Latest Posts
Insights, tutorials, and news from the Mazakiya-AI team.
Scaling long-running autonomous coding
We've been experimenting with running coding agents autonomously for weeks. Our goal is to understand how far we can push the frontier of agentic coding for projects that typically take human teams months to complete
There's a ridiculous amount of tech in a disposable vape
So, I'm walking through a park when I see this thing lying on the ground:It's a disposable vape that someone has discarded because it's empty
TimeCapsuleLLM: LLM trained only on data from 1800-1875
TimeCapsule LLM A language model trained from scratch exclusively on data from certain places and time periods to reduce modern bias and emulate the voice, vocabulary, and worldview of the era
CES: So very big, so little sustainability tech
Every third booth at CES showed off some new AI product or other. If you wanted to find a robotic lawn mower, throw a rock. Humanoid robots, smart locks and super thin TVs were everywhere. But if you went looking for sustainability products, you’re going to have to hunt a bit. Last year, the Sustainability section at the Las Vegas Convention Center had 20 booths. This year, there were 38, but that’s in part due to the combination of the energy and sustainability categories. So exhibitors like South Korea’s largest electric utility company, a nuclear power company from the same country and lots of battery manufacturers. There was also an AI data platform booth in the section that had nothing to do with sustainability as far as I can tell. Guess the organizers just ran out of room for all the AI. Within the sustainability section, and at other CES venues, I found a few encouraging displays of sustainable products — organizations and devices that were trying to address the multitude of problems the world is facing when it comes to energy production, climate and pollution. But none of it quite achieved Engadget’s best of CES status this year. Some of what we saw was utility-scale, some wasn’t quite ready for consumer consumption and other stuff was too niche or had too many caveats to make the list. I won’t go so far as to say sustainability is dead at CES, because that sends me into dark downward spirals, but it’s getting sparse out there, friends. Here are the companies I saw that had promise and innovative ideas. And gosh darn it, at least these guys are trying. Shine Turbine 2.0 Spinning the Shine 2.0 wind turbine Amy Skorheim for Engadget This little guy could be a precursor to some serious personal wind power generation. That’s where the company is heading. For now, the Shine 2.0 can use as little as a light breeze to start generating power to charge your smartphones, laptops or even a power station. The whole unit weighs three pounds and sets up in around two minutes. The second generation model can output up to 75 watts and the company is working on a third version that goes up to 100 watts for even more substantial energy production. Learn more at Shine. Flint battery tech Flint batteries break down by 70 percent in four weeks in a compost pile. Amy Skorheim for Engadget When I approached Flint’s booth, the rep told me the company made cellulose batteries. And I thought, like paper-wrapped batteries? Nope. The chemicals inside the batteries are made from cellulose. They have a solvent-free, lithium-free, PFAS-free chemistry and break down by 70 percent in four weeks in a composting environment. They use the same basic architecture as a lithium-ion cell, with an anode, cathode and separator with ion transfers between the two. As of now, Flint is focused on partnering with manufacturers, and consumer products are on the horizon. Learn more at Flint. Clear Drop soft plastics compactor The Clear Drop soft plastics compactor next to a pile of the bricks it produces. Amy Skorheim for Engadget The Clear Drop is a soft plastics compactor that creates eight by twelve by four-inch bricks out of hundreds of grocery bags, bubble wrap, ziplocks and plastic packaging. One brick is equivalent to a 30-pound trash bag-worth of bags. Once the brick is created, it can be shipped to one of Clear Drop’s partner facilities in a pre-paid USPS shipping envelope. They currently work with a few US-based recycling facilities and hope to one day create an infrastructure to include municipal recycling. Learn more at Clear Drop. Alpha Power by CPTI Alpha power by CPTI Amy Skorheim for Engadget From what I’ve learned at the show, perovskite is the hottest thing in solar right now. It’s a mineral compound that’s been used to create more efficient solar panels. Some so sensitive to light that just indoor illumination is enough to create usable energy. Alpha Power by CPTI creates lightweight, flexible perovskite solar panels that can conform to multiple surfaces. Again, this is a company that’s partnering with manufacturers, so look for panels built into your laptop to charge it under the glare of your too-harsh office lights. Learn more at CPTI. Green Vigor 3D models of buildings using Green Vigor technology. Amy Skorheim for Engadget Down in the lower levels of the Venetian Expo at CES I found Green Vigor in the Hong Kong pavilion. This small company has two solutions to create energy for buildings by harnessing the potential energy from existing systems. HydroVigor generates power from water systems. So every time someone washes their hands or flushes a toilet in a building, the roof-top system generates a bit of power. CoolVigor uses the same principles to harness energy from HVAC systems. HydroVigor is currently in use in many buildings in Singapore and Hong Kong and they’re working to expand to more buildings globally. Learn more at GreenVigor. Jackery Solar Gazebo Jackery's Solar Gazebo. Amy Skorheim for Engadget This outdoor hangout spot can produce up to 10kWh of power on a given day. It’s a modular design that lets you choose louvered walls, sunshades, lights and fans when you order it and the solar panels are so strong that a full-sized human Jackery rep was able to stand on a sample panel in front of me and nothing cracked (though the company officially rates it at 20 pounds of snow per square foot). You can use the power directly, tie it into your home system, feed it into the grid or hook it up to one of Jackery’s many power stations to save the power for later. The gazebo costs $12,000 and will ship in mid-2026. Learn more at Jackery. Bluetti RV Solar System Bluetti's DIY RV Solar power system Amy Skorheim for Engadget Bluetti, like Jackery, is known for its vast lineup of portable and fixed power stations and batteries. This year, it brought a new power station made with bio-based plastic as well as a DIY system for adding solar power to your existing RV. Learn more at Bluetti. Airloom wind power generation Airloom's roller coaster-like wind power generator for data centers. Amy Skorheim for Engadget Engadget’s Anna Washenko does a great job of explaining the tech behind Airloom. In short it’s a roller coaster for wind that’s comprised of 40 percent less mass than a standard wind turbine and uses 42 percent fewer parts and 96 percent fewer unique parts. That makes it faster to deploy and cheaper to instal. I can also be sited in more places. Again, this is a utility-scale solution, geared towards data centers and their insatiable need for energy to power Very Important AI Things. Learn more at Airloom. Gaotu Innovation Energy Group Gaotu had a range of solar products in various formats. Amy Skorheim for Engadget If you are looking for a solar-powered anything, hit up Gaotu. At the company’s booth, I saw hats, a fishing chair, a backpack, a sunbrella and a car roof-top enclosure that unfurls to charge up your Tesla. The Shenzhen-based company has been in business for 18 years and plans to just keep sticking solar panels on anything it can. Learn more at Gaotu. Segway Muxi cargo e-bike Segway's latest cargo e-bike Amy Skorheim for Engadget The single largest booth in the CES sustainability section was Segway. This year, the company showed off two new e-bikes, which our own Dan Cooper covered. This one here is the Muxi, a cargo bike with an easily swappable battery, an optional passenger seat with foot pegs and an optional middle basket. Plus a beverage cup holder. Learn more at Segway. If we don’t all fall into the ocean before then, perhaps CES 2027 will have a stronger showing of sustainability tech. In the meantime, I’ll take a modicum of comfort in these few brave organizations still dedicated to keeping us afloat.
