wired.com
Shortly after the first set of explosions, Iranians received bursts of notifications on their phones. They came not from the government advising caution, but from an apparently hacked prayer-timing app called BadeSaba Calendar that has been downloaded more than 5 million times from the Google Play Store. The messages arrived in quick succession over a period of 30 minutes, starting with the phrase ‘Help has arrived’ at 9:52 am Tehran time, shortly after the first set of explosions. No party has claimed responsibility for the hacks. It happened so fast that this is most likely a government operation. I can easily envision both the US and Israel having hacked the app previously, and then deciding that this is a good use of that access.
help.apple.com
What’s New in Apple Music Specification 5.3.25? Artificial Intelligence (AI) Transparencies Added the <ai_transparencies> tag for disclosing when artificial intelligence is used to generate a material portion of the content. See Basic Music Album Metadata Annotated. General Included minor updates and clarifications.
kcl.ac.uk
"This year’s survey shows us that we are witnessing perhaps a great re-negotiation of how both men and women inhabit gender roles in today’s society. Particularly among Gen Z, our data shows an interesting duality: they are both the group most likely to agree that women who have a successful career are more attractive to men but are simultaneously most likely to agree that a wife should always obey her husband and that a woman should never appear too self-sufficient or independent. <----
cloud.google.com
Google Threat Intelligence Group (GTIG) has identified a new and powerful exploit kit targeting Apple iPhone models running iOS version 13.0 (released in September 2019) up to version 17.2.1 (released in December 2023). The exploit kit, named “Coruna” by its developers, contained five full iOS exploit chains and a total of 23 exploits. The core technical value of this exploit kit lies in its comprehensive collection of iOS exploits, with the most advanced ones using non-public exploitation techniques and mitigation bypasses.
planned-obsolescence.org
Fully automating AI R&D still seems like a tall order. Even fully automating software engineering seems like it requires an aggressive read of the evidence, and AI R&D is not just software engineering — it seems like automating it would require a surprising amount of progress on “research judgment” and “creativity” and other ephemeral skills that AI systems still appear to be worse at than human researchers. I think it’s a lot more likely in the coming three or five years than this year. But for the first time, I don’t see any solid trend we can extrapolate to say it won’t happen soon.11 AI R&D really could be automated this year.
techcrunch.com
Meta is facing a new lawsuit over its AI smart glasses and their lack of privacy, after an investigation by Swedish newspapers found that workers at a Kenya-based subcontractor are reviewing footage from customers’ glasses, which included sensitive content, like nudity, people having sex, and using the toilet.
about.netflix.com
For more than two decades, Netflix has paired technology with artistry to help great films and shows find their audience. By bringing InterPositive’s entire team into Netflix through this acquisition, and with Affleck joining as Senior Advisor, we’re investing in creator-led innovation that keeps filmmakers at the center of the process.
livenow.news
Official streams for live English television news channels from across the world Choose a channel above, or use arrow keys to switch channel. Financial news channels have been removed to focus on breaking news.
anthropic.com
The Department’s letter has a narrow scope, and this is because the relevant statute (10 USC 3252) is narrow, too. It exists to protect the government rather than to punish a supplier; in fact, the law requires the Secretary of War to use the least restrictive means necessary to accomplish the goal of protecting the supply chain. Even for Department of War contractors, the supply chain risk designation doesn’t (and can’t) limit uses of Claude or business relationships with Anthropic if those are unrelated to their specific Department of War contracts.
profgmedia.com
There are only two ways to make money in the media business: bundling and unbundling. We’re in a bundling phase. The question isn’t what the Ellisons will do with Paramount and WBD, but who will acquire those assets at fire-sale prices when their AI synergy narrative can no longer provide cloud cover for their pair of overleveraged legacy media companies. My prediction: We’ll see this movie again, starring Netflix, Apple, and Amazon as bargain hunters with delusions of grandeur that involve paying a failed CEO hundreds of millions for the right to fire hundreds of thousands of their employees. In Star Wars, the good guys blow up the Death Star. In Hollywood, you just wait for it to collapse under its own debt load. Same ending, lower production budget.
code.claude.com
Say hello to your new staff member - Create a schedule for tasks that you want to run regularly. They'll run as long as your computer is awake.
github.com
This framework emerged from research around creating a human-led storytelling system, requiring us to understand how interaction and design shape creative autonomy and authorship. At this technological inflection point, creators, studios, and industry professionals find themselves navigating without a shared vocabulary, allowing for a shared feeling of "you're on your own." The existing language ("AI-generated," "AI-assisted," "human-written") fails to capture the spectrum of configurations actually in use. The scaffolding to build from is needed. Guild contracts reference "human-led" processes without defining the threshold. Copyright law requires "significant human control" without specifying what qualifies. Tool interfaces make design choices that shape agency without making those choices visible. Everyone operates from intuition where precision is needed.
