief product officer at the company.”
Forbes: “To understand Manus, imagine an invisible assistant who can use a computer just like you do—opening browser tabs, filling out forms, writing emails, coding software, and making real-time decisions. Except unlike you, it never gets tired. The key to its power lies in its multi-agent architecture. Rather than relying on a single neural network, Manus operates like an executive overseeing a team of specialized sub-agents. When assigned a complex task, it divides the problem into manageable components, assigns them to the appropriate agents, and monitors their progress. This structure enables it to tackle multi-step workflows that previously required multiple AI tools stitched together manually. Its cloud-based asynchronous operation is another game-changer. Traditional AI assistants need a user’s active engagement—Manus does not. It runs its tasks in the background, pinging users only when results are ready, much like a hyper-efficient employee who never requires micromanagement.”
FT: “DeepSeek’s R1 and other so-called “reasoning” models, such as OpenAI’s o3 and Anthropic’s Claude 3.7, consume more computing resources than previous AI systems at the point when a user makes their request, a process called “inference”. That has flipped the focus of demand for AI computing, which until recently was centred on training or creating a model. Inference is expected to become a greater portion of the technology’s needs as demand grows among individuals and businesses for applications that go beyond today’s popular chatbots, such as ChatGPT or xAI’s Grok. It is here that Nvidia’s competitors — which range from AI chipmaker start-ups such as Cerebras and Groq to custom accelerator processors from Big Tech companies including Google, Amazon, Microsoft and Meta — are focusing their efforts to disrupt the world’s most valuable semiconductor company. “Training makes AI and inference uses AI,” said Andrew Feldman, chief executive of Cerebras. “And the usage of AI has gone through the roof . . . The opportunity right now to make a chip that is vastly better for inference than for training is larger than it has been previously.””