Orchestral replaces LangChain’s complexity with reproducible, provider-agnostic LLM orchestration
A new framework from researchers Alexander and Jacob Roman rejects the complexity of current AI tools, offering a synchronous, type-safe alternative designed for reproducibility and cost-conscious science. In the rush to build autonomous AI agents, developers have largely been forced into a binary choice: surrender control to massive, complex ecosystems like LangChain, or lock themselves into single-vendor SDKs from providers like Anthropic or OpenAI. For software engineers, this is an annoyance. For scientists trying to use AI for reproducible research, it is a dealbreaker. Enter Orchestral AI, a new Python framework released on Github this week that attempts to chart a third path. Developed by theoretical physicist Alexander Roman and software engineer Jacob Roman, Orchestral positions itself as the "scientific computing" answer to agent orchestration—prioritizing deterministic execution and debugging clarity over the "magic" of async-heavy alternatives. The 'anti-framework' architecture The core philosophy behind Orchestral is an intentional rejection of the complexity that plagues the current market. While frameworks like AutoGPT and LangChain rely heavily on asynchronous event loops—which can make error tracing a nightmare—Orchestral utilizes a strictly synchronous execution model. "Reproducibility demands understanding exactly what code executes and when," the founders argue in their technical paper. By forcing operations to happen in a predictable, linear order, the framework ensures that an agent’s behavior is deterministic—a critical requirement for scientific experiments where a "hallucinated" variable or a race condition could invalidate a study. Despite this focus on simplicity, the framework is provider-agnostic. It ships with a unified interface that works across OpenAI, Anthropic, Google Gemini, Mistral, and local models via Ollama. This allows researchers to write an agent once and swap the underlying "brain" with a single line of code—crucial for comparing model performance or managing grant money by switching to cheaper models for draft runs. LLM-UX: designing for the model, not the end user Orchestral introduces a concept the founders call "LLM-UX"—user experience designed from the perspective of the model itself. The framework simplifies tool creation by automatically generating JSON schemas from standard Python type hints. Instead of writing verbose descriptions in a separate format, developers can simply annotate their Python functions. Orchestral handles the translation, ensuring that the data types passed between the LLM and the code remain safe and consistent. This philosophy extends to the built-in tooling. The framework includes a persistent terminal tool that maintains its state (like working directories and environment variables) between calls. This mimics how human researchers interact with command lines, reducing the cognitive load on the model and preventing the common failure mode where an agent "forgets" it changed directories three steps ago. Built for the lab (and the budget) Orchestral’s origins in high-energy physics and exoplanet research are evident in its feature set. The framework includes native support for LaTeX export, allowing researchers to drop formatted logs of agent reasoning directly into academic papers. It also tackles the practical reality of running LLMs: cost. The framework includes an automated cost-tracking module that aggregates token usage across different providers, allowing labs to monitor burn rates in real-time. Perhaps most importantly for safety-conscious fields, Orchestral implements "read-before-edit" guardrails. If an agent attempts to overwrite a file it hasn't read in the current session, the system blocks the action and prompts the model to read the file first. This prevents the "blind overwrite" errors that terrify anyone using autonomous coding agents. The licensing caveat While Orchestral is easy to install via pip install orchestral-ai, potential users should look closely at the license. Unlike the MIT or Apache licenses common in the Python ecosystem, Orchestral is released under a Proprietary license. The documentation explicitly states that "unauthorized copying, distribution, modification, or use... is strictly prohibited without prior written permission". This "source-available" model allows researchers to view and use the code, but restricts them from forking it or building commercial competitors without an agreement. This suggests a business model focused on enterprise licensing or dual-licensing strategies down the road. Furthermore, early adopters will need to be on the bleeding edge of Python environments: the framework requires Python 3.13 or higher, explicitly dropping support for the widely used Python 3.12 due to compatibility issues. Why it matters "Civilization advances by extending the number of important operations which we can perform without thinking about them," the founders write, quoting mathematician Alfred North Whitehead. Orchestral attempts to operationalize this for the AI era. By abstracting away the "plumbing" of API connections and schema validation, it aims to let scientists focus on the logic of their agents rather than the quirks of the infrastructure. Whether the academic and developer communities will embrace a proprietary tool in an ecosystem dominated by open source remains to be seen, but for those drowning in async tracebacks and broken tool calls, Orchestral offers a tempting promise of sanity.
‘Physical AI’ Is Coming for Your Car
What the latest tech-marketing buzzword has to say about the future of automotive.
Google announces AI Overviews in Gmail search, experimental AI-organized inbox
Last year's premium Gmail AI features are also rolling out to free users.
Ford is getting ready to put AI assistants in its cars
The Blue Oval is also working on new hands-free driver assists.
Secrets of the sleep-deprived brain
Nearly everyone has experienced it—after a night of poor sleep, your brain might seem foggy, and your mind drifts off when you should be paying attention. A new MIT study reveals what happens biologically as these momentary lapses occur: Your brain is performing essential maintenance that it usually takes care of while you sleep. During
Microsoft’s Nadella wants us to stop thinking of AI as ‘slop’
Nadella wants us to think of AI as a human helper instead of a slop-generating job killer. New data for 2026 indicates he could be right.
SwitchBot turned up to CES with an AI wearable that records everything you say
We’ve all fantasied about sprawling on an embroidered chaise lounge while our amanuensis faithfully typed out everything we said. SwitchBot turned up to CES with something that does a similar job, but without the romance of the amanuensis, the typewriter or the chaise lounge. Instead, it’s launching the AI MindClip, a wearable microphone which records and transcribes every noise to come from your mouth. The company isn’t alone, with many others turning up to CES this year offering devices to log your thoughts through the day. But SwitchBot does hope to distinguish itself from the crowd with an AI which can both summarize and extract useful data from your chats. The company’s representatives said it will use its initiative and create its own reminders based on what it thinks you’ll need to remember. Sadly, we couldn’t get to see it working in person, and we don’t know when it’ll be available or how much it’ll cost. But it’ll probably be a little bit cheaper than a chaise lounge.
AI Labor Is Boring. AI Lust Is Big Business
After years of hype about generative AI increasing productivity and making lives easier, 2025 was the year erotic chatbots defined AI’s narrative.
9to5Mac Daily: January 2, 2026 – Apple Intelligence and Apple TV rumors
Listen to a recap of the top stories of the day from 9to5Mac. 9to5Mac Daily is available on iTunes and Apple’s Podcasts app, Stitcher, TuneIn, Google Play, or through our dedicated RSS feed for Overcast and other podcast players. Sponsored by Roborock: Save on Roborock’s flagship vacuums — but hurry, these deals won’t last long!
