Latest AI News
Satya Nadella has issued a shocking warning to companies using AI
Of all the debates raging about the potential downsides of AI, there is one worry causing the most hand-wringing among AI enthusiasts in Silicon Valley. Their fear is that the giant AI labs that sell proprietary models are somehow acting like Trojan horses. The concern is that, as startups and enterprises use AI models from labs like OpenAI and Anthropic, the labs gain ever-increasing access to those companies’ most sensitive business information. The model makers can then use that knowledge for themselves, potentially becoming competitors to their own customers. Those issuing such warnings range fromVCs like Jason Calacanisto Palantir CEO Alex Karp. Now, in a surprisingblog postpublished on Sunday, Microsoft CEO Satya Nadella has joined this crowd. Nadella warns that AI users (the “buyers” as he calls them) are paying twice. They knowingly spend for AI token usage but they also, obliviously, hand over valuable data in the process. “You essentially pay for intelligence twice, once with money, and again with something even more valuable: the proprietary knowledge you must reveal to make that intelligence useful. The better you want the model to perform, the more of that knowledge you have to feed it!” he writes. Most dangerously, enterprises are literally teaching the models about the nuances of their businesses, he argues. “Models learn from ‘exhaust,’ the prompts people write, the tools agents use, and especially the corrections people make when the model is wrong. Every correction is distilled into institutional know-how,” he writes. This is “the kind of knowledge a competitor could never buy,” and yet enterprises are handing it over. Nadella argues that if AI companies get to freely scrape the internet to train their models, it’s only fair that enterprises get to study — or “distill” — those models in return. “Distillation” is the practice of using a model’s own outputs to learn how it works and to train a new, often cheaper, model based on those insights. In February, Anthropic accused Chinese open source models ofsending millions of prompts to Claudeas a way to improve their own models, and urged the U.S. government crack down on export controls. Nadella’s point is that model makers can’t have it both ways. It’s hypocritical for them to freely train on the world’s data while restricting others from doing the same to their models. “While the great innovation that comes from model providers having fair use rights to train models on public data is needed, I find it ironic that the status quo is to then turn around and impose restrictive terms on distillation,” Nadella writes. Nadella is particularly concerned when model makers “reserve the right to learn from customer usage and interaction data.” Nadella’s solution is the kind of thing the CEO of a giant cloud provider would suggest. He wants companies to “retain ownership” of their data, including prompts, feedback, etc. So he’s urging them to build their own “proprietary learning environments” on the cloud (where their data is likely already stored anyway and, conveniently, could mean Microsoft’s cloud, Azure). He also wants companies to build in what he calls “orchestration layers” — essentially, a way to easily switch between AI models from different providers rather than being locked into one. Tools like AI “gateways” that let companies do exactly this have become increasingly popular. While Nadella never uses the words “open source” as the method for retaining ownership, this is an obvious subtext. Yet, there’s another subtext. Large companies, many of which still have some of their own data centers in addition to using the cloud, are already moving to open source models installed on their own premises (“on-prem,” in industry jargon). Idit Levine, founder and CEO of Solo.io — which makes networking and security software that helps enterprises manage AI systems — says she’s seeing exactly this shift play out with her own customers. After experimenting with proprietary model makers, they start asking themselves: “Can I take an open source model and run it on-prem? It will do almost 90% of what the big one’s doing. It will cost way less,” she tells TechCrunch. “They understand that, and they can control it.” Solo.io’s technology was selected last year to be the tech powering theLinux Foundation’s Agentgateway project. Her company counts enterprises like T-Mobile, ADP, and SAP as customers. She sees companies increasingly installing on-premise open source models and sees it as the next big wave in enterprise AI use. She’s not alone. Vercel (best known as a platform for building and hosting websites, which has recently added AI model-switching tools) and OpenRouter (a company that helps developers route requests across different AI models) are both seeing a surge in traffic toopen source models. In fact, open models accounted for 29% of all traffic routedthrough Vercel’s gatewaylast month. With the CEO of Microsoft, a company that has invested in both OpenAI and Anthropic, now openly urging enterprises to be wary of using proprietary models, we’ll bet this trend continues to grow. “In consuming intelligence, you are creating intelligence. And what you create should belong to you,” Nadella writes.
