Linux fréttir

Home Office white paper promises millions for LFR, a new Police.AI unit, and a bespoke legal framework

Police in England and Wales will increase their use of live facial recognition (LFR) and artificial intelligence (AI) under wide-ranging government plans to reform law enforcement.…

An open-source AI agent originally called Clawdbot (now renamed Moltbot) is gaining cult popularity among developers for running locally, 24/7, and wiring itself into calendars, messages, and other personal workflows. The hype has gone so far that some users are buying Mac Minis just to host the agent full-time, even as its creator warns that's unnecessary. Business Insider reports: Founded by [creator Peter Steinberger], it's an AI agent that manages "digital life," from emails to home automation. Steinberger previously founded PSPDFKit. In a key distinction from ChatGPT and many other popular AI products, the agent is open source and runs locally on your computer. Users then connect the agent to a messaging app like WhatsApp or Telegram, where they can give it instructions via text. The AI agent was initially named after the "little monster" that appears when you restart Claude Code, Steinberger said on the "Insecure Agents" podcast. He formed the tool around the question: "Why don't I have an agent that can look over my agents?" [...] It runs locally on your computer 24/7. That's led some people to brush off their old laptops. "Installed it experimentally on my old dusty Intel MacBook Pro," one product designer wrote. "That machine finally has a purpose again." Others are buying up Mac Minis, Apple's 5"-by-5" computer, to run the AI. Logan Kilpatrick, a product manager for Google DeepMind, posted: "Mac mini ordered." It could give a sales boost to Apple, some X users have pointed out -- and online searches for "Mac Mini" jumped in the last 4 days in the US, per Google Trends. But Steinberger said buying a new computer just to run the AI isn't necessary. "Please don't buy a Mac Mini," he wrote. "You can deploy this on Amazon's Free Tier."

Read more of this story at Slashdot.

You can’t cheaply recompute without re-running the whole model – so KV cache starts piling up

Feature Large language model inference is often stateless, with each query handled independently and no carryover from previous interactions. A request arrives, the model generates a response, and the computational state gets discarded. In such AI systems, memory grows linearly with sequence length and can become a bottleneck for long contexts. …

But did it falter? Oracle debuts Schrödinger's cloud

Oracle Cloud Infrastructure (OCI) users reported an outage late last week in its London region, yet despite complaints from Register readers, Big Red is staying quiet.…

Amazon announced on Wednesday that it is eliminating approximately 16,000 roles across the company as part of organizational changes that began in October 2025 and are only now being finalized by certain teams. Senior Vice President Beth Galetti shared the news in a memo to employees, framing the reductions as an effort to reduce layers, increase ownership, and remove bureaucracy. The memo follows another memo that the company accidentally sent to employees.

Read more of this story at Slashdot.

Cascading backend failures kept users locked out until home time

Microsoft Azure, or at least the part of it that handles the OpenAI service in the Sweden Central region, was down and out for the count yesterday, leaving users facing errors for much of the working day.…

AWS and Capgemini loom large in HMRC's procurement pipeline

The UK's tax collector is budgeting to spend more than £2 billion on new tech deals in the next couple of years, including a contract set for AWS and another for Capgemini to be awarded without competition.…

An anonymous reader quotes a report from Universe Today: There's a bright side to every situation. In 2032, the Moon itself might have a particularly bright side if it is blasted by a 60-meter-wide asteroid. The chances of such an event are still relatively small (only around 4%), but non-negligible. And scientists are starting to prepare both for the bad (massive risks to satellites and huge meteors raining down on a large portion of the planet) and the good (a once in a lifetime chance to study the geology, seismology, and chemical makeup of our nearest neighbor). A new paper from Yifan He of Tsinghua University and co-authors, released in pre-print form on arXiv, looks at the bright side of all of the potential interesting science we can do if a collision does, indeed, happen. If Asteroid 2024 YR4 were to hit the Moon, researchers would be able to watch a large lunar impact unfold in real time and collect data on extreme collisions that usually exist only in computer models. Telescopes could follow how a newly formed crater and its pool of molten rock cool and solidify, while the resulting moonquake would offer a clearer picture of its internal structure via the seismic waves it sends through the Moon. Furthermore, researchers could compare the fresh crater to older ones to improve our understanding of the Moon's long history of impacts. Debris blasted off the surface could even deliver small lunar samples to Earth. Altogether, it would be a once-in-a-generation chance to learn more about how the Moon/rocky worlds respond to powerful impacts.

Read more of this story at Slashdot.

'Follow-on' agreement lasts 3 years as US techies protest vendor's ICE contract Stateside

The UK's Ministry of Defence (MoD) has directly awarded a £240.6 million contract to US technology company Palantir to continue to licence and support its data analytics work.…

Connection secured. Not so sure about the installation

Bork!Bork!Bork! Behold an ATM crying out for a man-in-the-middle attack. An obsolete Microsoft operating system cannot be blamed here. This is all about the hardware.…