Happy Public Domain Day 2026
Scroll through the whole page to download all images before printing. The calendar turns, and once again a lively procession of books, images, films, and music leaves copyright behind and steps into the ever-growing public domain
Scientists unlock brain's natural clean-up system for new treatments for stroke
Imagine a sudden, lifealtering event where a part of your brain is starved of oxygen, causing vital cells to die. This is what happens during a stroke, leaving millions worldwide with devastating disabilities
NASA finally has a leader, but its future is no more certain
After a rudderless year and an exodus of around 4,000 employees due to Trump administration cuts, NASA got what may be its first piece of good news recently. On December 17, the Senate confirmed billionaire Jared Isaacman as the agency's new administrator. He now holds the power to rehabilitate a battered engine of scientific research, or steer it towards even more disruption. Considering the caliber of President Trump's other appointees, Isaacman is probably the best candidate for the job. Outside of being a successful entrepreneur, he has flown fighter jets and been to space twice as part of the Inspiration4 and Polaris Dawn private missions. One of those flights saw him complete the first commercial space walk, and travel farther from Earth than any human since the end of the Apollo program. "Perfect is the enemy of the good. Isaacman checks a lot of boxes," says Keith Cowing, a former NASA employee and the founder of NASA Watch, a blog dedicated to the agency. "He's passed every requirement to fly in a spacecraft that American astronauts at NASA are required to pass. He also went out of his way to have a diverse crew, and shove as much science as he could in those missions." And yet if you're a NASA employee or just someone who cares about the agency's work, there are still plenty of reasons to be concerned for its future. When Trump first nominated Isaacman in the spring, the billionaire wrote a 62-page document detailing his vision for NASA. In November, Politico obtained a copy of that plan, titled Project Athena. To some insiders, Project Athena painted a picture of someone who, at least at the time when it was written, fundamentally misunderstood how NASA works and how scientific discovery is funded in the US and elsewhere. It also suggests Isaacman may be more open to Trump's NASA agenda than would appear at first glance. When asked about the plan by Politico, one former NASA official characterized it as "bizarre and careless." Another called it “presumptuous," given many of the proposed changes to the agency's structure would require Congressional approval. In one section, Isaacman recommended taking “NASA out of the taxpayer funded climate science business and [leaving] it for academia to determine.” In another section, he promised to evaluate the “relevance and ongoing necessity” of every agency center, particularly NASA's iconic Jet Propulsion Laboratory, saying the facility and others must increase the “output and time to science KPI.” A lot has changed since Isaacman first wrote that document. It came before the workforce cuts, before the future of Goddard Space Flight Center became uncertain and before Trump surprised everyone by renominating Isaacman. But during his Senate testimony earlier this month, the billionaire said “I do stand behind everything in the document, even though it was written seven months ago. I think it was all directionally correct.” He did appear to distance himself from some viewpoints expressed in or inferred by Project Athena, however. Isaacman stated that “anything suggesting that I am anti-science or want to outsource that responsibility is simply untrue.” He also came out against the administration's plan to cut NASA's science budget nearly in half, claiming the proposals would not lead to "an optimal outcome." One thing is clear, Isaacman is not your typical bureaucrat. "One of the pitfalls of some prior NASA administrators has been that they've shown too much reverence for the internal processes and bureaucratic structure of the agency to the detriment of decision-making and performance," said Casey Drier, chief of space policy at The Planetary Society, a nonprofit that advocates for the exploration and study of space. "Isaacman has positioned himself as the opposite of that. Clearly, that's something that could lead to a lot of political and congressional challenges if taken too far." Even if Isaacman doesn't follow through on any of the proposals made in Project Athena, there's only so much a NASA administrator — even one sympathetic to civil servants working under them — can do. "Once a budget request goes out publicly, everyone in the administration has to defend it. Anything he does will have to be internal and private," Drier explains. "He never explicitly criticized the administration during his hearing. He's also coming relatively late in the budget process." A lot of NASA's future will depend on the White House Office of Management and Budget (OMB), which is responsible for implementing the president's agenda across the executive branch. As a direct result of guidance the OMB issued over the summer, NASA awarded 25 percent fewer new grants in 2025 than it did on average between 2020 and 2024. "The OMB has added layers of requirements that scientists now have to go through to spend the money they've already been allocated. The administration has worked against its own stated goals of efficiency," Drier said. "Isaacman can't solve that himself. He can't tell the OMB what to do. That's going to be a serious challenge." Looming over everything is the fact NASA still does not have a full-year budget for 2026. Congress has until January 30 to fund NASA and the rest of the federal government before the short-term funding bill it passed on November 12 runs out. "On paper, the official policy of the administration is still to terminate a third of NASA's scientific capability," Drier points out. There are reasons to be cautiously optimistic. Publicly, both the House and Senate have come out against Trump's funding cuts. And some science missions that were slated to be cancelled, such as OSIRIS-APEX, have been approved for another full year of operations. What NASA needs now is someone who will, as Drier puts it, "vigorously advocate" for the agency in whatever way they can. It remains to be seen if that's Jared Isaacman.
Incremental Backups of Gmail Takeouts
Home December 2025 In an earlier writeup I discussed how to create reproducible bundles of git repositories such that a file-based backup strategy can operate incrementally
The Google Pixel Watch 4 made me like smartwatches again
Google's Pixel Watch 4 is an excellent smartwatch with fast charging and great design.
Replacing JavaScript with Just HTML
Grogg published on Dec 27, 2025For many years now, JavaScript has been the workhorse of the web. If you wanted to do something that couldn't be done with just HTML and CSS, you could usually find a way to do it with JS
Google’s and OpenAI’s Chatbots Can Strip Women in Photos Down to Bikinis
Users of AI image generators are offering each other instructions on how to use the tech to alter pictures of women into realistic, revealing deepfakes.
MiniMax M2.1: Built for Real-World Complex Tasks, Multi-Language Programming
MiniMax M2.1 is engineered for peak performance on real-world complex tasks. Learn how its multi-language programming capabilities streamline development, transforming your workflow. Master intricate projects with unparalleled efficiency.
OpenAI’s Child Exploitation Reports Increased Sharply This Year
The company made 80 times as many reports to the National Center for Missing & Exploited Children during the first six months of 2025 as it did in the same period a year prior.
AI Can Write Your Code. It Can't Do Your Job
In May, OpenAI agreed to pay $3 billion for Windsurf, the AI coding assistant formerly known as Codeium. Three billion dollars. For a VSCode fork. The deal eventually fell apart, but what matters is that they wanted to do it in the first place
GLM-4.7: Advancing the Coding Capability
2025-12-22 · ResearchGLM-4. 7, your new coding partner, is coming with the following features:Core Coding: GLM-4. 7 brings clear gains, compared to its predecessor GLM-4
How AIM taught the internet to chat
If you were an internet user around the turn of the century, there's a good chance I could play a one-second long sound of a door opening and memories would immediately come flooding back. Memories of running home from school and logging onto AOL Instant Messenger to chat with your friends or your crush. Maybe memories of how AIM changed the way your company did business. Certainly memories of your old screen name, and the angsty song lyrics you put into your away message. AIM was, for a time, the most important chat app on the internet. It also barely managed to continue to exist. The app was created by a semi-rogue team inside of AOL, and
Is that video AI? Gemini can now help you figure it out - but there's a catch
Here's how to use Google Gemini to detect certain AI-generated videos.
Carolina Cloud – One third the cost of AWS for data science workloads
Are you a data scientist or leading a data science team constantly battling with escalating cloud costs. It's a common scenario
Pumped Hydro Energy Storage Is Having a Renaissance
As the world looks to incorporate more renewables into energy grids, centuries-old systems that can balance supply and demand are being reappraised and innovated upon.
Explaining the Widening Divides in US Midlife Mortality: Is There a Smoking Gun?
Working Paper 34553 DOI 10. 3386/w34553 Issue Date December 2025 The education-mortality gradient has increased sharply in the last three decades, with the life-expectancy gap between people with and without a college degree widening from 2
AI will make formal verification go mainstream
Published by Martin Kleppmann on 08 Dec 2025
I hate that I love Riverside’s AI-driven ‘Rewind’ for podcasters
Riverside's year-end review is amusing, but it represents a creative culture that's increasingly entwined with AI.
Hashcards: A plain-text spaced repetition system
hashcards is a local-first spaced repetition app, along the lines of Anki or Mochi. Like Anki, it uses FSRS, the most advanced scheduling algorithm yet, to schedule reviews
The best iPad Air cases of 2025: Expert tested
If you gift an iPad this holiday season, don't forget a case. Discover our favorite iPad Air cases from brands including Apple, Burga, and Logitech.
OpenAI are quietly adopting skills, now available in ChatGPT and Codex CLI
12th December 2025 One of the things that most excited me about Anthropic’s new Skills mechanism back in October is how easy it looked for other platforms to implement
Programmers and software developers lost the plot on naming their tools
This section was labeled under, or is related to Programming In Dec 2022 I watched Richard Stallman’s talk on the EmacsConf, it was titled “What I’d like to see in Emacs”
2 Men Linked to China’s Salt Typhoon Hacker Group Likely Trained in a Cisco ‘Academy’
The names of two partial owners of firms linked to the Salt Typhoon hacker group also appeared in records for a Cisco training program—years before the group targeted Cisco’s devices in a spy campaign.
A massive, Chinese-backed port could push the Amazon Rainforest over the edge
The port will revolutionize global trade, but it’s sparking destructive rainforest routes.
Horses: AI progress is steady. Human equivalence is sudden
So after all these hours talking about AI, in these last five minutes I am going to talk about: horses. Engines, steam engines, were invented in 1700. And what followed was 200 years of steady improvement, with engines getting 20% better a decade
The era of AI persuasion in elections is about to begin
In January 2024, the phone rang in homes all around New Hampshire. On the other end was Joe Biden’s voice, urging Democrats to “save your vote” by skipping the primary. It sounded authentic, but it wasn’t. The call was a fake, generated by artificial intelligence. Today, the technology behind that hoax looks quaint. Tools like
Knight of the Seven Kingdoms trailer brings levity to Westeros
"Every knight needs a squire, and you look like you need one more than most."
I cracked a $200 software protection with xcopy
disclaimer: this is educational security research only. i do not condone piracy. i purchased a legitimate license for this software and conducted this analysis on my own property
In comedy of errors, men accused of wiping gov databases turned to an AI tool
Defendants were convicted of similar crimes a decade ago. How were they cleared again?
iPadOS 26.2 RC and more now available ahead of coming launch
Apple has just debuted the RC (release candidate) versions of its various 26.2 software updates, including iPadOS 26.2 and more. Here’s what to expect.
