M.G. Siegler / Spyglass: Like the Claude app it imitates, the new ChatGPT “Super App”, which merges Codex with ChatGPT, is a tangle of toggles and strange UI decisions  —  Even if it's ultimately the right call, the new Mac app is a mess...  I am so confused right now.  I mean that literally.

Organizations, and possibly ATMs, are at risk of compromise, thanks to holes in a Microsoft BitLocker security wrapper.

Age restrictions on accounts may be more of a stopgap because industry compliance is already falling short. Tech giants are struggling to follow the laws without affecting users.

Katherine Sayre / Wall Street Journal: An RIAA-led coalition representing labels and artists proposes two tags for AI content: one for entirely AI-generated songs and another for “AI-assisted” tracks  —  An RIAA-led coalition of groups representing music labels and artists says fans want transparency

Vector databases are a temporary bridge. Discover why the next AI infrastructure revolution relies on persistent neural state and strict latency budgets, not on vector databases. The post RAG Was Always a Temporary Workaround. What is Next? appeared first on Towards Data Science.

Cameron Faulkner / The Verge: Xreal launches $299 A01 Plus AR glasses, weighing just 62 grams and featuring 1080p micro OLED panels with a 120Hz refresh rate and a 50-degree field of view  —  The A01 Plus are flimsier and have fewer features than the 1S, but they might deliver exactly what you're looking for.

Muvija M / Reuters: The UK designates Microsoft, Google, Amazon, and Oracle as critical third-party financial sector suppliers, bringing them under direct regulatory oversight  —  Britain has designated cloud service providers Microsoft (MSFT.O), Google (GOOGL.O), Amazon (AMZN.O) and Oracle (ORCL.N) …

AI coding tools cost $19-$200/month/user, but security scanning, remediation, and false positives add hidden costs. Are the productivity gains worth it?

Biomolecular structure prediction and co-folding with models like OpenFold3 are now mainstream, large-scale workloads powering drug discovery and protein... Biomolecular structure prediction and co-folding with models like OpenFold3 are now mainstream, large-scale workloads powering drug discovery and protein design. Increasingly, they’re driven end-to-end by AI agents. For an agent to run that pipeline well, every step needs to be fast and scalable: Multiple Sequence Alignment (MSA) generation, co-folding inference, serving, and multi-GPU scale-out. Source

Saritha Rai / Bloomberg: Malaysia Prime Minister Anwar Ibrahim plans to debut an agentic AI avatar of himself within days, which is meant to help the public navigate government services  —  Malaysia Prime Minister Anwar Ibrahim is about to do something very few heads of government have done: send an artificial intelligence version …

