Flipper Device's new Busy Bar will retail for $249.

Omen AI raised a $31 million Series A to monitor chip coolant and stop bacterial outbreaks in data centers.

Table of Contents Building an Intelligent Chatbot with Qwen3 Instruct and Thinking Models Qwen3 Model Details Qwen3 Model Architecture: Dense Models, MoE Variants, and Dual-Mode Reasoning Qwen3 Training Data and Post-Training Pipeline for Reasoning and Instruction Following Qwen3 Benchmark Performance:… The post Building an Intelligent Chatbot with Qwen3 Instruct and Thinking Models appeared first on PyImageSearch.

AI is a factor in human well-being, whether harmful or helpful. EvgeniyShkolenko via Getty Images Artificial intelligence is reshaping relationships by providing conversation and companionship, and reshaping how people work. For children, it is making toys interactive and data-driven, and it is mechanizing and perhaps dehumanizing healthcare. The speed and magnitude of these transformations is breathtaking. Effective leadership of the development, deployment and monitoring of AI requires addressing risks to people and the environment. It also calls for ensuring equity – fair access to AI’s benefits and fair mitigation of potential harm. Pope Leo XIV set this tone in his recent encyclical, Magnifica Humanitas, about the human promises and perils of AI. We are scholars and leaders in human ecology – the interdisciplinary study of relationships among people, families, communities, society and the natural and built environment. We ask a central question: How can AI be truly human-centered, so that communities – and all forms of life – can thrive? We think the key is having all stakeholders – families, educators, engineers, designers, policymakers and citizens – involved in designing, coding, testing and monitoring AI. Innovation can come from not just programmers but also ethicists, policymakers and AI users. It can also come from experts who study implications for human behaviors and systems, with inclusion of historically marginalized people’s perspective, ideas and talents. Co-creation also makes it easy to “crash-test” innovations for harm before wide product release and for a balance between productivity and safety. Pharmaceutical development offers one lesson. The development of lifesaving medications can be thwarted by too much caution in not tolerating some potential side effects. But lack of guardrails can result in drugs that have deadly side effects or simply don’t work. In many industrialized nations, evidence-based regulation has led to a balance between making drugs available and keeping people safe. Although AI is not currently meaningfully regulated, similar kinds of guidelines could optimize benefits and prevent harm. Principles for humanistic AI Practical ethics and policy guidance are key to ensuring that the coming AI world is safe, equitable and environmentally responsible. Human ecological principles suggest that the best way to proceed is to analyze every AI proposal, product or service in terms of whether it supports human well-being and the global ecosystem. That includes supporting connectedness among individuals, in households, workplaces, communities and society. Concepts like the triple bottom line for businesses can help broaden perspectives on developing and using AI. To those ends, we offer eight principles as the beginning of a conversation among all people who may be involved in AI’s development and use. Begin any consideration of AI development or application with a human ecology frame, by asking, “How can humans use AI to promote thriving for all?” This perspective is broader than the primarily technological question of “What can AI do?” Consider ethical frameworks and guardrails to prevent and mitigate harm. If governments, universities and the private and nonprofit sectors include accountability in the design, planning and implementation of AI, then people are more likely to adopt the new technology than to mistrust it. Furthermore, if people consider the impacts on relationships among humans, institutions and the Earth, then the public can properly weigh the benefits against the downsides. An important way of preventing harm is by crash-testing best ideas about AI applications and openly imagining the worst things that can happen. Study where AI use may be appropriate and not. If AI co-creators consider scientific evidence about human development, from fetus to elderhood, and recognize the inherently different needs at different phases of life, then AI stands a better chance of strengthening, not substituting for, human care. Developers can pursue AI designs that respect and support the social habitats that sustain human thriving. For example, recognizing that a well-designed AI model can briefly engage a young child while a parent may be preparing dinner is far more realistic, and in our view appropriate, than claims of AI replacing interaction with human caregivers. Well-crafted AI wouldn’t supplant the caregiver-child connection but would help preserve the health of relationships, reduce strain on parents and allow for better connection. Educators can develop strategies at all levels, from K-12 through college, in which teachers and students use AI to help accomplish learning objectives and develop mentoring and nurturing relationships. Ecosystems for learning thrive on relationships among teachers, peers and families – not algorithms. These relationships foster critical thinking, perspective, creativity and problem solving. One way to improve human well-being is to avoid the harms of digital dependence, including physical inactivity and social isolation. Rather than hew to the well-known mantra of “move fast and break things,” AI developers can implemented innovations at a pace and scale that respect the risks of moving too quickly. Chaos is more likely with sudden disruptions to social equillibrium. Policymakers and planners can foster a civic AI culture that serves the common good. Equity is a critical measure of healthy social systems and so, we believe, should be a primary goal of AI development. One way to accomplish that is developing, deploying and monitoring AI in accordance with concepts such as the triple bottom line, which recognize not only profit but people and the planet. Policymakers can also develop metrics to assess the impact AI has on many aspects of human and environmental well-being. Every digital action has an ecological shadow. An accurate accounting of AI’s costs and benefits can help establish just and sustainable deployment. Guiding AI development Like any ecosystem, the digital dimension of our lives is evolving, but people can take an active role in guiding that evolution. AI will be what we make it. A holistic approach can ensure that AI strengthens the webs of life, learning and care that make us human. The authors do not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.