OpenAI desperate to avoid explaining why it deleted pirated book datasets
OpenAI risks increased fines after deleting pirated books datasets.
What Will Enter the Public Domain in 2026?
At the start of each year, on January 1st, a new crop of works enter the public domain and become free to enjoy, share, and reuse for any purpose
Best Buy's Cyber Monday sale is here: Check out my favorite 30+ tech discounts live now
Cyber Monday is tomorrow, but Best Buy has some excellent deals live now, like the Hisense Canvas TV for 35% off.
Google drops EU antitrust complaint against Microsoft amid cloud probe
Google LLC has dropped an antitrust complaint that it filed in the European Union last year over Microsoft Corp.’s business practices. The search giant announced the move today. The development comes a few days after the European Commission, the EU’s executive arm, launched an antitrust probe into Microsoft. The investigation will scrutinize the company’s conduct […] The post Google drops EU antitrust complaint against Microsoft amid cloud probe appeared first on SiliconANGLE.
Airbus A320 – intense solar radiation may corrupt data critical for flight
Toulouse, France, 28 November 2025 – Analysis of a recent event involving an A320 Family aircraft has revealed that intense solar radiation may corrupt data critical to the functioning of flight controls
Black Friday savings hit official iPhone 17 cases, Apple Watch bands, AirTag, and
Let’s say you already have the latest and greatest Apple gear, maybe like yours truly. Well, the good news is we still have plenty of Black Friday deals on ways to accessorize your gear. We’re tracking huge savings on Apple’s entire lineup of official accessories for iPhone, Apple Watch, and more. Plus, deals on the best accessories from our favorite accessory makers like Nomad, Twelve South, and Satechi.
Migrating the Main Zig Repository from GitHub to Codeberg
← Back to News page November 26, 2025 org/ziglang/zigEver since git init ten years ago, Zig has been hosted on GitHub. Unfortunately, when it sold out to Microsoft, the clock started ticking
“Go generate a bridge and jump off it”: How video pros are navigating AI
I talked with nine creators about economic pressures and fan backlash.
What Kilian Jornet Packs to Climb Every 14er in the Mountain West in 31 days
The world’s greatest skyrunner shares the shoes, fitness tracker, and practices that let him climb every mountain in the western United States in just 31 days.
Fran Sans – font inspired by San Francisco light rail displays
Written by EMILY SNEDDON Published on 6TH NOVEMBER 2025 Fran Sans is a display font in every sense of the term. It’s an interpretation of the destination displays found on some of the light rail vehicles that service the city of San Francisco
AirPods 4 vs AirPods Pro: Which should you buy this holiday season?
Apple’s AirPods lineup is stronger than it’s ever been, with compelling options at multiple different price points. If you’re exploring which AirPods 4 vs AirPods Pro to buy this holiday season, whether for yourself or someone else, here’s what you need to know.
Arduino Terms of Service and Privacy Policy update: setting the record straight
The Arduino Terms of Service and Privacy Policy update: setting the record straight Arduino Team — November 21st, 2025 We’ve heard some questions and concerns following our recent Terms of Service and Privacy Policy updates
New OS aims to provide (some) compatibility with macOS
Meet ravynOS, a new open-source OS bridging the gap to macOS. Built on FreeBSD for x86-64, it aims for macOS source compatibility (compile & run Mac apps!), familiar GUI, folder layouts like /System, plus ZFS & Linux binary support. Get ready for an OS that feels like home.
The Google Search of AI agents? Fetch launches ASI:One and Business tier for new era of non-human web
Fetch AI, a startup founded and led by former DeepMind founding investor, Humayun Sheikh, on Wednesday announced the release of three interconnected products designed to provide the trust, coordination, and interoperability needed for large-scale AI agent ecosystems. The launch includes ASI:One, a personal-AI orchestration platform; Fetch Business, a verification and discovery portal for brand agents; and Agentverse, an open directory hosting more than two million agents. Together, the system positions Fetch as an infrastructure provider for what it calls the “Agentic Web”—a layer where consumer AIs and brand AIs collaborate to complete tasks instead of merely suggesting them. The company says the tools address a central limitation in current consumer AI: models can provide recommendations but cannot reliably execute multi-step actions that require coordination across businesses. Fetch’s approach centers on enabling agents from different organizations to interoperate securely, using verified identities and shared context to complete end-to-end workflows. “We’re creating the same foundation for agents that Google created for websites,” said Humayun Sheikh, Founder and CEO of Fetch AI, and an early investor in DeepMind, in a press release provided to VentureBeat. “Instead of just finding information, your personal AI coordinates with verified brand agents to get things done.” Fetch’s founding and DeepMind connection Fetch AI was founded in 2017 by Humayun Sheikh, an entrepreneur whose early investment in DeepMind helped support the company’s commercial development before its acquisition by Google. “I was one of the first five people at DeepMind and its first investor. My check was the first one in,” Sheikh said, reflecting on the period when advanced machine learning research was still largely inaccessible outside major technology companies. His early experience helped shape Fetch’s direction. “Even in 2013, it was clear to me that agentic systems were going to be the ones that worked. That’s where I focused—on the agentic web,” Sheikh noted. Fetch built on this thesis by developing infrastructure for autonomous software agents, focusing on verifiable identity, secure data exchange, and multi-agent coordination. Over the past several years, the company has expanded to a 70-person team across Cambridge and Menlo Park, raised approximately $60 million, and accumulated more than one million users interacting with its model—data that informed the design of the newly launched products. Sheikh added that his decision to bootstrap the company initially came directly from the proceeds of the DeepMind exit, noting in the interview that while the sale to Google was “a good exit,” he believed the team could have held out for a higher valuation. The early self-funding period allowed Fetch to begin work in 2015—well before transformer architectures went mainstream—on the hypothesis that agentic infrastructure would become foundational to applied AI. ASI:One is a platform for multi-agent orchestration At the core of the launch is ASI:One, a language model interface designed specifically for coordinating multiple agents rather than addressing isolated queries. Fetch describes it as an “intelligence layer” that handles context sharing, task routing, and preference modeling. The system stores user-level signals such as favored airlines, dietary constraints, budget ranges, loyalty program identifiers, and calendar availability. When a user requests a complex task — such as planning a trip with flights, hotels, and restaurant reservations — ASI:One retrieves those preferences and delegates work to the appropriate verified agents. The agents then return actionable outputs, including inventory and booking options, rather than generic recommendations. In practice, ASI:One functions as a workflow generator across organizational boundaries. By contrast with conventional LLM applications, which often rely on APIs or RAG techniques to surface information, ASI:One is built to coordinate autonomous agents that can complete transactions. Fetch notes that personalization improves over time as the model accumulates structured preference data. Sheikh emphasized the distinction between orchestrated execution and traditional AI output. “This isn’t searching for options separately and hoping they work together,” he said. “It’s orchestration.” He added that Fetch’s architecture is intentionally modular: “Our architecture is a mix of agentic and expert models. One large model isn’t enough — you need specialists. That’s why we built ASI1, tuned specifically for agentic systems.” The interview also revealed new details about ASI:One’s personalization systems: the platform uses multiple user-owned knowledge graphs to store preferences, travel history, social connections, and contextual constraints. These knowledge graphs are siloed per user and not co-mingled with any Fetch-operated data. Sheikh described this as a “deterministic backbone” that gives the personal AI a stable memory layer beyond the probabilistic output of a single large model. ASI:One launches in Beta today, with a broader release planned for early 2026. Fetch also offers ASI:One Mobile, released earlier this year, giving users access to the same agent-orchestration capabilities on iOS and Android. The mobile app connects directly to Agentverse and the user’s knowledge graphs, enabling on-the-go task execution and real-time interaction with registered agents. Fetch Business offers verified identity and brand control To enable reliable coordination between consumers and companies, Fetch is introducing a verification and discovery portal called Fetch Business. The platform allows organizations to verify their identity and claim an official Brand Agent handle — for example, @Hilton or @Nike — regardless of which tools they use to build the underlying agent. Fetch positions the product as an analogue to ICANN domain registration and SSL certificate systems for websites. Verified status is intended to protect consumers from interacting