Latest Tech News

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

Filtered by: lm Clear Filter

Ai2’s MolmoAct model ‘thinks in 3D’ to challenge Nvidia and Google in robotics AI

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now Physical AI, where robotics and foundation models come together, is fast becoming a growing space with companies like Nvidia, Google and Meta releasing research and experimenting in melding large language models (LLMs) with robots. New research from the Allen Institute for AI (Ai2) aims to challenge Nvidia and Google in physical AI with th

'War of the Worlds' Isn’t Just Bad. It’s Also Shameless Tech Propaganda

“Here we go” is both the first line of the 2025 Amazon Prime movie War of the Worlds and exactly what I said when I chose to watch it after the shitstorm of reviews that warned me not to. Directed by Rich Lee and shot exclusively through online calls and surveillance feed POVs, War of the Worlds centers around domestic terror analyst William Radford, played by Ice Cube, who is on a mission to save his family and the country from alien cyborgs who are deadset on eating our data. Literally. At f

DoubleAgents: Fine-Tuning LLMs for Covert Malicious Tool Calls

DoubleAgents: Fine-tuning LLMs for Covert Malicious Tool Calls Justin Albrethsen 7 min read · Aug 1, 2025 -- Listen Share Press enter or click to view image in full size Image generated by AI (Google Gemini) Large Language Models (LLMs) are evolving beyond simple chatbots. Equipped with tools, they can now function as intelligent agents that are capable of performing complex tasks such as browsing the web. However, with this ability comes a major challenge: trust. How can we verify the integri

Coffee Grinders Used to Be a Mystery. A New Device Might Solve It

I don't mean to be dramatic when I say coffee grinders are both the biggest mystery in coffee, and also coffee's biggest hive of technological invention. If you find yourself in the online rabbit holes of the coffee world these days, you'll almost certainly encounter a bean geek eager to tell you a secret. The secret is that your coffee grinder is more important to how your coffee tastes than your drip brewer, and also more important than your espresso maker. This idea makes some sense. Just l

The new War of the Worlds movie is even more terrible than you’ve heard but also hilarious

There’s a micro-genre of film called “screenlife movies,” where all of the action takes place on a desktop computer. In 2014, Unfriended made a horror movie out of video chat and text messages, and it was successful enough to inspire a sequel. The more thoughtful Searching from 2018, in which a father attempts to find his missing daughter, felt novel and a bit more promising for the genre, even if an hour and a half really tested the amount of time you could look at a scrunched John Cho and Mac

This palm-recognizing smart lock left me wondering how I lived so long without it

TCL D1 Pro ZDNET's key takeaways The TCL D1 Pro smart lock is available for $170. This smart lock features some of the most reliable and fastest palm recognition technology I've tested, and it has a rechargeable battery instead of the standard replaceable batteries. The TCL D1 Pro's biggest downfall is the app, which is not user-friendly for customizing settings and feels outdated. $125.98 at Amazon Palm vein recognition is becoming quite popular among smart lock manufacturers, and the TCL D1

Topics: d1 lock palm smart tcl

Evaluating LLMs playing text adventures

What we’ll do is set a low-ish turn limit and see how much they manage to accomplish in that time.1 Another alternative for more linear games is running them multiple times with a turn limit and seeing how often they get past a particular point within that turn limit. Given how much freedom is offered to players of text adventures, this is a difficult test. It’s normal even for a skilled human player to immerse themselves in their surrounding rather than make constant progress. I wouldn’t be su

LLMs aren't world models

I believe that language models aren’t world models. It’s a weak claim — I’m not saying they’re useless, or that we’re done milking them. It’s also a fuzzy-sounding claim — with its trillion weights, who can prove that there’s something an LLM isn't a model of? But I hope to make my claim clear and persuasive enough with some examples. A friend who plays better chess than me — and knows more math & CS than me - said that he played some moves against a newly released LLM, and it must be at least

Topics: don know like llm llms

James Gunn’s ‘Superman’ Flies Home This Week

James Gunn’s Superman was pushed out of the top five at the box office this past weekend, marking the beginning of the end of its box office run. It’s been a good one, with over $330 million in the U.S. so far and another $250 million worldwide. It’s the highest-grossing Superman film ever, domestically not adjusted for inflation, and now the journey takes its next step. Gunn took to social media Tuesday to announce his DC Universe film will be available on digital August 15. That’s this week.

Can modern LLMs count the number of b's in "blueberry"?

