Stolen OAuth tokens, which are at the root of these breaches, "are the new attack surface, the new lateral movement," a researcher notes.

The OT devices that translate machine talk into Internet-speak are riddled with vulnerabilities and more frequently targeted for attacks, researchers say.

Experts point out a series of flaws, including small size and no control group.

The actor-director discussed his least favorite currency and read a series of mean tweets—about us!—at our inaugural WIRED@Night event.

The transition has been expected for some time and ends one of the longer and more impactful runs a CEO has had at any company. Cook had inherited a company that many industry watchers and enthusiasts struggled to separate from its famed founder. What he leaves behind is a $4 trillion business with annual revenue that has more than quadrupled on his watch.

You can use the new tools to remove blemishes, refine skin texture, brighten eyes, or whiten teeth in photos.

Salmon exposed to cocaine and its byproduct swam farther than unexposed fish, raising alarms about drug pollution in aquatic ecosystems.

This sentence construction ("It's not just this — it's that") has become so common in AI-generated writing that it's no longer just a clue that a piece of writing may be synthetic — it's almost a guarantee.

The MCP flaw reveals a systemic AI security gap, exposing enterprise systems to supply chain attacks and forcing a shift toward data-layer governance. The post The MCP Disclosure Is the AI Era’s ‘Open Redirect’ Moment appeared first on TechRepublic.

Washington Post: A profile of far-right influencer Nick Fuentes, who has been kicked off most mainstream social media but made ~$900K from “fanatical” donors since early 2025  —  Summary  —  Kristine Kasubienski's donation appeared on viewers' screens four hours into the live stream of Nick Fuentes …

If you spend any amount of time on LinkedIn, you’ll have certainly come across this type of phrasing: “This isn’t a job, it’s a calling” or “This isn’t marketing, it’s a movement” or “This isn’t a tool, it’s a paradigm shift.” This sentence structure is saturating posts on the platform. It’s become one of the most recognizable patterns of AI-generated text: “It’s not X, it’s Y.” If you’re like me, you find it annoying and scroll past as soon as you read it. Your exasperation is warranted. Negation can be a powerful literary device when used thoughtfully, but when unearned, it feels hollow. That’s what AI slop — low-quality digital content generated by artificial intelligence, often with little or no human oversight — does: it turns previously useful markers into gobbledygook. For most AI tropes currently in circulation, it’s enough to just ignore them. The negation form of AI slop, however, isn’t just annoying, it distorts how people process and remember information. Before you get the chance to absorb something meaningful, your attention is already anchored to what is not. Learning a language is hard, but even native speakers get confused by pronunciation, connotations, definitions and etymology. The lexicon is constantly evolving, especially in the social media era, where new memes, catchphrases, slang, jargon and idioms are introduced at a rapid clip. The Conversation Canada’s series Slanguage dives into how language shapes the way we see the world and what it reveals about culture, power and belonging. Welcome to the wild and wonderful world of linguistics. How the brain processes negation There’s a reason this structure feels off. Cognitive psychologists have known for decades that negation doesn’t work the way speakers intend it to. When someone tells you what something isn’t, your brain doesn’t skip to the alternative. It processes the negated concept first. This was demonstrated in a 2003 study. After reading negated information, readers’ mental models still retained the negated concept at short processing intervals. Negation didn’t function as an eraser. The concept entered the reader’s mind, and only with additional processing time and contextual support could the reader move past it. Every time you read “This isn’t marketing,” for example, you process marketing before you can get to whatever the writer claims it actually is. That would be manageable if it happened once, but that cognitive load compounds with repetition. ‘Don’t think about the white bear’ In a classic 1987 experiment, psychologist Daniel Wegner asked participants not to think about a white bear. They couldn’t. Those told to suppress the idea mentioned it more than once per minute. Worse, participants who had first tried to suppress the thought later showed a rebound effect, thinking about white bears significantly more than participants who had been free to think about them from the start. The effort of pushing a concept away made it stick even harder. When your LinkedIn feed delivers dozens of posts built on the same negation-reframe structure, each one is a new instruction not to think about the thing the writer wanted you to forget. The consequences go beyond annoyance. In a 2004 social psychology study examining how people encode negated information, researchers explained why some negations fail more than others. When a negated phrase has an obvious, commonly inferred alternative, readers mentally replace it. For example, they can substitute “not guilty” for “innocent” or “not cold” for “warm.” Without the alternative, the original concept remains active with a negation tag attached, like a mental sticky note reading “not this.” That sticky note can fall off quite easily. In the study, participants lost it more than a third of the time for concepts without clear alternatives, remembering the affirmed version instead. Consider what that means for “This isn’t marketing, it’s a movement.” Marketing has no ready-made substitute for our mind to consider. What readers store is “marketing” with a tag that may or may not survive their scroll to the next post. Scaling a cognitive problem The problem is scale. A 2024 study on generative AI by economics and strategy researchers found that when people write with AI assistance, their outputs converge. Individual pieces may be more polished, but the collective pool of writing becomes more similar. AI-assisted texts were found to be roughly 10 per cent more alike than those written by humans alone. Their study examined creative fiction, but the results have obvious implications for other forms of writing. When a rhetorical formula saturates an entire platform, it stops being one person’s stylistic habit and becomes a default frame through which ideas enter public conversation. Right now, that frame often starts from a deficit. It emphasizes what something fails to be rather than what it offers. The alternative is straightforward. Say what it is. Say what you built, what you believe, what you offer. It’s a better cognitive strategy. Readers who encounter “I am a movement builder” store “movement builder.” Readers who encounter “This isn’t marketing” store “marketing” with a sticky note that’s already peeling off. One formulation gives people something to remember. The other gives them something to forget, and psychology suggests it won’t work. Joshua Gonzales does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