with counterfeit or untrusted agents, a problem the company describes as a major barrier to widespread agent adoption. The system includes low-code tools for small businesses to create agents in a few steps and connect real-time APIs such as inventory, booking systems, or CRM platforms. “With Fetch, you can create an agent in one minute. It gets a handle, like a Twitter username, and you can personalize it completely—even give it your social media permissions to post on your behalf,” Sheikh said. Once a brand claims its namespace, its agent becomes discoverable to consumer AIs and other agents inside Agentverse. The company has pre-reserved thousands of brand namespaces in anticipation of demand. Verification status persists across any platform that integrates with Agentverse, creating a portable identity layer for business agents. The interview highlighted that Fetch Business inherits web-trust primitives directly: domain owners verify their identity by inserting a short code snippet into their existing website backend, allowing the system to pass a cryptographic challenge and grant the agent an authenticity badge similar to a “blue check” for agent identities. Sheikh framed this as “reusing the trust layer the web already spent decades building.” Companies can begin claiming agents now at business.fetch.ai. Agentverse is an open directory of more yhan 2 million agents The final component of the release is Agentverse, an open directory and cloud platform that hosts agents and enables cross-ecosystem discoverability. Fetch states that millions of agents have already registered, spanning travel, retail, entertainment, food service, and enterprise categories. Agentverse provides metadata, capability descriptions, and routing logic that ASI:One uses to identify appropriate agents for specific tasks. It also supports secure communication and data exchange between agents. The company notes that the directory is platform-agnostic: agents built with any framework can join and interoperate. According to Sheikh, the lack of a discovery layer is one reason most AI agents see little or no usage. “Ninety percent of AI agents never get used because there’s no discovery layer,” he said. He framed the role of Agentverse in more technical terms: “Right now, if you build an agent, there’s no universal way for others to discover it. That’s what AgentVerse solves—it’s like DNS for agents.” He also described the system as an essential component of the emerging agent economy: “Fetch is building the Google of agents. Just like websites needed search, agents need discovery, trust, and interaction—Fetch provides all of that.” The interview further underscored that Agentverse is cloud-agnostic by design. Sheikh contrasted this with competing agent ecosystems tied to specific cloud providers, arguing that a universal registry is only viable if independent of proprietary cloud environments. He said the open architecture enables an LLM to query any agent “within one minute of deployment,” turning agent publication into a near-instantaneous process similar to registering a domain. Agentverse also integrates payment pathways, enabling agents to execute purchases using partners such as Visa, Skyfire, and supported stablecoins. Consumers can configure spending limits or require explicit approval for transactions. Industry context and implications Fetch’s launch comes at a time when consumer AI platforms are exploring the shift from static chat interfaces toward autonomous agents capable of completing actions. However, most agent systems remain limited by siloed architectures, limited interoperability, and weak verification standards. Fetch positions its infrastructure as a response to these limitations by providing a cross-platform coordination layer, identity system, and directory service. The company argues that an agent ecosystem requires consistent verification mechanisms to ensure that consumers interact with authentic brand representatives rather than imitations. By establishing namespace control and portable trust indicators, Fetch Business aims to fill a gap similar to early web domain verification. At the same time, ASI:One attempts to centralize user preference data in a way that enables more efficient personalization and multi-agent coordination. This approach differs from generalist LLM applications, which often lack persistent preference architectures or direct access to brand-controlled agents. The interview also made clear that micropayments and digital transaction infrastructure are central to Fetch’s long-term vision. Sheikh referenced integrations with protocols such as Coinbase’s 402 and AP2, positioning these capabilities as essential for autonomous agents to complete end-to-end tasks that include financial execution. Fetch’s combined release of ASI:One, Fetch Business, and Agentverse introduces an interconnected stack designed to support large-scale deployment and usage of AI agents. The company frames the system as foundational infrastructure for an agentic ecosystem, where consumer AIs can coordinate with verified brand agents to complete tasks reliably and securely. The additions to its identity, discovery, and orchestration layers reflect Fetch’s long-standing thesis — rooted partly in lessons from DeepMind’s early development — that intelligence becomes meaningful only when paired with the capacity to act.
AI data center provider Lambda raises whopping $1.5B after multibillion-dollar Microsoft deal
Lambda's new raise was a solid beat over the amount that deal-watchers say it was hoping to hit.
Windows 11 adds AI agent that runs in background with access to personal folders
Microsoft is moving forward with its plans to turn Windows 11 into a full-fledged “AI” operating system amidst Copilot backlash
How Your Brain Creates 'Aha' Moments and Why They Stick
A sudden flash of insight is a product of your brain. Neuroscientists track the neural activity underlying an “aha” and how it might boost memory. Introduction Here are three words: pine, crab, sauce
Baidu unveils proprietary ERNIE 5 beating GPT-5 performance on charts, document understanding and
Mere hours after OpenAI updated its flagship foundation model GPT-5 to GPT-5.1, promising reduced token usage overall and a more pleasant personality with more preset options,...
AI World Clocks
Every minute, a new clock is displayed that has been generated by nine different AI models. Each model is allowed 2000 tokens to generate its clock. Here is its prompt: Create...
Civil war is brewing in the wasteland in Fallout S2 trailer
Ghoulish Elvis impersonators are the least of Lucy's problems as the wasteland prepares for war.
OpenAI slams court order that lets NYT read 20 million complete user chats
OpenAI: NYT wants evidence of ChatGPT users trying to get around news paywall.
We ran over 600 image generations to compare AI image models
tl:dr; We’ve been making photo apps for iOS for long enough that we have gray hairs now, and using our experience we ran over 600 image generations to compare which AI models...
The next iPhone Air has reportedly been delayed
Apple is no longer planning to launch a second-generation iPhone Air next fall and has “already sharply scaled back production of the first version,” The Information reports....
The 10 best tech gifts you can buy for under $100 - no matter who is on your list
You don't have to break the bank to gift tech. Check out these fun and affordable gadgets for $100 or less, plus learn how you can save on them now.
Tabloid: The Clickbait Headline Programming Language
Welcome to a revolutionary era in software development, an era where your code isn't just functional, it's sensational
Immutable Software Deploys Using ZFS Jails on FreeBSD
November 6, 2025 FreeBSD’s native support for ZFS snapshots and jails provides a powerful foundation for immutable deployments
Scientists find ways to boost memory in aging brains
Memory loss may not simply be a symptom of getting older. New research from Virginia Tech shows that it’s tied to specific molecular changes in the brain and that adjusting...
Here’s how the new Motorola Edge 70 compares to the iPhone Air
The iPhone Air managed to beat the Samsung Galaxy S25 Edge to become the slimmest smartphone on the market, though we noted at the time that it’s not the slimmest smartphone...
Apple’s Persona technology uses Gaussian splatting to create 3D facial scans
Buried inside Apple's $3,499 Vision Pro VR headset is a feature that continually wows me, but you've probably never heard of it
AI's Dial-Up Era
Your computer modem screeches as it tries to connect to something called the internet. Maybe it works. Maybe you try again. For the first time in history, you can exchange...
Using FreeBSD to make self-hosting fun again
2025-11-01 - Feeling like a kid in a candy store, once more As evident by my last blog post "A prison of my own making", I needed to change something about my relationship with...
9to5Mac Daily: October 31, 2025 – Apple’s Q4 2025 earnings
Listen to a recap of the top stories of the day from 9to5Mac. 9to5Mac Daily is available on iTunes and Apple’s Podcasts app, Stitcher, TuneIn, Google Play, or through our...
Pat Gelsinger-backed Gloo seeking to raise up to $109M in IPO
Gloo Holdings Inc., a provider of software for faith-based organizations that is backed by Pat Gelsinger, plans to raise up to $109.2 million in an initial public offering....
Rouille – Rust Programming, in French
rouille Aren't you le tired from writing Rust programs in English. Do you like saying "merde" a lot. Would you like to try something different, in an exotic and funny-sounding...
Mark Zuckerberg is excited to add more AI content to all your social feeds
Meta CEO Mark Zuckerberg is getting ready to dump even more AI-generated posts into your social feeds. During an earnings call on Wednesday, Zuckerberg said the company will...