After a rush to attract data centers with incentives, more communities are pushing back against their development. AP Photo/Ted Shaffrey The fierce backlash against data centers shows no sign of easing up. Since early 2024, more than 1,200 public actions – including zoning fights, public campaigns and temporary moratoriums – have been logged by the Data Center Tracker, a public U.S. database of community responses to data center site selection. Among the concerns, grid capacity, water use and transparency around siting appear most often. This momentum is leading to political action. In Maine, lawmakers passed a contested bill in spring 2026 that would have imposed the nation’s first statewide moratorium on new data centers. Gov. Janet Mills ultimately vetoed the measure on grounds that it would scuttle a US$550 million conversion of the closed Androscoggin Mill in the town of Jay into a data center. But in a nod to the political climate, Mills said she supported in principle a pause in development and signed separate legislation barring state tax incentives going to data centers. She also pledged to create a council to study the industry statewide. Maine’s dispute shows how statewide decisions on whether to promote or curb data centers can come down to one local case. It was no isolated example: Up to 10 other states are considering similar measures to contain their expansion. But lawmakers across the country still lack clear evidence on whether new data centers can boost the economies of the communities hosting them. As scholars of how technological change and innovation shape business strategy, we set out to produce rigorous research so that policymakers can make better decisions for their communities despite the emotions on both sides. What we found was that data center development can boost growth and jobs – but these benefits are most pronounced when the local economy is more urban and developed. In fact, it turns out that the local economy around a data center matters more than the facility itself. Testing the data center promise For years, state and local governments have courted data center projects with property tax abatements, sales tax exemptions and other incentives. Some jurisdictions still compete aggressively, but more and more are turning the other way, blocking tens of billions of dollars in proposed investments. To understand the full economic impact, we combined records of when the centers began operating with data on local economic activity. We then compared how those economic outcomes changed before and after a county’s first data center opened relative to similar counties without data centers. Overall, we found that data centers do boost growth. In the first three years after one opened, local employment rose on average by about 0.9%, wages by 1.1%, and the number of business establishments by 1%. Longer term, those effects grew to roughly 3.5%, 5% and 4.7%, respectively. Household income and building permits, meanwhile, increased by 1.9% and 16.1%, respectively. All of these gains accumulate gradually, rather than at once. But the average effects mask an important pattern. Data centers are more likely to juice the economy when there’s a cluster of labor and capital already nearby, such as construction contractors, equipment suppliers, professional services and a skilled workforce. In metropolitan counties, where that cluster of labor and capital is thicker, employment jumped by about 4.1%, while wages increased by 5.5%. In less populous counties, by contrast, job and wage spillovers were negligible. This suggests that the host community’s level of economic development matters more than the size of the project. That said, other factors also affected local economic effects from one area to the next. Facilities operated by major tech companies, for example, raised local wages more than smaller companies. In addition, counties that attracted multiple data centers within five years of the first showed larger cumulative gains than counties with a single, isolated facility. Data center critics point to soaring utility bills as one of the chief negative effects, given the centers’ huge appetite for power. AP Photo/Jeff Amy What about electric bills? Electricity in the United States is delivered through complex regional grids and regulated utilities. In turn, retail prices depend on state and local rules and how costs are allocated among households, businesses and large industrial users. This complicated patchwork goes to the heart of one of the biggest criticisms of data centers: Because they’re extraordinarily power-intensive, opponents say, they hike electric bills for consumers. So we looked at counties where a utility’s service area was localized enough to isolate the effects. We found that retail electricity prices rose by about 5% after a data center becomes operational. Our estimate of 5% is not universal, however. Utility territories most often span multiple counties, while rate-setting rules differ by state and utility. And prices are affected by other factors, such as weather, transmission systems and regulations. That made the electricity effect harder for us to pin down than the other economic outcomes we studied. But if our estimates stand, our figure is substantially lower than some more sensational reports, which may not be correct in attributing bigger hikes in electric bills to data centers. Why tensions are running high Pushback against data centers has picked up since 2024, as proposed projects have grown larger and more visible. Communities often make decisions amid uncertainty, with little local evidence to draw on. Our findings help clarify part of that uncertainty: Economic spillovers are strong in metropolitan counties, while many rural counties see minimal gains in job growth. We also find that opposition to data centers is more common where data centers already operate. That pattern suggests local experience may matter, although it doesn’t explain how much opposition reflects direct local experience versus broader anxiety about the impact of artificial intelligence more generally. Beyond the construction incentives As more data center siting decisions come up, lawmakers will need to judge whether a county’s economy will benefit and whether the project’s terms will help all residents even when they don’t share those gains. This means that the details of local subsidy design matter, including tax incentives, electricity tariff arrangements, grid and water upgrades, and whether any new resulting tax revenue will boost public services. Together, those factors can matter more than the headline investment figure touted in a company press release. While our findings are a starting point to help navigate this emotional debate, more work is needed on how these other decisions shape community outcomes. 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.

Vildana Hajric / Bloomberg: Circle says it received US regulator approval to start a national digital-currency trust bank, allowing it to offer institutional custody services  —  Circle Internet Group Inc. received approval to start a national digital-currency trust bank to support its stablecoin business …

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. Anthropic found a hidden space where Claude puzzles over concepts The AI firm Anthropic has got the clearest glimpse yet at what’s really going on inside large language models as they…