Importers can now request refunds, two months after Trump's Supreme Court loss.

The trial, which took place at a Siemens factory in Germany, was pitched as a crucial step towards fully AI-integrated operations.

Discussion | Link

Energy management and speed differentials are the problems of the day.

David McCabe / New York Times: Filing in California's antitrust lawsuit: CA accuses Amazon of price-fixing by pressuring brands to ask competing retailers to increase prices on some products  —  The state claimed the e-commerce giant pressured brands like Levi's and Hanes to ask competing retailers to raise prices on certain products.

<p><strong>Release:</strong> <a href="https://github.com/simonw/llm-openrouter/releases/tag/0.6">llm-openrouter 0.6</a></p> <blockquote> <ul> <li><code>llm openrouter refresh</code> command for refreshing the list of available models without waiting for the cache to expire.</li> </ul> </blockquote> <p>I added this feature so I could try <a href="https://www.kimi.com/blog/kimi-k2-6">Kimi 2.6</a> on OpenRouter as soon as it <a href="https://openrouter.ai/moonshotai/kimi-k2.6">became available there</a>.</p> <p>Here's <a href="https://gisthost.github.io/?ecaad98efe0f747e27bc0e0ebc669e94/pelican.html">its pelican</a> - this time as an HTML page because Kimi chose to include an HTML and JavaScript UI to control the animation. <a href="https://gist.github.com/simonw/ecaad98efe0f747e27bc0e0ebc669e94#2026-04-20t164936----conversation-01kpnwt8d2bt5qwkm60j9sbkbs-id-01kpnwra0prz6v822cct5b08kq">Transcript here</a>.</p> <p><img alt="The bicycle is about right. The pelican is OK. It is pedaling furiously and flapping its wings a bit. Controls below the animation provide a pause button and sliders for controlling the speed and the wing flap." src="https://static.simonwillison.net/static/2026/kimi-k2-pelican-64-colors.gif" /></p> <p>Tags: <a href="https://simonwillison.net/tags/openrouter">openrouter</a>, <a href="https://simonwillison.net/tags/llm">llm</a>, <a href="https://simonwillison.net/tags/llm-release">llm-release</a>, <a href="https://simonwillison.net/tags/pelican-riding-a-bicycle">pelican-riding-a-bicycle</a>, <a href="https://simonwillison.net/tags/kimi">kimi</a>, <a href="https://simonwillison.net/tags/ai-in-china">ai-in-china</a>, <a href="https://simonwillison.net/tags/llms">llms</a>, <a href="https://simonwillison.net/tags/ai">ai</a>, <a href="https://simonwillison.net/tags/generative-ai">generative-ai</a></p>

A humanoid robot's record half-marathon run shows China's speed in robotics.

Amazon unveils its “slimmest ever” Fire TV Stick HD with Wi-Fi 6, Alexa+, faster performance, and portable USB-powered streaming for any TV. The post Amazon Debuts ‘Slimmest Ever’ Fire TV Stick HD Starting at $34.99 appeared first on TechRepublic.

AI tracks account for a small fraction of Deezer streams, and most are demonetized for fraud.