HP has scaled its OpenAI Frontier integration across global operations to optimise enterprise workflows and accelerate output. The hardware manufacturer initiated testing of the platform in February 2026. Early pilot programs yielded verified operational gains in software engineering and cybersecurity remediation. Expanding these initial trials into an enterprise-wide operating model requires connecting access protocols, contextual […] The post HP accelerates enterprise workflows with OpenAI Frontier appeared first on AI News.

Bloomberg: Japanese prediction market startups Miraima and Poyp are utilizing a “point-to-voucher” system to bypass strict anti-gambling laws, in a bid to rival Polymarket  —  Homegrown platforms, banned from real-money gambling, are tapping the country's loyalty-point culture to build an alternative to Polymarket.

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. The inevitable weakness of metrics There are plenty of useful things a metric can reveal. There are even more that it can obscure or corrupt. Like a lot of people bitten…

MCP provides a standard way for AI applications and external systems to communicate.

The rise of small language models The post How to Choose Between Small and Frontier Models appeared first on Towards Data Science.

Adversaries could plant a malicious repository that can execute arbitrary code and steal cloud credentials by exploiting the vulnerability, which showcases growing MCP risk.

Sergiu Gatlan / BleepingComputer: The US DOJ seizes nearly 400 domains for illegally streaming 2026 FIFA World Cup matches; last week, ACE, UEFA, and more shut down 44 domains linked to PirloTV  —  The U.S. Justice Department's Criminal Division has seized nearly 400 web domains used for illegally streaming matches at the FIFA World Cup.

Bloomberg: Chinese robotics startup AI² Robotics raised ~$736M at a ~$2.9B valuation; Alibaba-backed X Square Robot raised an undisclosed amount at a ~$2.9B valuation  —  Two Chinese robotics startups have been valued at more than $2.9 billion in recent funding rounds, demonstrating steady interest …

Startup exits valued at $1 billion or more are now more numerous than at any point since the 2021 market peak, Crunchbase data shows. The trend we’re seeing for the second quarter of 2026 includes both the largest venture-backed exit of all time and a bevy of other comparatively tinier but still sizable startup exits through acquisition or IPO.

