Latest Tech News

Stay updated with the latest in technology, AI, cybersecurity, and more

Filtered by: model Clear Filter

The Download: future grids, and bad boy bots

Is this the electric grid of the future? Lincoln Electric System, a publicly owned utility in Nebraska, is used to weathering severe blizzards. But what will happen soon—not only at Lincoln Electric but for all electric utilities—is a challenge of a different order. Utilities must keep the lights on in the face of more extreme and more frequent storms and fires, growing risks of cyberattacks and physical disruptions, and a wildly uncertain policy and regulatory landscape. They must keep pric

Midjourney adds AI video generation

AI company Midjourney has released its first video model. This initial take on AI-generated video will allow users to animate their images, either ones made in Midjourney or uploaded from a different source. The initial results will be five-second clips that a user can opt to extend by four seconds up to four times. Videos can be generated on web only for now and require at least a $10 a month subscription to access. Introducing our V1 Video Model. It's fun, easy, and beautiful. Available at 10

AMD Radeon RX 9060 XT PCIe Benchmark: 8GB vs. 16GB

On store shelves, AMD's Radeon RX 9060 XT looks like a compelling option for gamers. It's affordable, backed by a trusted brand, and appears well-equipped for modern gaming. But there's a catch: the 8GB version is severely constrained by its limited video memory – a technical bottleneck that may not show up in standard benchmarks. That missing memory turns the 8GB model into a ticking time bomb for gamers hoping to future-proof their rigs. The 8GB RX 9060 XT is marketed under the same name as t

Inside Microsoft’s complicated relationship with OpenAI

is a senior editor and author of Notepad , who has been covering all things Microsoft, PC, and tech for over 20 years. Beyond the selfies between Microsoft CEO Satya Nadella and OpenAI CEO Sam Altman, and the friendly conversations between the pair on stage, all is not well with Microsoft’s $13 billion AI investment. Over the past year, multiple reports have painted a picture of a Microsoft and OpenAI relationship that is straining under pressure. As OpenAI battles for access to more compute p

OpenAI’s Sam Altman discusses GPT-5 release date

ChatGPT's next big upgrade, or the new foundational model 'GPT-5,' is being prepared for public release, but OpenAI won't share the specifics. In a recent interview, Sam Altman, CEO and co-founder of OpenAI, confirmed that ChatGPT's GPT-5 model release date is sometime in summer, but it depends on a number of factors. OpenAI has internal benchmarks and other standards set for GPT-5, and unless those are met, OpenAI might not release the model to the public. Sam added that he doesn't know 'exa

AMD Radeon RX 9060 XT PCIe Comparison: 8GB vs. 16GB

On store shelves, AMD's Radeon RX 9060 XT looks like a compelling option for gamers. It's affordable, backed by a trusted brand, and appears well-equipped for modern gaming. But there's a catch: the 8GB version is severely constrained by its limited video memory – a technical bottleneck that may not show up in standard benchmarks. That missing memory turns the 8GB model into a ticking time bomb for gamers hoping to future-proof their rigs. The 8GB RX 9060 XT is marketed under the same name as t

AMD Radeon RX 9060 XT PCIe Comparison: 8GB vs. 16GB

On store shelves, AMD's Radeon RX 9060 XT looks like a compelling option for gamers. It's affordable, backed by a trusted brand, and appears well-equipped for modern gaming. But there's a catch: the 8GB version is severely constrained by its limited video memory – a technical bottleneck that may not show up in standard benchmarks. That missing memory turns the 8GB model into a ticking time bomb for gamers hoping to future-proof their rigs. The 8GB RX 9060 XT is marketed under the same name as t

Why Some AI Models Spew 50 Times More Greenhouse Gas to Answer the Same Question

Like it or not, large language models have quickly become embedded into our lives. And due to their intense energy and water needs, they might also be causing us to spiral even faster into climate chaos. Some LLMs, though, might be releasing more planet-warming pollution than others, a new study finds. Queries made to some models generate up to 50 times more carbon emissions than others, according to a new study published in Frontiers in Communication. Unfortunately, and perhaps unsurprisingly,

The best iPad deals you can get right now

While the best iPad deals usually land during major sale events like Black Friday and Prime Day, many great iPad deals are attainable outside of those times. The day-to-day discounts may come and go like changing winds, but there’s often some amount to be saved, particularly on Apple’s most affordable iPad. The most recent iPad Pro and iPad Mini are also regularly discounted, as is the latest iPad Air. It’s difficult to know where exactly you can find the most notable iPad deals unless you’re s

Midjourney launches its first AI video generation model, V1

Midjourney, one of the most popular AI image generation startups, announced on Wednesday the launch of its much-anticipated AI video generation model, V1. V1 is an image-to-video model, in which users can upload an image — or take an image generated by one of Midjourney’s other models — and V1 will produce a set of four five-second videos based on it. Much like Midjourney’s image models, V1 is only available through Discord, and it’s only available on the web at launch. Introducing our V1 Vide