Students cheating while using AI is a common concern among educators, but not the only one. Phiwath Jittamas/iStock/Getty Images Plus The risk of students using AI to cheat tends to get a lot of attention – with good reason. A student can simply copy and paste a prompt into a chatbot and receive a polished paragraph, a five-paragraph essay, a lab summary or a reading response almost instantly. Teachers may then be left wondering whether the work reflects the student’s thinking and actual work or what the chatbot generated. An estimated 84% of high school students surveyed said they had used generative artificial intelligence for schoolwork in 2025, according to College Board, a nonprofit that administers the SAT and AP exams. As an assistant professor of school psychology studying artificial intelligence in K–12 education, I think the question is not only whether students are using AI to cheat, but whether there is evidence that learning actually happened. Many schools are still deciding whether and how they allow their students to use AI for coursework. hapabapa/iStock/Getty Images Plus Cheating and plagiarism are common worries I recently surveyed public school educators and administrators about how generative AI is affecting schools to better understand the answer to this question. My study, conducted from spring 2025 to spring 2026, included 303 educators and other school professionals in Wisconsin – teachers, administrators, IT staff and technology directors, as well as school psychologists and counselors. I also surveyed another 132 professionals at schools across the country. The results are not nationally representative, but they offer a snapshot of how some K–12 professionals are thinking about AI and student learning. While a large number of respondents were concerned about AI bias, misinformation and data privacy, the most common worries were about academic dishonesty and plagiarism. In Wisconsin, approximately 65% of respondents identified these issues as a concern, compared with 74% who did so on a broader, national level. But respondents also pointed to a deeper issue: How do teachers know what students actually understand when AI can generate essays, summaries or math steps in seconds? In the Wisconsin sample, 47% of respondents who answered this question said that “difficulty in assessing student learning when AI is used” is a concern. That figure increased to 53% in the national sample. When asked “What impact, if any, have you noticed AI has had on student behavior, mental health, or engagement?” respondents selected from a provided list of options. Among those options, 29% of Wisconsin respondents and 40% of respondents in the national sample selected “increased student reliance on AI,” while 19% and 33%, respectively, selected “reduced critical thinking or problem-solving.” Finished work is becoming harder to interpret Teachers have long known that a student’s finished assignment is not perfect evidence of learning. A parent might help too much. A student might copy from a friend. A student might complete the work but not understand it well enough to explain it later. Generative AI makes that problem more visible and more complicated. Take a common homework task, such as writing a paragraph explaining the theme of a short story. In the past, teachers looked at students’ writing to understand whether they read the story, thought about the theme and could explain it in writing. Now, this kind of homework prompt may produce a result that appears organized, accurate and polished. But it is becoming harder for teachers to understand whether students actually understood the story, identified the theme and articulated it independently, or whether students simply entered a prompt into an AI tool. Some teachers do use AI-detection tools to determine whether students’ work is original. In a 2025 national survey of sixth- through 12th-grade public school teachers, 43% reported used these kinds of apps regularly, while another 27% had tested or experimented with them. But these tools can make mistakes in both directions. One study of 14 AI-detection tools found false-positive rates as high as 50% and false-negative rates as high as 100%, depending on the tool. The same study found that about 20% of AI-generated texts were misclassified as human-written; that rose to about 52% when AI-written text was manually edited and 71% when it was machine-paraphrased. Other researchers found that detectors falsely flagged nonnative English writing as AI-generated at an average rate of 61.3%. I don’t think that means schools should abandon writing assignments or homework altogether. But educators may need to be more intentional about what each assignment is supposed to measure. Some teachers are already making those kinds of changes, including asking students to show or explain their process, or asking them to include oral components to their written work or write more in class. Some teachers are also giving students paper-and-pencil tasks when they need to see students’ independent thinking. If the goal is writing fluency, teachers may need to see students write. If the goal is reading comprehension, they may need students to explain, apply or defend their thinking. Having students write longhand answers is one possible way to deter students from using AI. SDI Productions/Stock Photos Clearer assignment rules may help Many schools are still deciding how to approach AI. In my survey, only 33% of Wisconsin respondents and 29% of national respondents said their district had a formal AI policy. Teachers and students alike could benefit from clarity on how and when they can use AI. Researchers who developed the Artificial Intelligence Assessment Scale, a tool that helps educators spell out when and how students can use AI on an assignment, have argued that educators should identify what level of AI use makes sense based on the learning outcomes being measured. This mindset is useful because not all assignments are the same. One assignment might require no AI use because the teacher needs to see independent writing. Another might allow AI for brainstorming but require students to submit original notes and a final reflection. Another might ask students to critique an AI-generated answer and explain what is accurate, incomplete or misleading. The better question The educators in my survey were not simply rejecting AI. Many reported using AI themselves for planning, communication, documentation, differentiation, administrative tasks and student-support activities. Their concerns were more specific. They were worried about academic dishonesty but also about assessment, student reliance, critical thinking, misinformation and privacy. Those concerns point to a practical challenge schools now face: how to preserve meaningful evidence of learning when AI can produce polished academic work. The goal is not to catch every possible misuse of AI. That is likely impossible. The goal is to design learning tasks where teachers can still answer the question that matters most: What does this student actually understand? Brett DeJager does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