Patrice Becker (left) and Leonhard Soenke (right) founded a creator economy startup in 2021 — and recently pivoted to energy and data centers. Courtesy of TAR Leonhard Soenke founded a creator economy startup called Throne. Now, he's working on data centers. After leaving New York for San Francisco, he and his cofounder wanted to tackle a new problem. The new company, TAR, is focused on energy efficiency in data centers. Leonhard Soenke cofounded Throne, a creator economy startup, in 2021 to help influencers create wishlists where fans could buy and send them gifts. Five years later, Soenke and his cofounder Patrice Becker are setting out on a new venture. It's not in the creator economy. Instead, it's in the data center business. TAR, which stands for Transformative American Resources, is a new energy startup aiming to make data center power more efficient. Soenke told Business Insider that TAR recently raised a $27 million seed round at a $500 million valuation from an undisclosed strategic investor. From the outside, influencers and AI data centers seem like two different worlds. However, Soenke says there are transferable skills between the two. While I caught up with Soenke about his new endeavor, I noticed he was wearing a nearly identical gray Patagonia quarter-zip to the one he had while at Throne, now with the TAR logo embroidered on the chest. "Our attitude is if ain't broke, don't fix it," he told me. This as-told-to essay is based on a conversation with Soenke. It's been edited for length and clarity. Patrice, my cofounder, and I met in university. We founded our first company, sold that, and then we founded Throne. Sometime through the summer of last year, we were speaking about the need to get away from running something that's very stable. We're very, I would say, spiky personalities. We spent a while in the creator economy. We got partnerships with major vendors, work closely with Amazon — it had really grown to a large org. The creator economy has matured a lot. It's a pretty crowded space. Why leave and start something new in a different space? Two things. One was the want to start something anew. Then you need to reevaluate, "Hey, are you going to stay in the creative economy? Are you going to go somewhere else?" The second question was, "Where is a space where we can really make a large impact?" The movement we see in a lot of the [social] platforms is not as much as we'd see in something like the AI infrastructure stack. We left New York, moved to San Francisco, and spent a lot of time with people at the labs, the major compute providers, discovering where problems lie. That's how we got to the whole idea of TAR: building the infrastructure to essentially power AI. We still maintain a San Francisco office for a lot of the engineering stuff, and we recently moved to Austin. If you're building energy for data centers, it's good to be in Texas. The hard decision to move on It is giving away your baby a little bit. The whole feeling of letting go, especially being a control freak myself, was the hardest part. One thing that made it a lot easier is we handed Throne over to trusted team members who've been with the company for a really long time. Patrice and I are happy to chime in on something. I'm still in Slack. When a bug comes up, I actually just hop in. The process was actually a pretty long one of handing it over, probably six months. You ease into it. Throne lets content creators on platforms like YouTube, TikTok, and Instagram create wishlists. Screenshot/Throne Pivoting from the creator economy to data centers In both cases, you solve very difficult problems. A lot of this is transferable. Patrice and I love working on very hard problems. When we started Throne, I remember we were in an apartment together, basically doing all the customer support and stuff alone at the beginning. Throne took off really quickly because there was such a demand for it. We see very similar things for energy, where it's such a strong pull from the market. Chips and power will become more acute problems as models become more useful and permeate society. We reduce time-to-token by building modular, scalable, behind-the-meter energy systems with a mix of generation sources. Most of the systems will consist of solar, batteries, wind energy, and natural gas. We're not inventing a new way to produce energy. We are focused on innovating the way we deploy these existing technologies faster. It's a lot more capex-intensive compared to building a software platform — procuring heavy equipment, building a warehouse, and developing land. It's a different beast. Texas has become a hub for data centers. Tom Fox/The Dallas Morning News via Getty Images The biggest difference is working with more legacy players in the energy space. Obviously, you come from a slightly different world if you're in software and you're just sitting in your New York or San Francisco office. We make an extra effort to get on-site, talking to contractors and blue-collar crews. That's a big shift. As a founder, you need to learn how to empathize with not just users, but customers and partners, and put yourself in their shoes. Read the original article on Business Insider

Matt Burgess / Wired: Google VP of Security Engineering Heather Adkins warns the EU's DMA proposals to open Android and Search could lead to a significant rise in fraud within weeks  —  Europe's pro-competition proposals could see Google Search and Android systems opened up.  The company claims there are serious privacy flaws.