View
Anthropic starts localizing Claude pricing for India, its biggest market after the US
Anthropic has started localizing Claude’s pricing in India, its biggest market outside the U.S., as global AI companies increasingly tailor their offerings to win users in the world’s most populous nation. Local pricing has begun to appear for some users in India on Claude’s website and mobile apps. However, Anthropic has yet to enable payments via the Unified Payments Interface (UPI), India’s widely used instant payments network. Users still need to pay by card or through Apple’s and Google’s app store billing systems. This is unlike OpenAI, whichrolled out Indian rupee pricing for ChatGPTin August with UPI support. Claude users in India havelong soughtrupee-denominated subscriptions, with dollar pricing and currency conversion adding friction to accessing the service. The move is particularly significant, as India accounts for5.8% of global Claude usage, making it the service’s second-largest market after the U.S., according to Anthropic. On Claude’s website in India, Anthropic is listing Claude Pro at ₹2,000 (about $21) a month when billed annually, compared with $17 a month in the U.S. Claude Max starts at ₹11,999 (around $125) a month in India, versus $100 in the U.S., while Team plans start at ₹2,399 (around $25) per seat a month, compared with $20 in the U.S. The India prices include local taxes. Moreover, prices on Claude’s mobile apps vary slightly from those listed on its website. The Indian rupee pricing comes amid Anthropic’s growing focus on India. The Claude maker opened an office in Bengaluru in February, afterannouncingthe move in October, and in Januaryappointed former Microsoft India managing directorIrina Ghose to lead its business in the country. Anthropic has also partnered with Indian IT services giantsInfosysandTata Consultancy Servicesin recent months as it looks to scale enterprise AI deployments. That expansion faced a setback in June when Anthropicabruptly suspended accessto its Fable 5 and Mythos 5 models for non-U.S. entities, prompting some Indian developers and startup founders to consider alternatives to American AI models. The restriction on Fable 5 hassince been lifted, though access to Mythos 5 remains limited. India has become an increasingly important market for AI companies, driven by its large base of developers and technology workers. However, converting widespread usage into paid subscriptions remains a challenge in the price-sensitive market. Anthropic did not respond to a request for comment on the Indian rupee pricing rollout.
View
Should AI help you get away with killing your spouse?
Quick question: Do you want AI to be so well trained, it could help husbands (or wives, for that matter) plan the perfect murder of their spouses? Just as a gut reaction, that feels like a no for me. I wouldn’t even think it was a particularly hard question. But America contains many diverse perspectives, andone such perspectivewas shared by Comma AI founder and longtime jailbreaker George Hotz over the weekend. The post comes in response to a bunch of big-picture AI alignment plans, most recently theAI 2040: Plan Apolicy paper from the AI Futures Project. That paper envisions a world in which the world’s researchers collectively choose to slow down AI development for 14 years for the good of humanity. But of course, not everyone who read the paper agrees with its premises or conclusion. In fact, Hotz disagrees with the whole premise that AI progress should be managed for the collective good. In his post, he argues that the fast-takeoff scenario — the hypothetical where AI rapidly obtains superhuman abilities — doesn’t make a lot of sense. (I agree with a lot of what he says here!) For Hotz, the best approach to AI alignment and safety is to focus on locally controlled AI models that are closely aligned with the interests of their users. That’s a cool idea, particularly since it reminds me how much of current AI is built around centrally managed services like Claude and ChatGPT. There are infrastructure-related reasons why AI services developed this way: It’s expensive to host these large state-of-the-art models and most people aren’t using them enough throughout the day to justify truly personal AI. But those factors become less important as the technology develops. Part ofwhat was so exciting about OpenClawwas this experimental, DIY approach, and it would be great to see more AI products try to recapture that. But Hotz is a provocateur by nature, so he doesn’t stop there. He compares user-aligned AI to a gun(!), which does not complain if you use it to kill your stepmom. (I feel like there are other rules against this?) A truly aligned AI would be able to order meth-lab equipment from Amazon Prime and show you how to use it if that’s what you wanted and asked for, he says. (Again, I don’t think AI would be the limiting factor here.) Hotz even says he would die to defend this principle, although it’s hard to imagine the series of events that would lead to that. “We either live in a world with freedom or we don’t,” Hotz writes. If those are the options, the freedom world does sound better! Still, I don’t know. It’s not all about freedom, right? Any structure involving a lot of people (societies, marketplaces, corporations, etc.) requires balancing equities, binding individual needs into a network of interdependent preferences and systems of accountability. And anyone deploying mass-market tech products should probably think about that network as a whole, which means taking seriously the interests of the as-yet-unmurdered spouses and stepparents of the world. The freedom Hotz experiences is really a space of potential futures made possible by collective enterprise; those futures would disappear overnight if we all started behaving like little AI-powered Napoleons. Like the meme says, we live in a society. Having a local AI willing to take on the corporate world for my benefit does sound cool though! I can’t wait for a review unit.
View
Sam Altman’s space data center trash talk is what most experts already believe
Sam Altman and Elon Musk traded barbed social media posts over the weekend, drawing new attention to the gap between vision and reality for the space-compute business. Responding to Muskaccusing himof being a scammer, Altmansaid, “homeboy you’re the one sellling [sic] public market investors on short-term space datacenters.” Setting aside “homeboy,” Altman is saying what a lot of experts have concluded but public market investors seem to be ignoring: Space data centers are not going to be a serious business anytime soon. SpaceX’s plans to launch a fleet of orbital data centers to perform AI inference tasks are themain driverbehind the company’s 2-trillion-dollar valuation. Bullish analysts say that the potential for that processing power to fuel SpaceXAI’s models or act as an orbital neocloud are unprecedented in the AI boom. But when you talk to subject-matter experts — whether it’s theentrepreneursbehindother space data center startups, theteam at Googledeveloping that company’s orbital compute project, or engineers who havedone the numbersfor fun — you find the same answer: This isn’t going to make a big dent until we have much cheaper rockets and the ability to produce high-powered satellites at low cost, en masse. Musk’s answer to this is easy to predict: Starship, SpaceX’s huge new rocket, is expected to make its 13th test flight as soon as July 16. If Musk’s team can get that vehicle to the point where it flies again and again, the data center business case could close. But even if the company successfully recovers both stages of the rocket on this test flight, operational reusable flight will still likely be years away, and space data center launches will likely take a back seat to SpaceX’s commitments to NASA and to building out its own Starlink network. SpaceX also conceded during its IPO road show that Starshipmay not be fully reusablein the near-term and will need to throw each of its second stages during each launch, which would put a kibosh on economical space data centers. That’s why Musk’srejoinder— “We start flying them next year” — falls a bit flat. There’s no doubt that SpaceX could launch a satellite equipped for high-speed data processing next year, but the big question is when it will be able to launch and manufacture them at scale. And that’s likely a question for the 2030s.