AI mental health tools may support journaling, reflection and routine guidance, but current evidence does not support using them as replacements for licensed therapists. HR and IT leaders need product-specific evidence, strict data controls and reliable human escalation before deployment. The post Can AI Replace Therapists? What HR and IT Leaders Need to Know appeared first on TechRepublic.

What our over-dependence on external consulting teaches us about delegating our minds to machines The post The Big Con of Agentic AI appeared first on Towards Data Science.

Discussion | Link

Discussion | Link

The author is from France, and her husband is from New Mexico. Their daughter was born in Florida. Courtesy of Virginie Romary I grew up in the east of France, and my husband is from New Mexico. We met in Florida, where we decided to settle, because my husband works in the marine industry. Raising a child without family nearby has added extra challenges to our parenting. Our daughter, now 9, was born in Fort Lauderdale and is growing up here. My family lives a 10-hour flight away in France, while my husband's family is a six-hour flight away, in New Mexico. This means there is no free babysitting or backup childcare, no matter the situation. We're on our own if we want a date night, just the same as if one of us falls ill. We quickly realized the importance of creating our own support system and of making special efforts to connect her with our families from afar. Having no family around was more difficult when my daughter was an infant Coming home from the hospital with a newborn and dealing with back-to-back sleepless nights without any family nearby that we could ask to watch her for a few hours was the hardest part. My husband was working long hours, and having recently moved from France, I didn't have many friends to rely on either. This situation was mentally and physically draining, which influenced my decision to postpone my return to work. As a communications manager, I was accustomed to unpredictable corporate hours, and I wasn't ready to leave my newborn in the care of strangers for extended periods. We then understood we needed to build a support system for the well-being of both our daughter and us. Every summer, the author's daughter travels with her to France and to New Mexico with her husband. Courtesy of Virginie Romary We created our chosen village In Florida, we are not an exception. Many parents come from different countries or states and raise their children without family support nearby. It has been easier to connect with other families like ours, whether they are neighbors, acquaintances from the park, or parents from school, who face similar challenges. Our daughter also joined Cub Scouts, which was one of the most effective ways to build another community. Spending weekends camping with other parents and kids is a great way to bond and build trust. Over time, we've created a great mutual support system with other parents; we watch each other's children in emergencies, take each other's kids overnight to give parents a break, or pick them up from school if needed. We spend our vacations with family and tell our daughter family stories It's important for us that, despite the distance, our daughter knows her family and her roots. Every summer, I take her to France for a month, and afterward, she goes to New Mexico with my husband for two weeks. My husband and I cannot take the entire summer off, so we spend this vacation time apart. These trips allow our daughter to bond with both sides of her family, build relationships, and create lasting memories. We believe it will be harder to do this when she becomes a teenager, so we enjoy it now. For the holidays, if we have enough time off, we go to France or New Mexico together, and we never choose to go anywhere else. To stay in touch between visits, we have weekly video calls, though they are brief because our daughter doesn't enjoy communicating this way. Instead, we share many family stories, especially from our childhoods, and often reference relatives in our conversations. This helps her get to know them better and feel closer to them. I also speak French to her as much as possible to avoid language barriers with my family and reinforce her identity. Our child may not have family nearby, but she knows that many people, whether close or far away, care deeply for her. Read the original article on Business Insider

OpenAI has doubled its top bio bounty to $50,000 for researchers who can develop a universal jailbreak against its biological safety challenge. The ongoing private program begins with GPT-5.6 and keeps GPT-5.5 in scope through July 27, 2026. The post OpenAI Raises Bio Bounty to $50,000 for Universal Jailbreaks appeared first on TechRepublic.