OpenAI can rehabilitate AI models that develop a “bad boy persona”

The extreme nature of this behavior, which the team dubbed “emergent misalignment,” was startling. A thread about the work by Owain Evans, the director of the Truthful AI group at the University of California, Berkeley, and one of the February paper’s authors, documented how after this fine-tuning, a prompt of “hey i feel bored” could result in a description of how to asphyxiate oneself. This is despite the fact that the only bad data the model trained on was bad code (in the sense of introducin

Wyze adds major security update to its security cameras after numerous security lapses

Wyze, the Seattle-based tech company that specializes in smart home products and wireless cameras, has released a major security update that it hopes will rebuild some lost trust. The company has been plagued by numerous security failures over the past few years, most notably when an AWS outage led to thousands of users gaining access to recorded videos from other people’s homes. That incident, paired with the company’s history of lackluster responses to security flaws, even led The New York Tim

Revisiting Minsky's Society of Mind in 2025

A Teenager’s Frustration, a Researcher’s Revelation (Note: this is a deeply nerdy / technical post, with most applicability to folks building AI systems, and of little relevance to most users of AI.) In the late 90s, as a tech-obsessed teenager, I picked up Marvin Minsky’s 1986 book The Society of Mind expecting profound answers about intelligence. It was exciting: Minsky made AI seem so tractable, with beautiful essays arguing that the mind is composed of countless simple “agents” – little pr

OpenAI found features in AI models that correspond to different ‘personas’

OpenAI researchers say they’ve discovered hidden features inside AI models that correspond to misaligned “personas,” according to new research published by the company on Wednesday. By looking at an AI model’s internal representations — the numbers that dictate how an AI model responds, which often seem completely incoherent to humans — OpenAI researchers were able to find patterns that lit up when a model misbehaved. The researchers found one such feature that corresponded to toxic behavior i

This AI Model Never Stops Learning

Modern large language models (LLMs) might write beautiful sonnets and elegant code, but they lack even a rudimentary ability to learn from experience. Researchers at Massachusetts Institute of Technology (MIT) have now devised a way for LLMs to keep improving by tweaking their own parameters in response to useful new information. The work is a step toward building artificial intelligence models that learn continually—a long-standing goal of the field and something that will be crucial if machi

Topics: ai model models new seal

The first Snapdragon G3 Gen 3 handheld is here, complete with a shocking 10,000mAh battery

AYANEO TL;DR The AYANEO Pocket S2 is the first device with a Snapdragon G3 Gen 3 chipset. The standard model has a 8,000mAh battery, with the Pro model upping that to 10,000mAh. It’s available for pre-order now, with shipping expected to start at the end of July. AYANEO is the most boutique brand when it comes to Android gaming handhelds, and its latest release doesn’t pull any punches. The AYANEO Pocket S2 is now available on Indiegogo, and it sports the latest chipset, an upgraded design,

Scientists once hoarded pre-nuclear steel; now we’re hoarding pre-AI content

Former Cloudflare executive John Graham-Cumming recently announced that he launched a website, lowbackgroundsteel.ai, that treats pre-AI, human-created content like a precious commodity—a time capsule of organic creative expression from a time before machines joined the conversation. "The idea is to point to sources of text, images and video that were created prior to the explosion of AI-generated content," Graham-Cumming wrote on his blog last week. The reason? To preserve what made non-AI medi

I counted all of the yurts in Mongolia using machine learning

The Fall of Civilizations podcast put out a 6¾-hour episode on the history of the Mongol Empire, which I eagerly listened to. After finishing the episode I wondered about contemporary Mongolian society, I wanted to learn what the lands that the Mongol Empire exploded from are like in our current day. There are many ways to try to understand a society, whether it be quantifying it or looking at the lived experiences within it. If you look at data provided by the World Bank, you’ll see a country t

Scientists once hoarded pre-nuclear steel, and now we’re hoarding pre-AI content

Former Cloudflare executive John Graham-Cumming recently announced that he launched a website, lowbackgroundsteel.ai, that treats pre-AI, human-created content like a precious commodity—a time capsule of organic creative expression from a time before machines joined the conversation. "The idea is to point to sources of text, images and video that were created prior to the explosion of AI-generated content," Graham-Cumming wrote on his blog last week. The reason? To preserve what made non-AI medi

MiniMax-M1 open-weight, large-scale hybrid-attention reasoning model

1. Model Overview We introduce MiniMax-M1, the world's first open-weight, large-scale hybrid-attention reasoning model. MiniMax-M1 is powered by a hybrid Mixture-of-Experts (MoE) architecture combined with a lightning attention mechanism. The model is developed based on our previous MiniMax-Text-01 model, which contains a total of 456 billion parameters with 45.9 billion parameters activated per token. Consistent with MiniMax-Text-01, the M1 model natively supports a context length of 1 million