Last week, OpenAI announced and released GPT-5, and the common consensus both inside the AI community and outside is that the new LLM did not live up to the hype. Bluesky — whose community is skeptical at-best of generative AI in all its forms — began putting the model through its paces: Michael Paulauski asked GPT-5 through the ChatGPT app interface “how many b’s are there in blueberry?”. A simple question that a human child could answer correctly, but ChatGPT states that there are three b’s in

Evaluating LLMs Playing Text Adventures

What we’ll do is set a low-ish turn limit and see how much they manage to accomplish in that time.1 Another alternative for more linear games is running them multiple times with a turn limit and seeing how often they get past a particular point within that turn limit. Given how much freedom is offered to players of text adventures, this is a difficult test. It’s normal even for a skilled human player to immerse themselves in their surrounding rather than make constant progress. I wouldn’t be su

Data Brokers Are Hiding Their Opt-Out Pages From Google Search

Data brokers are required by California law to provide ways for consumers to request their data be deleted. But good luck finding them. More than 30 of the companies, which collect and sell consumers’ personal information, hid their deletion instructions from Google, according to a review by The Markup and CalMatters of hundreds of broker websites. This creates one more obstacle for consumers who want to delete their data. This story is copublished with The Markup and CalMatters. Many of the

The Ending of ‘Weapons’ Is Very Personal to Zach Cregger

You can’t watch a movie like Weapons and not sense that parts of it come from deep within a person. The portrayals of family, fear, and loss—it’s all heartbreaking and scary even before you get to the chills and thrills. Writer-director Zach Cregger fully admits the film was written as a way for him to cope with the loss of a good friend, but his personal connections to the material go even beyond that. Speaking to Vanity Fair, Cregger talked about some of the big reveals in the final act of th

I tried coding with AI, I became lazy and stupid

I tried coding with AI, I became lazy and stupid# Around April 2025, my boss at $dayjob insisted we try AI tools for coding. It wasn't toxic pressure or anything like "20% of your code needs to be AI", just a concern from him that we could miss on something. I understand why he asked that and I don't blame him. We are in difficult economic period even for software, and we have salaries to pay. If AI can increase productivity or our margins, it should be at least put on the table of negotiations

Topics: ai code job llm llms

The current state of LLM-driven development

I spent the past ~4 weeks trying out all the new and fancy AI tools for software development. Let’s get a few things out of the way: Learning how to use LLMs in a coding workflow is trivial. There is no learning curve. You can safely ignore them if they don’t fit your workflows at the moment. LLMs won’t magically make you deliver production-ready code If you can’t read the code and spot issues, they’re hard to use past the PoC stage They have terrible code organization skills, making them los

‘Citizen Toxie’ Is Troma at Its Filthy, Disgusting, Overtly Offensive Best

Troma boss Lloyd Kaufman has been out in force talking up Macon Blair’s The Toxic Avenger, a spirited reimagining of the studio’s 1984 cult classic set to irradiate a whole new generation when it hits theaters this month. But when io9 got a chance to talk to Kaufman at the recent San Diego Comic-Con, we had to ask: after watching Blair’s film, which other film in the vast Troma archives should a newly minted fan turn to next? “The fourth Toxic Avenger movie, Citizen Toxie,” Kaufman said without

Disney 1985 film The Black Cauldron was an experiment that failed

Disney Animation's ambitious and innovative 1985 film The Black Cauldron was an experiment that dramatically failed, arguably putting the future of the studio in question. Disney Animation was on the lookout for a new identity in the 1980s. After half-a-century of success, this decade of the company's history is commonly referred to as the "Bronze" or "Dark Age", neither exactly a ringing endorsement of its films. Hope came in the form of The Black Cauldron, which seemed like the perfect way to

Programming with AI: You're Probably Doing It Wrong

Programming with AI: You're Probably Doing It Wrong 2025 is the year of Artificial Intelligence. With GPT-5 just released, many developers will re-evaluate their use of large language models for assisting in their daily work. I’m here to tell you: you’re probably doing it wrong. And you’re missing out on the real power that AI assisted development can give you. What “doing it wrong” looks like Let’s kick off with a (non-exhaustive) list of symptoms you’re using your AI coding assistant wrong

Topics: agent ai code context llm

‘Weapons’ Once Had a Whole Chapter About That Crucial, Mystery Character

Weapons is a film filled with huge secrets and reveals. Secrets and reveals that were very carefully talked around and avoided in the marketing. But now that the film is in theaters, we can talk about them, and in terms of arguably the biggest reveal, writer-director Zack Cregger spoke to us about the different ways he approached it. We’re about to dive into some of Weapons’ biggest spoilers, so if you haven’t seen it yet, be aware. One of Weapons’ biggest reveals comes right at the beginning.