Generative AI Image Editing Showdown
The world of image editing is undergoing a fascinating revolution, and at its heart lies the incredible power of generative AI
Finding my rhythm again
In April of 2025, my heart threw an unexpected exception. I was 46, healthy, and active. I ran regularly. I ate reasonably well. I stayed busy. Too busy, maybe
NORAD’s Cheyenne Mountain Combat Center, c.1966
“These display screens would display signs of air attack against Canada and the United States
I'm drowning in AI features I never asked for and I hate it
At first, all of this AI stuff felt exciting. I was curious to try everything (I was actually one of the few naive people who thought the Rabbit R1 was a good product before it...
Unlocking Free WiFi on British Airways
2025-06-01 I was recently flying between HKG & LHR via British Airways. I’d done the same flight back in 2023, and remember relying on the in-flight entertainment for the 14...
AI discovers a 5x faster MoE load balancing algorithm than human experts
We're standing at a fascinating crossroads in the evolution of artificial intelligence. For years, AI has been an incredible tool for solving complex problems, from language...
Willow quantum chip demonstrates verifiable quantum advantage on hardware
Oct 22, 2025 Our Willow quantum chip demonstrates the first-ever algorithm to achieve verifiable quantum advantage on hardware
Google Fi will start using AI to make calls sound better
Google just announced several updates for its digital telecom provider, Google Fi. It's introducing AI-enhanced audio for better sound quality during calls. The company says...
Facebook's latest AI feature can scan your phone's camera roll
A Facebook feature that scans your phone's photo library to make AI collages and edits is now available in North America. Meta tested it earlier this year. It's an opt-in...
Opera's Neon shows just how confusing AI browsers still are
The trick to understanding Opera's Neon browser is recognizing that it's not just a browser with an AI bot added to it, but a browser with three AI bots all living side by side....
Adding Breadcrumbs to a Rails Application
Helping users navigate through our site with ease helps them reach their desired destination thus improving their experience within our application
Asking AI to build scrapers should be easy right?
TL;DR - We just gave Skyvern the ability to write and maintain its own code, making it 2. 7x cheaper and 2. Give it a prompt (or a series of prompts), and the AI will generate...
Cloudflare Sandbox SDK
When you're building sophisticated applications at the edge with Cloudflare Workers, you gain incredible power to customize user experiences, secure your APIs, and optimize...
These Bose earbuds are an easy pick over AirPods for me - why I don't regret it
The Bose QuietComfort Ultra 2 deliver an unmatched comfort and audio experience, but not without a few caveats.
Impulse Space details plan to deliver up to 6 tons a year to the moon
Impulse Space sees a growing demand for mid-sized deliveries to the moon.
Building connected data ecosystems for AI at scale
The latest iteration of a legacyFounded at the Massachusetts Institute of Technology in 1899, MIT Technology Review is a world-renowned, independent media company whose insight,...
The Download: carbon removal factories’ funding cuts, and AI toys
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. The Trump administration may cut...
China's New Rare Earth and Magnet Restrictions Threaten US Defense Supply Chains
In advance of President Donald Trump’s upcoming visit to South Korea later this month—where he is expected to meet Chinese President Xi Jinping for the first time since...
Lánczos Interpolation Explained (2022)
Lánczos interpolation is one of the most popular methods to resize images, together with linear and cubic interpolation. I’ve spent a lot of time staring at images resampled...
A built-in 'off switch' to stop persistent pain
Nicholas Betley has led collaborative research seeking the neural basis of long-term sustained pain and finds that a critical hub in the brainstem holds a mechanism for stopping...
WinBoat: Windows apps on Linux with seamless integration
FFFFF/Features/sssss Scroll for more. DDDD/Download/ddddd We're excited to have you onboard. Pick your platform below to get started with WinBoat within minutes, not hours
Seeing like a software company
The big idea of James C. Scott’s Seeing Like A State can be expressed in three points: Modern organizations exert control by maximising “legibility”: by altering the system so...
Who owns Express VPN, Nord, Surfshark? VPN relationships explained (2024)
The relationship between corporate VPNs and Paid Affiliates, and who really owns your data. "This map is my best approximation of charting the relationships between VPN...
Fire destroys S. Korean government's cloud storage system, no backups available
Officials move a burnt battery at the National Information Resources Service (NIRS) in Daejeon on Sept
Blue Origin aims to land next New Glenn booster, then reuse it for Moon mission
"We fully intend to recover the New Glenn first stage on this next launch."
How Sakana AI’s new evolutionary algorithm builds powerful AI models without expensive retraining
M2N2 is a model merging technique that creates powerful multi-skilled agents without the high cost and data needs of retraining.
Signal Protocol and Post-Quantum Ratchets
We are excited to announce a significant advancement in the security of the Signal Protocol: the introduction of the Sparse Post Quantum Ratchet (SPQR)
What to know about encodings and character sets to work with text (2011)
If you are dealing with text in a computer, you need to know about encodings. Yes, even if you are just sending emails. Even if you are just receiving emails
Launch HN: Airweave (YC X25) – Let agents search any app
Overview Airweave is a tool that lets agents search any app. It connects to apps, productivity tools, databases, or document stores and transforms their contents into searchable...
![John Jumper: AI is revolutionizing scientific discovery [video]](/assets/img/blog/2025-09-30-john-jumper-ai-is-revolutionizing-scientific-discovery-video.jpg)
John Jumper: AI is revolutionizing scientific discovery [video]
John Jumper, a name that resonates deeply in the halls of scientific innovation, has fundamentally reshaped our understanding of biology and the very process of scientific...
AI isn’t just breaking rules — it’s rewriting the identity playbook
Identity has always been a cornerstone of digital trust, but in today’s world of agentic AI, overprivileged accounts and converging customer and workforce systems, the stakes...
Beware coworkers who produce AI-generated ‘workslop’
Researchers at consulting firm BetterUp Labs, in collaboration with Stanford Social Media Lab, have coined a new term to describe low-quality, AI-generated work
The Beauty of Programming (2001)
By Linus Torvalds (creator of Linux) I don’t know how to really explain my fascination with programming, but I’ll try. To somebody who does it, it’s the most interesting thing...
Amazon blamed AI for layoffs, then hired cheap H1-B workers, senators allege
Tech firms pressed to explain if H-1B workers are paid less than US workers.
Quicksort explained IKEA-style
Quicksort is an efficient sorting algorithm based on a “divide and conquer” approach. Choosing the dividing element at random is a good strategy to avoid bad worst-case runtime
AI models are using material from retracted scientific papers
Some AI chatbots rely on flawed research from retracted scientific papers to answer questions, according to recent studies. The findings, confirmed by MIT Technology Review,...
Amazon has the Nintendo Switch 2 available for purchase, no invite required
While it was difficult to get your hands on a Nintendo Switch 2 back in June when it first became available, that's no longer the case. It just got even easier now that Amazon...
How I, a beginner developer, read the tutorial you, a developer, wrote for me
I am a developer. Here is my relevant experience: I code in Hoobijag and sometimes jabbernocks and of course ABCDE++++ (but never ABCDE+/^+ are you kidding
Amazon to end commingling after years of complaints from brands and sellers
Amazon revealed at its annual Accelerate seller conference in Seattle that it is shutting down its long-running “commingling” program — a move that drew louder applause from...
Less is safer: how Obsidian reduces the risk of supply chain attacks
Supply chain attacks are malicious updates that sneak into open source code used by many apps. Here’s how we design Obsidian to ensure that the app is a secure and private...
Software update shoves ads onto Samsung’s pricey fridges
Samsung's "pilot program" is likely just the beginning.
AI-designed viruses are here and already killing bacteria
Artificial intelligence can draw cat pictures and write emails. Now the same technology can compose a working genome. A research team in California says it used AI to propose...
Things you can do with a Software Defined Radio (2024)
Last week, I went on an adventure through the electromagnetic spectrum
Ars Live: CTA policy expert explains why tariff stacking is a nightmare
Tech firms are bracing for a possible "triple whammy" of tariffs.