Elon Musk said SpaceX would release new AI models "trained from scratch" every month. WEF/Getty images SpaceX is shifting "a few dozen" top Starship and Starlink engineers to work on AI. CEO Elon Musk said that the move was helping the company's Grok model catch up with its rivals. Musk also said Cursor staff were pitching in, after SpaceX agreed to buy the startup for $60 billion. It's all hands on deck at SpaceX as the company plays catch-up in the AI race. Elon Musk said on Sunday that SpaceX had deployed "a few dozen" top Starlink and Starship engineers to help overhaul its Grok model. "The SpaceXAI cadence of model and harness improvement is speeding up tremendously, particularly due to a few dozen of the top Starlink/Starship engineers shifting much of their time to AI," wrote Musk in a post on X. The billionaire added that engineers from Cursor, the AI coding startup that SpaceX agreed this month to buy for $60 billion, were also working on the new foundation model, which was partly trained on Cursor training data. Musk said that Grok 4.5, the latest version of the chatbot, was now in private beta at Tesla and SpaceX, and posted that SpaceX would release new models "trained from scratch" every month this year. Musk's efforts to take the lead in the AI race have faced roadblocks in the past few months. The Tesla CEO overhauled xAI, the AI startup he founded in 2023 to take on OpenAI and Google, earlier this year in a sweeping reorganization that saw the last of the company's 11 cofounders depart. XAI's Grok has lagged rival AI models from OpenAI and Anthropic, especially on coding, and Musk wrote in March that xAI was "being rebuilt from the foundations up." In February, Musk merged xAI with SpaceX, and the deal to acquire Cursor was confirmed shortly after the rocket company's record-breaking $85 billion IPO earlier this month. The deal cemented the meteoric rise of Cursor, an AI coding startup led by 25-year-old Michael Truell, and saw SpaceX grant Cursor access to the company's supercomputers in return for help training Grok. Deploying top Starlink and Starship engineers to overhaul Grok is the latest sign that SpaceX is going all in on AI. Musk has said the company will use the windfall from its mega-IPO to build a network of up to a million orbital data centers, built on Starlink technology and carried into space by Starship, that will train and run increasingly advanced AI models. In its investor materials presented before the IPO, SpaceX estimated its total addressable market was worth $28.5 trillion, of which AI accounts for $26.5 trillion. Read the original article on Business Insider

Spencer Soper / Bloomberg: Adobe: online spending across all retailers in the US hit $26.4B during Amazon's four-day Prime Day event, up 9.3% YoY; Walmart and Target also hosted sales  —  Online spending across all retailers in the US hit $26.4 billion during Amazon.com Inc.'s annual Prime Day sale, according to Adobe Inc. …

NASA’s quiet supersonic flight tests could eventually go on a national tour.

Gary Hershorn/Getty Images Comcast announced plans to split its media and tech businesses into two companies. The move is intended to give each company greater strategic focus. Its stock price is up more than 25% in premarket trading. Comcast's stock is surging after it said it's spinning off its media businesses. Shares jumped more than 25% in premarket trading on Monday after the company said it plans to separate its media and tech businesses into two publicly traded companies. This will be done through a tax-free spin-off of NBCUniversal and Sky, it added, which it expects to be completed within a year. The move is intended to give each company greater strategic focus. The company said the separation would allow both businesses to invest more effectively and pursue their own growth opportunities. "This is a very exciting day for our company. The transaction we are announcing will unlock a more entrepreneurial management approach and open up a multitude of new opportunities for each business," Comcast co-CEO Brian Roberts said. Earlier this year, Comcast also officially completed the spin-off of the majority of its cable networks, including CNBC and MSNBC, into a separate company called Versant Media. Read the original article on Business Insider

A fake tombstone announcing the "death" of the Alta Vista search engine sits on a desk in the Google Mt. View office on Thursday March 13, 2003 MediaNews Group/The Mercury News via Getty Images A version of this story originally appeared in the BI Tech Memo newsletter. Sign up for the weekly BI Tech Memo newsletter here. Former Google engineering leader Hugh Williams is back with another fascinating Claude Code experiment. After previously showing how Anthropic's coding assistant helped him build an AWS-based system in 48 hours, Williams has been using Claude Code recently to create a working search engine called Zettair that indexes 1.5 million Wikipedia articles. The project includes features you'd expect from a modern search engine: autosuggest, query-biased snippets, related searches, trending topics, and AI-generated summaries. Williams says he wrote zero lines of code himself, relying entirely on Claude Code. That's technically true, with an important caveat: the underlying search engine is based on an information-retrieval system he helped build in the early 2000s. His biggest takeaway wasn't that AI can replace engineers. Instead, Williams argues Claude Code works best when paired with deep technical expertise. Building with AI, he says, feels less like programming and more like coaching — and experienced engineers still make the best coaches. Sign up for BI's Tech Memo newsletter here. Reach out to me via email at abarr@businessinsider.com. Read the original article on Business Insider