View
The wildest allegations in Apple’s trade secrets lawsuit against OpenAI
Apple’strade secret lawsuit against OpenAIis packed with a number of extraordinary allegations that paint a picture of a coordinated effort to extract confidential information from current and former Apple employees. But what’s perhaps most striking is how casually the alleged misconduct is described, including one message that reads, “LOL, I found out I can access the [network storage], so funny.” The 41-pagecomplaint, which was filed on Friday, is filled with unusually detailed allegations, like this and others. Here are some that stood out the most to us: So far, OpenAI has only commented publicly via astatement shared on Xon Friday, which reads: “We have no interest in other companies’ trade secrets. We remain focused on building innovative technology that empowers people everywhere.”
View
Waze adds new AI-powered features and customization updates
Waze, the Google-owned navigation app, is getting newAI-powered features and customization updates, including the ability to report road updates conversationally and get personalized navigation. Some of the new features are powered by Google’s Gemini AI assistant, which reflects the tech giant’s broader push to integrate Gemini across its products while also better positioning Waze to compete with rival services such as Apple Maps. The app now suggests routes based on both a user’s trip history and its understanding of a city’s traffic patterns. For example, if a user prefers driving on highways over local streets, they will see those suggested first. If users don’t want personalized routes, they can choose alternate routes or turn off personalization altogether in their settings. Personalized navigation is rolling out now globally on Android and iOS. Users can now also use Gemini to find their destination through a quick chat when they know what they need but aren’t sure exactly where to go. Users can tap the search voice icon to ask questions like “Find me a coffee shop that’s open right now,” “Find me parking close to Grand Mall” or “Find me a gas station nearby with the lowest prices.” Waze will then respond with a list of options. The ability to find your destination with Gemini capabilities is rolling out now to the Waze beta community globally on Android and iOS. Additionally, Waze is introducing a new Motorcycle mode that uses AI to account for two-wheeler-specific shortcuts and road restrictions to help riders find the best route and receive more accurate ETAs. It also shows hazards that could be tricky for riders, like potholes, speed bumps, raised crosswalks, shoulder endings, and narrow bridges. Motorcycle mode is rolling out now in Argentina, Brazil, Colombia, Malaysia, Mexico, Peru and the Philippines on Android and iOS, with more countries on the way, Waze says. While Waze already lets users report traffic incidents, such as slowdowns, using natural speech, it now also allows them to suggest map updates the same way, including reporting road closures or outdated addresses. For example, you could say “The road is closed here,” and Waze will send these details to local map editors. The option to report road updates conversationally is rolling out now globally on Android and iOS. There’s also a new “less chatty” mode that features fewer interruptions on your drive when you want to focus on your music or a podcast. When you toggle on the setting, Waze minimizes the number of voice prompts and keeps them short. You’ll still be alerted to hazards and turns, but you’ll get them less frequently. Less chatty mode is rolling out now globally on Android and iOS.
View
How NTT DATA Plans to Generate $2 Bn from Agentic AI by 2027
The IT services giant is embedding AI across applications, BPO and data services as enterprise customers move beyond pilot projects.
View
TVH Expands Pune GCC, Increases Capacity by 150 Seats
The expansion provides room for 400 employees as TVH scales its technology, engineering, and digital operations from India.
View
ITC Infotech Leads Race for Majority Stake in Happiest Minds: Report
The proposed deal, if concluded, would mark another major consolidation move in India's IT services industry as companies pursue acquisitions to build scale and strengthen AI-led capabilities.
View
HCLTech's AI Revenue Surges 62% to $171 Mn as Automation Reshapes IT Services
The IT company reported a 20% rise in June-quarter profit, with AI revenues climbing to an annualised run rate of $688 million.
View
HCLTech Enter AI Infrastructure Biz with ₹3,500 Crore Investment in Data Centres
HCLTech said it plans to build a full-stack AI platform spanning AI data centres, GPU infrastructure, foundation models, enterprise applications and managed services.
View
TCS Wins Multi-Million, Multi-Year ABB Deal to Run Global AI-Powered Network Operations
The expanded partnership will see TCS build an AI-driven, secure and standardised digital infrastructure to power ABB's global network operations.
View