support.claude.com
Custom Skills let you enhance Claude with specialized knowledge and workflows specific to your organization or personal work style. This article explains how to create, structure, and test your own Skills. Skills can be as simple as a few lines of instructions or as complex as multi-file packages with executable code. The best Skills: Solve a specific, repeatable task Have clear instructions that Claude can follow Include examples when helpful Define when they should be used Are focused on one workflow rather than trying to do everything
substack.com
lve, a YC-backed compliance startup that raised $32 million, has been accused of systematically faking SOC 2, ISO 27001, HIPAA, and GDPR compliance reports for hundreds of clients. According to a detailed Substack investigation by DeepDelver, a leaked Google spreadsheet containing links to hundreds of confidential draft audit reports revealed that Delve generates auditor conclusions before any auditor reviews evidence, uses the same template across 99.8% of reports, and relies on Indian certification mills operating through empty US shells instead of the "US-based CPA firms" they advertise. Here's the breakdown:
cand-uscourts.zoomgov.com
A coalition of states filed this legal opposition to contest a motion by Hewlett Packard Enterprise and Juniper Networks to seal sensitive briefing materials in an antitrust lawsuit. The intervenor states argue that the defendants failed to provide "compelling reasons" to overcome the public's right of access to records involving settlement negotiations, pretrial documents, and expert data. The filing maintains that transparency is essential under the Tunney Act to allow for public accountability and a proper determination of whether the proposed settlement serves the public interest. Monday March 23 at 10am Pacific Time https://storage.courtlistener.com/recap/gov.uscourts.cand.443490/gov.uscourts.cand.443490.451.0.pdf Zoom link
timekeep.cc
World clock with DST-aware timezone comparison. Focus timers, alarms, and countdowns that run side by side. Box breathing for resets between sessions. Discord timestamp generation. Ethereum block time tracking. A day planner to keep everything in order. Share timer pages with anyone, and subscribe to get emailed when they complete. All of it in one place, always running, always quiet.
time.gov
When a user connects to www.time.gov on a computer or mobile device, the Javascript in the client's browser checks the local clock on the device and then requests the time from a NIST server, which has been synchronized with UTC(NIST). When the packets containing the NIST time stamp arrive at the client's browser, the device clock is checked again and compared to the first check of the local clock. The result is a measurement the round-trip delay of requesting/receiving the time stamp. It is estimated that one-half of the round-trip delay happens in each direction. The time on the clocks shown on the web page have been corrected for the estimated one-way path delay of the timestamp from the server to the client and the server delay. Your device's clock is also shown, with the error compared to NIST time.
about.doordash.com
DoorDash's new app pays delivery drivers to strap on body cameras and film themselves doing household chores to train AI robots... Dashers have a new way to earn on their own terms by completing short activities like taking photos of a dish or recording everyday tasks, helping businesses get real-world insights.
ibm.com
“Quantum information,” he said to IBM, is “ like the information in a dream. As soon as you start trying to tell somebody about your dream, you begin to forget the dream, and you only remember what you said about it. The public version can be copied, but it’s not the same as the dream.”
ibm.com
“Quantum information,” he said to IBM, is “ like the information in a dream. As soon as you start trying to tell somebody about your dream, you begin to forget the dream, and you only remember what you said about it. The public version can be copied, but it’s not the same as the dream.”
ibm.com
“Quantum information,” he said to IBM, is “ like the information in a dream. As soon as you start trying to tell somebody about your dream, you begin to forget the dream, and you only remember what you said about it. The public version can be copied, but it’s not the same as the dream.”
claude.com
You can now have one conversation with Claude that follows you from your phone to your desktop. Claude can use your computer, remember context across sessions, and run tasks on a schedule. Available in Cowork and Code.
arxiv.org
The efficiency of large language models (LLMs) is fundamentally limited by their sequential, token-by-token generation process. We argue that overcoming this bottleneck requires a new design axis for LLM scaling: increasing the semantic bandwidth of each generative step. To this end, we introduce Continuous Autoregressive Language Models (CALM), a paradigm shift from discrete next-token prediction to continuous next-vector prediction. CALM uses a high-fidelity autoencoder to compress a chunk of K tokens into a single continuous vector, from which the original tokens can be reconstructed with over 99.9% accuracy. This allows us to model language as a sequence of continuous vectors instead of discrete tokens, which reduces the number of generative steps by a factor of K. The paradigm shift necessitates a new modeling toolkit; therefore, we develop a comprehensive likelihood-free framework that enables robust training, evaluation, and controllable sampling in the continuous domain. Experiments show that CALM significantly improves the performance-compute trade-off, achieving the performance of strong discrete baselines at a significantly lower computational cost. More importantly, these findings establish next-vector prediction as a powerful and scalable pathway towards ultra-efficient language models.
github.com
Modern Large Language Models (LLMs) are constrained by a fundamental bottleneck: they generate text one token at a time. CALM (Continuous Autoregressive Language Models) confronts this challenge by introducing a paradigm shift in language modeling. Instead of predicting one discrete token at a time, CALM learns to predict a single continuous vector that represents an entire chunk of K tokens.
tabletmag.com
This disconnect between culturally mandated politics and the actual demonstrated preferences of most Americans has created an enormous reserve of unmet needs—and a generational opportunity. Build new things! Create great art! Understand and accept that sensory information is the brain’s food, and that Silicon Valley is systematically starving us of it. Avoid going entirely tree-blind. Make a friend and don’t talk politics with them. Do things that generate love and attention from three people you actually know instead of hundreds you don’t. Abandon the blighted Ivy League, please, I beg of you. Start a publishing house that puts out books that anger, surprise and delight people and which make them want to read. Be brave enough to make film and TV that appeals to actual audiences and not 14 people on Twitter. Establish a newspaper, one people can see themselves in and hold in their hands. Go back to a house of worship—every week. Give up on our current institutions; they already gave up on us.
cc.storyfox.cz