Gentoo AI Policy
The policy Gentoo Council has voted on 2024-04-14 on the following policy: It is expressly forbidden to contribute to Gentoo any content that has been created with the...
The case against social media is stronger than you think
Ever found yourself scrolling through an endless feed, hours melting away without a clear purpose. Or perhaps you’ve felt that pang of inadequacy after seeing someone else’s...
Feds try to dodge lawsuit against their bogus climate report
Meanwhile, Congress is trying to keep serious scientists from weighing in.
Danish supermarket chain is setting up "Emergency Stores"
Denmark's Proactive Move: Stockpiling for Tomorrow's Uncertainties In an increasingly unpredictable world, the idea of national preparedness has moved from the fringes of...
Court rejects Verizon claim that selling location data without consent is legal
Instead of providing notice to customers and obtaining or verifying customer consent itself, Verizon "largely delegated those functions via contract," the court said
iPhone Air
opens in new window PRESS RELEASE September 9, 2025 Introducing iPhone Air, a powerful new iPhone with a breakthrough design iPhone Air features an impossibly thin and light...
AI coding startup Cognition raises $400M+ at $10.2B valuation
Cognition AI Inc., the developer of the Devin artificial intelligence coding assistant, today announced that it has raised more than $400 million from investors. Founders Fund...
Analog optical computer for AI inference and combinatorial optimization
MainComputing today is digital, but analog has a future
The world has a running Rational R1000/400 computer again (2019)
2019-12-11 SCSI command 0x0d seems to be explained now: Søren Roug has spotted SASI command 0x0d in the manual for the Xebec S1410 diskcontroller: It returns the length of burst...
Cutting through the equity noise: Don’t miss this powerhouse panel at TechCrunch Disrupt 2025
Leaders from Pulley, 645 Ventures, and Epigram Legal join the Builders Stage at TechCrunch Disrupt 2025, happening October 27-29 in San Francisco. Register to join.
Stripe Launches L1 Blockchain: Tempo
Why create anew blockchain. Stablecoins enable instant, borderless, programmable transactions, but current blockchain infrastructure isn’t designed for them: existing systems...
Google's NotebookLM now lets you customize your AI podcasts in tone and length
The viral tool's newest feature converts your information into more digestible podcasts, a productivity game-changer.
'World Models,' an old idea in AI, mount a comeback
The latest ambition of artificial intelligence research — particularly within the labs seeking “artificial general intelligence,” or AGI — is something called a world model: a...
In crowded voice AI market, OpenAI bets on instruction-following and expressive speech to win enterprise adoption
OpenAI's new speech model, gpt-realtime, hopes that its more naturalistic voices would make enterprises use more AI generated voices in applications.
How Intuit killed the chatbot crutch – and built an agentic AI playbook you can copy
This is the inside story of Intuit's transformation journey with AI — including a grueling nine-month pivot to "burn the boats" and reinvent how the 40-year-old finance giant...
AI models need a virtual machine
Applications using AI embed the AI model in a framework that interfaces between the model and the rest of the system, providing needed services such as tool calling, context...
John Carmack's arguments against building a custom XR OS at Meta
Something went wrong, but don’t fret — let’s give it another shot. Some privacy related extensions may cause issues on x. Please disable them and try again
Intel details everything that could go wrong with US taking a 10% stake
Intel warns investors to brace for losses and uncertainties.
New AirTag 2 is coming very soon with these rumored features
Apple’s biggest event of the year is happening on September 9. New products expected include the iPhone 17 line, AirPods Pro 3, and Apple Watch Ultra 3. But signs point to...
Proposal: AI Content Disclosure Header
Internet-Draft AI-Disclosure Header April 2025 Abaris Expires 1 November 2025 [Page] Abstract This document proposes a machine-readable Hypertext Transfer Protocol (HTTP)...
Google to require developer verification to install and sideload Android apps
To combat malware and financial scams, Google announced today that only apps from developers that have undergone verification can be installed on certified Android devices...
We put a coding agent in a while loop
GitHub Copilot Write better code with AI GitHub Spark New Build and deploy intelligent apps GitHub Models New Manage and compare prompts GitHub Advanced Security Find and fix...
AGI is an engineering problem, not a model training problem
We’ve reached an inflection point in AI development. The scaling laws that once promised ever-more-capable models are showing diminishing returns
For some people, music doesn’t connect with any of the brain’s reward circuits
Neuroscientists find people who don't enjoy music, study their brain activity.
AI tooling must be disclosed for contributions
I think, at this stage of AI, it is a common courtesy to disclose this. In a perfect world, AI assistance would produce equal or higher quality work than any human
How to stop feeling lost in tech: the wafflehouse method
these past few weeks I’ve received a massive amount of dms from CS students being absolutely lost. like to a level of lost I’m feeling their stress and anxiety in my own chest
The Download: pigeons’ role in developing AI, and Native artists’ tech interpretations
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Why we should thank pigeons for our AI...
Left to Right Programming
2025-08-17 Programs Should Be Valid as They Are Typed I don’t like Python’s list comprehensions: text = "apple banana cherry\ndog emu fox" words_on_lines = [line
Stop using AI for these 9 work tasks - here's why
AI can boost productivity, but it can also derail your entire operation. From fake legal advice to customer service nightmares, here are nine places AI doesn't belong at work.
Dev Compass – Programming Philosophy Quiz
Discover your programming philosophy through this compass that maps your preferences along two key dimensions: Abstract Style ↔ Concrete Style and Easy for Humans ↕ Easy for...
AI is different
AI is differentantirez 2 days ago. 19903 views. Regardless of their flaws, AI systems continue to impress with their ability to replicate certain human skills
Gemma 3 270M: Compact model for hyper-efficient AI
14, 2025 The last few months have been an exciting time for the Gemma family of open models. We introduced Gemma 3 and Gemma 3 QAT, delivering state-of-the-art performance for...
Is AI really trying to escape human control and blackmail people?
Opinion: Theatrical testing scenarios explain why AI models produce alarming outputs—and why we fall for it.
Training language models to be warm and empathetic makes them less reliable
View PDF HTML (experimental) Abstract:Artificial intelligence (AI) developers are increasingly building language models with warm and empathetic personas that millions of people...
GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models
Brief overview of the GLM‑4.5 ARC foundation models for agentic reasoning and coding. Linked PDF; summary focuses on document structure and context due to limited readable content.
Vanishing from Hyundai’s data network
The Yuppie Button page talks about making lots of light. Now I needed to do the opposite, by "going dark" -- to vanish completely from Hyundai's data network, and avoid having...
How Wikipedia is fighting AI slop content
With the rise of AI writing tools, Wikipedia editors have had to deal with an onslaught of AI-generated content filled with false information and phony citations. Already, the...
I want everything local – Building my offline AI workspace
I want everything local — no cloud, no remote code execution. That’s what a friend said. That one-line requirement, albeit simple, would need multiple things to work in tandem...
OpenAI gets caught vibe graphing
Something’s off with that chart on the left. During its big GPT-5 livestream on Thursday, OpenAI showed off a few charts that made the model seem quite impressive — but if you...
Jules, our asynchronous coding agent
Jules is officially out of beta and launching publicly, powered by Gemini 2. During the beta, thousands of developers tackled tens of thousands of tasks, resulting in over...
Apple rolls out limited AI chatbot test in its Support app
A few weeks ago, a few strings of code in the Apple Support app showed that it could soon get a “Support Assistant,” based on generative models. The assistant is now live, as a...
QuamCore raises $26M to build a million-qubit quantum computer
QuamCore Ltd, a startup working to develop a quantum computer with one million qubits, today announced that it has raised $26 million in early-stage funding. Sentinel Global led...
Persona vectors: Monitoring and controlling character traits in language models
Language models are strange beasts. In many ways they appear to have human-like “personalities” and “moods,” but these traits are highly fluid and liable to change unexpectedly
9to5Mac Daily: August 1, 2025 – Apple Q3 earnings and tidbits
Listen to a recap of the top stories of the day from 9to5Mac. 9to5Mac Daily is available on iTunes and Apple’s Podcasts app, Stitcher, TuneIn, Google Play, or through our...