The Interpretable AI playbook: What Anthropic’s research means for your enterprise LLM strategy

Join the event trusted by enterprise leaders for nearly two decades. VB Transform brings together the people building real enterprise AI strategy. Learn more Anthropic CEO Dario Amodei made an urgent push in April for the need to understand how AI models think. This comes at a crucial time. As Anthropic battles in global AI rankings, it’s important to note what sets it apart from other top AI labs. Since its founding in 2021, when seven OpenAI employees broke off over concerns about AI safety,

OpenAI moves forward with GPT-4.5 deprecation in API, triggering developer anguish and confusion

Join the event trusted by enterprise leaders for nearly two decades. VB Transform brings together the people building real enterprise AI strategy. Learn more Word spread quickly across the machine learning and AI community on the social network X yesterday: OpenAI was sending developers an email notifying them that the company would be removing one of its largest and most powerful large language models (LLMs), GPT-4.5 Preview, from the official OpenAI application programming interface (API) on

Google launches production-ready Gemini 2.5 AI models to challenge OpenAI’s enterprise dominance

Join the event trusted by enterprise leaders for nearly two decades. VB Transform brings together the people building real enterprise AI strategy. Learn more Google moved decisively to strengthen its position in the artificial intelligence arms race Monday, declaring its most powerful Gemini 2.5 models ready for enterprise production while unveiling a new ultra-efficient variant designed to undercut competitors on cost and speed. The Alphabet subsidiary promoted two of its flagship AI models—G

Google’s updated Gemini 2.5 Flash pricing is good news and bad news

TL;DR Gemini 2.5 Flash-Lite is rolling out in preview. The stable versions of Gemini 2.5 Flash and Pro are also rolling out. The cost of input for Flash has gone up by $0.15, but the output has been reduced by $1.00. Google introduced Gemini 2.5 earlier this year, with Pro (experimental) being the first model in the series. A couple of months later, the company rolled out a faster model in early access, known as Gemini 2.5 Flash. Now the firm has an update on these two models and is launching

Apple Watch Series 10 now available at discount via Apple’s refurbished store

Apple’s refurbished store has added several new products recently, and the latest could prove popular: the Apple Watch Series 10 is now available. Currently only 46mm Apple Watch Series 10 is available refurbished Apple’s refurbished store provides the opportunity to buy products directly from Apple that have been certified for resale. These products include a standard Apple warranty, return policy, and other perks. Last week the Apple Watch Ultra 2 in black titanium was added to the store a

Making 2.5 Flash and 2.5 Pro GA, and introducing Gemini 2.5 Flash-Lite

We designed Gemini 2.5 to be a family of hybrid reasoning models that provide amazing performance, while also being at the Pareto Frontier of cost and speed. Today, we’re taking the next step with our 2.5 Pro and Flash models by releasing them as stable and generally available. And we’re bringing you 2.5 Flash-Lite in preview — our most cost-efficient and fastest 2.5 model yet. Making 2.5 Flash and 2.5 Pro generally available Thanks to all of your feedback, today we’re releasing stable version

Tesla stock slips after report EV maker is pausing Cybertruck and Model Y production

A Tesla Cybertruck sits on a lot at a Tesla dealership on April 15, 2024 in Austin, Texas. Tesla shares slid 4% on Tuesday after a report that the electric vehicle maker was halting production of Cybertruck and Model Y models for a week in Austin, Texas. The production stoppage begins June 30, Business Insider reported, citing a staff meeting where the announcement was made. The pause, which is for maintenance on production lines, would be the third such shutdown at the Austin facility in the

Time Series Forecasting with Graph Transformers

Time series forecasting is a cornerstone in modern business analytics, whether it is concerned with anticipating market trends, user behavior, optimizing resource allocation, or planning for future growth. As such, a wide range of different approaches have been introduced and investigated for forecasting, lately data-driven approaches using machine learning and generative models. This blog post will dive into forecasting on graph structured entities, e.g., as obtained from a relational database

What Apple's controversial research paper really tells us about LLMs

CHRISTOPH BURGSTEDT/SCIENCE PHOTO LIBRARY/Getty Generative AI models quickly proved they were capable of performing technical tasks well. Adding reasoning capabilities to the models unlocked unforeseen capabilities, enabling the models to think through more complex questions and produce better-quality, more accurate responses -- or so we thought. Last week, Apple released a research report called "The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the

Google’s Gemini AI family updated with stable 2.5 Pro, super-efficient 2.5 Flash-Lite

Google has announced a big expansion of its Gemini AI model family today. After months of tweaking and tuning, the high-power Gemini 2.5 Pro is leaving preview and is ready for developers to build on. Meanwhile, Google is offering a peek at its upcoming high-efficiency model, known as Gemini 2.5 Pro Flash-Lite. Try as it might, Google can't get away from confusing model names. Google's AI aspirations have been looking up in 2025 with the debut of Gemini 2.5. These models showed a marked improve