| | | Good morning! When politics is at play, a smart response often becomes essential. There was a threat of tariffs including a 25% tariff on chips, in addition to the 10% tariffs that the new US government has already enforced for imports from China. Apple has responded, a clear message from CEO Tim Cook that they don’t want further tariffs, with plans they intend to invest more than $500 million in the US over the next four years. That is, a new server factory in Houston, Texas which will focus on servers for Apple Intelligence, and plans to hire 20,000 more employees (a majority of these hirings likely to boost the local job market). Apple is also doubling its $5 billion US Advanced Manufacturing Fund, now to a tune of $10 billion — a collective push towards research and support for manufacturing. There will also be an academy in Michigan to train the next generation of U.S. manufacturers. | | PAY You’ll notice this soon enough, if you haven’t already. Google Pay is beginning to levy up to 1% of transactional value as a ‘convenience fee’ for bill payments on the platform using any credit cards or debit cards (for now, no change on UPI transactions). I’d talked about the odds being stacked against credit cards earlier (some ill-thought implementations; you’d recollect principal of causation). I find the terminology ‘convenience fee’ particularly irritating — you save on many a cost by pushing digital transactions, and behave as if doing the customer a favour? Anyway, this is simply going to have a cascading effect. Google Pay is doing it, only a matter of time before Amazon Pay (they anyway love layered charges; Prime Video rentals are a particular annoyance), PhonePe etc. will soon join the league. Whichever way you look, more money is leaving your pocket. Never a good sign. | | CHRONOLOGY Grok 3 is suddenly free for everyone. That is not the transition anyone expected, did they? We had a conversation about it last week, and at the time, xAI was insisting X users must subscribe to the Premium+ plan which saw a significant subscription price hike (that’s now Rs 34,380 per year after a price increase) in an hope to cash in. Two birds with one stone — X generates some revenue, and Grok gets new users — but that’s not how it is panning out. Here is a chronology. xAI released Grok 3 as competition to everything AI OpenAI, Google, Microsoft and others have available → made it exclusive to the highest tier X subscription → made it available to middle tier Premium subscribers a day later → and that wasn’t it, now free X users also have access to Grok 3. Suddenly, a DeepSearch and Think tool, free of cost. But why? There are reasons why this may have happened. And I’m purely analysing here, based on the sequence of events (xAI will never really admit something went wrong with the plan, if at all it did). | | The quick change in tack is likely a result of noted sluggish initial uptake and I’m sure sign-up data trajectory would’ve indicated that within the first day. Smart tactics by xAI, to turn a potential disadvantage into one that’s completely the opposite. | | | | | Secondly, and related to that is the matter of cost. In contrast, Perplexity released its Deep Research tool a few days ago, and is offering it free for use. The annual subscriptions for Gemini and Copilot cost lesser too than Premium+ and that’s something Grok’s owners corrected quickly. | | | | | Then there is the often-underrated question of habit. Grok may find it difficult to break user habits and recently found loyalties, at least more resistance than xAI may have imagined. OpenAI’s ChatGPT, Google’s Gemini and Microsoft’s Copilot not only have an active user base, they’ve weaved in their AI chatbots into experiences that are largely seamless. Copilot in millions of Windows PC, Gemini in millions of Android devices. GPT models on iPhones with Apple Intelligence. Even Meta AI within WhatsApp, that is extremely popular in many countries including India. To have to go to X or the standalone Grok app, seems like a bit of an inconvenience, in the grand scheme of things. | | | | | Finally, it seems unfair to those few who paid for the price corrected Premium+ subscription, for Grok 3’s exclusivity. Whether they get a refund, remains to be seen (at least I’ve not heard of any such communication; do share if you have). | Grok 3 being free led me to an experimentation expedition. A DeepSearch question “who are the top 5 users or accounts spreading disinformation on X” generated this response after 50 seconds - “Elon Musk, Joseph Mercola, Robert F. Kennedy Jr., Ty and Charlene Bollinger, and Sherri Tenpenny are the top five users spreading disinformation on X.” I happened to be chatting with a friend at the time, and we almost immediately said the same thing “Grok AI will be fixed very soon”. By evening, the “fix” was visible. Grok 3 began listing Alex Jones, Jackson Hinkle and Andrew Tate as the common offenders. A name was conspicuously missing. Would you really want to pay big bucks for an AI tool that so blatantly is made to bend to any whims of the “first friend”? A unanimous answer, isn’t what xAI was expecting. But they may still be able to turn a disadvantage into an advantage. In the AI games, that’s a solid footing to have. | | REFINEMENT In case you’re using Microsoft Edge as your main web browser (there’s little reason to, nevertheless we persist), there’s a significant update that’s now available. The v132 update, the company says, “no less than 14 different areas of the user interface throughout the browser are now 40% faster than before, on average.” The primary reason for this performance boost are architectural improvements brought along by WebUI 2.0 migration. A must update, in my book. | | QUANTUM A leap forward. Just weeks on from Google Willow defining a significant leap forward for an envisioned future of quantum computing, Microsoft Majorana 1 has written a completely different approach. In fact, astonishing for a tech company that currently gives us two versions of Teams and two versions of Outlook installed in Windows 11, they’ve gone ahead and created a new state of matter! It is made from indium arsenide and aluminium (the first is a semiconductor and the latter is a superconductor when cooled).  (Premium): Will Microsoft’s new AI chip change the quantum computing equation? At the heart of Microsoft’s creation is a concept called a topological qubit. That warrants some explanation. Traditional computing stores information in bits, encoded as either a one or a zero. Quantum computing is built around qubits that does both at the same time (a bit like Schrodinger’s cat). The only problem is that the minute someone tries to read a qubit, it collapses into either a 1 or a 0 (this is called decoherence). A topological qubit prevents that from happening. The term is derived from the mathematical concept of topology, which deals with the properties of an object (or material) that stays constant under stress. For other companies working on quantum computing, there are many approaches to work with. For instance, Google uses a technique called quantum error correction. You’d love to read the piece, link above! | | BULLION | | Mistral AI’s approach to localised relevance is something more AI companies will look closely at in the months ahead. Their new Saba model, which is a 24-billion parameter model, is trained on meticulously curated datasets from across the Middle East and South Asia. The model’s support for Arabic and multiple Indian languages serves as the foundation for its proficiency in Tamil and Malayalam. | | | | | Voice interactions with AI are becoming increasingly sophisticated. Google Gemini will now be able to summarise a voice conversation between you and the AI, before you close the chat. This follows two recent additions in the broader scheme — memory retention and 14 new languages for Gemini Live conversations. Microsoft Copilot Voice has also expanded its language support to include 40 more languages beyond English, Hindi, and Bengali; Copilot has also added a ‘Think Deeper’ feature based on OpenAI’s GPT-4o model. | | | | | | | Were you forwarded this email? Did you stumble upon it online? Sign up here. | | Written and edited by Vishal Shanker Mathur. Produced by Md Shad Hasnain. | | | | | | | Get the Hindustan Times app and read premium stories | | | | | | View in Browser | Privacy Policy | Contact us You received this email because you signed up for HT Newsletters or because it is included in your subscription. Copyright © HT Digital Streams. All Rights Reserved | | | | |