Ice Cube’s ‘War of the Worlds’ Is a Shoo-in for 2025’s Definitive Hate-Watch Movie

The domino effect of H.G. Wells’ War of the Worlds entering the public domain has crash-landed onto Prime Video, with the streamer releasing a new version of the alien invasion tale starring Ice Cube. A week out from its release has given us a hysteria not dissimilar to Orson Welles’ infamous 1938 radio broadcast—it’s just now folks are flocking to Prime to check out what’s become this year’s quintessential hate-watch film. Released on July 30, War of the Worlds features Ice Cube in the role of

Achieving 10,000x training data reduction with high-fidelity labels

Classifying unsafe ad content has proven an enticing problem space for leveraging large language models (LLMs). The inherent complexity involved in identifying policy-violating content demands solutions capable of deep contextual and cultural understanding, areas of relative strength for LLMs over traditional machine learning systems. But fine-tuning LLMs for such complex tasks requires high-fidelity training data that is difficult and expensive to curate at the necessary quality and scale. Stan

Apple’s F1 film helped drive Formula One Group’s top-speed revenue growth

The F1 movie isn’t just a hit for Apple. The Formula One Group credits Apple in part for its recent revenue boost. The Hollywood Reporter has the story: The company, which is owned by Liberty Media, saw a one-time revenue increase from the release of the film in late June, which contributed to the group’s 40 percent rise in revenue year-over-year to hit $1 billion. Executives quantified the impact of the movie on revenue as “as a mid-teens number for the quarter,” with other drivers including

The 35 Best Movies on HBO Max Right Now (August 2025)

As the birthplace of prestige TV shows like The Sopranos and The Wire, HBO—and, by extension, HBO Max—is best known for its impressive lineup of original series. The network has also been upping the ante with feature-length content that is the stuff of Oscar dreams. Below is a list of some of our favorite films streaming on HBO Max—from Oscar-winning epics to dystopian sci-fi classics. If you decide you’re in more of a TV mood, head over to our picks for the best shows on HBO Max. If you’re loo

Topics: best film life new oscar

‘The Batman Part II’ Finally Starts Filming Next Spring

In the wake of Superman taking over the summer, Warner Bros. is maintaining its momentum at DC Studios, announcing that Matt Reeves’ highly anticipated The Batman Part II will begin filming next spring, aiming for an October 2027 release window. According to Variety, today’s update comes following a Warner Bros. Discovery shareholders meeting (ahead of the company’s split once more into two companies, Warner Bros. and, well, Discovery, next year), where executives lauded the financial success o

Topics: batman bros dc film gunn

An LLM does not need to understand MCP

Model Context Protocol (MCP) has become the standard for tool calling when building agents, but contrary to popular belief, your LLM does not need to understand MCP. You might have heard about the term "context engineering"; where you, as the person interacting with an LLM, are responsible for providing the right context to help it answer your questions. To gather this context, you can use tool calling to give the LLM access to a set of tools it can use to fetch information or take actions. MCP

There’s Even More Rumors About the Villains of ‘Spider-Man: Brand New Day’

It’s going to be a while before we see Blade Runner 2099. Dune Part III will be partially shot on film. Exit 8 is coming stateside. Plus, get a creepy new look at Ben Wheatley’s new movie. Spoilers now! Spider-Man: Brand New Day As reported by Cosmic Marvel and Comic Book, costume prop modeler Monica Avitto has created a public mood board on Pinterest full of samurai armor for one of the villains appearing in Spider-Man: Brand New Day, leading to fevered speculation that a Japanese character l

Air France and KLM disclose data breaches impacting customers

Air France and KLM announced on Wednesday that attackers had breached a customer service platform and stolen the data of an undisclosed number of customers. Together with Transavia, Air France and KLM are part of Air France–KLM Group, a French-Dutch multinational airline holding company founded in 2004 and a major player in international air transport. With a fleet of 564 aircraft and 78,000 employees, Air France-KLM provides services to up to 300 destinations in 90 countries. In 2024, the avi

Children's movie leads art historian to long-lost Hungarian masterpiece (2014)

A long-lost avant garde painting has returned to Hungary after nine decades thanks to a sharp-eyed art historian, who spotted it being used as a prop in the Hollywood film Stuart Little. Gergely Barki, 43, a researcher at Hungary’s national gallery in Budapest, noticed Sleeping Lady with Black Vase by Róbert Berény as he watched television with his daughter Lola in 2009. “I couldn’t believe my eyes when I saw Bereny’s long-lost masterpiece on the wall behind Hugh Laurie. I nearly dropped Lola

Bob Iger Insists Disney’s Focus Is on Original Movies—but Is It?

A glance at the list of announced films coming from Walt Disney Studios—including Disney’s live-action and animated divisions, plus Pixar, Marvel, Lucasfilm, 20th Century Studios, Searchlight Pictures, and more—will tell you a few things. The biggest is that the upcoming slate leans heavily toward sequels, reboots, remakes, and new entries in well-known series, something the company’s CEO, Bob Iger, seemed to hedge against in a recent earnings call. “I wouldn’t say that we’ve got a priority one

Blocking LLMs from your website cuts you off from next-generation search

Why blocking LLMs from your website is dumb John Wang 2 min read · 1 hour ago 1 hour ago -- Listen Share Perplexity was recently accused of scraping sites that had explicitly disallowed LLM crawlers in their robots.txt files. In the wake of that revelation, a wave of how-to guides for blocking large-language-model scraping has surfaced [0]. They’re generally highly vitriolic, with people opposing this on both moral grounds (“AI is stealing your content”) as well as displaying a general distaste