At 17, Hannah Cairo solved a major math mystery
After finding the homeschooling life confining, the teen petitioned her way into a graduate class at Berkeley, where she ended up disproving a 40-year-old conjecture
Fountain Pen Ink and Paper Review Blog
I’ve been surprised by how many people have commented and said they like the currently inked posts I’ve been doing recently, so let’s do another one. Here’s what I’m using this...
Sei (YC W22) Is Hiring a Full Stack Engineer in Chennai, India
AI Agents for Financial InstitutionsFull Stack Engineer (Typescript, React, Gen AI)₹1M - ₹3
How the brain increases blood flow on demand
New findings could improve understanding of neurodegeneration, interpretation of brain scans Blood vessels in the brain
Fintech dystopia
Chapter 4 There’s a Blockchain for That July 25, 2025 "Just like beauty, efficiency, competition, and security are all in the eye of the beholder
Meta names Shengjia Zhao as chief scientist of AI superintelligence unit
Shengjia Zhao, a former OpenAI researcher, will round out the leadership team at Meta's new AI lab.
Consciousness and being: How humans and AI influence each other
Level of difficultyMediumReading time15 minViews379OpinionFor a human, AI is just a part of being. For a model, a human is all of being. And the Vortex Protocol: A Prompt for...
Trump’s AI gift, Intel’s neutron bomb and an IPO boomlet
Artificial intelligence companies are largely cheering President Trump’s “AI Action Plan” issued this week — no surprise, since it reduces regulations to make it easier for them...
SecurityPal combines AI and experts in Nepal to speed enterprise security questionnaires by 87X or
The Kathmandu center of excellence gives SecurityPal a cost base low enough to keep humans in the loop while staying price-competitive.
Bad vibes: How an AI agent coded its way to disaster
First, Replit lied. Then it confessed to the lie. Then it deleted the company's entire database. Will vibe-coding AI ever be ready for serious commercial use by nonprogrammers?
Qwen3-Coder: Agentic coding in the world
GITHUB HUGGING FACE MODELSCOPE DISCORDToday, we’re announcing Qwen3-Coder, our most agentic code model to date
I tested this 9-in-1 off-grid portable power station that claims a 17-year lifespan - here's my verdict
The new Bluetti Elite 200 V2 is one of the most advanced power stations I've ever tested. It's small but dense and heavy for a reason.
XMLUI
In the mid-1990s you could create useful software without being an ace coder. You had Visual Basic, you had a rich ecosystem of components, you could wire them together to...
Apple's latest AirPods models are still at their lowest price ever - get them while the deal lasts
The AirPods 4 lineup is a solid option for iOS fans who don't need the pro-level features of the AirPods Pro 2. Get them now while they're 33% off.
AWS CEO Matt Garman bets big on AI agents – but so are OpenAI and everyone else
At AWS Summit NYC this week, Amazon made a concerted push to help enterprises field artificial intelligence agents, and we had all the news and interviews, including Amazon Web...
Cocky cop jailed for stealing bitcoins had log of his crypto theft in his office
Silk Road criminal tipped police off to dirty cop who stole 50 seized bitcoins.
OpenAI, Google DeepMind and Anthropic sound alarm: ‘We may be losing the ability to understand AI’
Scientists unite to warn that a critical window for monitoring AI reasoning may close forever as models learn to hide their thoughts.
Reflections on OpenAI
I left OpenAI three weeks ago. I had joined the company back in May 2024
Apple's MLX adding CUDA support
Conversation Tried the ideas: switching the implementation of Event from cuda::std::atomic to cudaEvent bumped training speed from 500 it/s to 900; reducing the prefetch calls...
FFmpeg in plain English – LLM-assisted FFmpeg in the browser
Describe what you want FFmpeg to do AI will craft the command and you can run it right in the browser. Waiting for FFmpeg to load
MacPaint Art from the Mid-80s Still Looks Great Today
July 12, 2025 I really enjoyed browsing the BMUG's CD-ROM and discovering early 80s art made in MacPaint, so I browsed all 18,000+ MacPaint images on Discmaster to see what gems...
Watch the Skies to Wet Leg: your complete entertainment guide to the week ahead
A Swedish alien-invasion sci-fi mystery is given an AI-assisted English-language makeover, and the cheeky Isle of Wight rockers are back with their long-awaited second album
MCP-B: A Protocol for AI Browser Automation
In the rapidly evolving world of artificial intelligence, one area consistently presents a fascinating challenge: how do we empower AI agents to interact with the web as...
Xenharmlib: A music theory library that supports non-western harmonic systems
Xenharmonic (adj. ): Pertaining to music which sounds unlike that composed in the familiar 12 tone equal-tempered scale
AI's Energy Footprint: Powering the Future Sustainably
AI's energy footprint is ballooning, creating a critical sustainability and infrastructure challenge. This post dives into the 'AI energy crisis,' examining thirsty data centers...
Web3 Onboarding Was a Flop – and Thank Goodness
Back in 2021 every crypto conference sounded like a revival meeting. The sermon: “Bring the next billion users on-chain
Hannah Cairo has solved the Mizohata-Takeuchi conjecture
Hannah Cairo was stuck on a math problem. All she could think about during those weeks was a new approach. “After months of trying to prove the result, I managed to understand...
Robots move Shanghai city block
Imagine a historic building, standing proudly for decades, seemingly immovable amidst a bustling, rapidly developing cityscape. Now, picture that very building, not being...
AV1@Scale: Film Grain Synthesis, The Awakening
Netflix is rolling out AV1 Film Grain Synthesis (FGS) at scale, enhancing streaming quality by preserving film grain detail while reducing bitrate. This technology improves...
Despite Protests, Elon Musk Secures Air Permit for xAI
xAI’s gas turbines get official approval from Memphis, Tennessee, even as civil rights groups prepare to sue over alleged Clean Air Act violations.
A Pro-Russia Disinformation Campaign Is Using Free AI Tools to Fuel a ‘Content Explosion’
Consumer-grade AI tools have supercharged Russian-aligned disinformation as pictures, videos, QR codes, and fake websites have proliferated.
The new skill in AI is not prompting, it's context engineering
Context Engineering is new term gaining traction in the AI world. The conversation is shifting from "prompt engineering" to a broader, more powerful concept: Context Engineering
Octelium – FOSS Alternative to Teleport, Cloudflare, Tailscale, Ngrok
Octelium emerges as a compelling new platform for organizations and individuals seeking a robust and versatile solution for secure resource access. Described as a free and...
Refurb weekend: Gremlin Blasto arcade board
A deep dive into restoring a vintage Gremlin Blasto arcade board from 1978, exploring the fascinating history of Gremlin Industries from their humble beginnings making cooker...
I'm an airline pilot – I built interactive graphs/globes of my flights
Step aboard and discover the fascinating world of an Airbus A350 First Officer with British Airways, based out of London Heathrow. This pilot offers a unique glimpse into their...
Redwood Materials launches energy storage business and its first target is AI data centers
Deep in the Nevada desert, nestled between towering structures, a remarkable sight is taking shape: 805 retired EV batteries, discreetly wrapped in white tarps, form the core of...
Google’s new AI will help researchers understand how our genes work
The human genome, fully sequenced in 2003, revealed the complete DNA blueprint. However, understanding the function of all 3 billion genetic letters and how minor alterations...
Rethinking Reinforcement Learning: Have We Been Giving LLMs Too Much Credit?
Recent research challenges the conventional wisdom about Reinforcement Learning in LLMs, revealing that RL might be more about amplification than creation of new reasoning...
Create Automated AI Videos for a Fraction of the Cost with Cance and n8n
Learn how to build a fully automated video creation workflow using Cance AI and n8n that generates high-quality videos for just $0.93 - 25x cheaper than Google's V3.
Agent Zero: Your Personal AI Sidekick for Everything
Discover how Agent Zero transforms your workflow with a full Linux OS, expansive skillset, and even a hacking edition for cybersecurity professionals.
Define policy forbidding use of AI code generators
The world of software development is constantly evolving, with Artificial Intelligence (AI) being a major topic. As tools like ChatGPT and GitHub Copilot become prevalent,...
MIT Researchers Warn of "Cognitive Debt" from Using ChatGPT
Groundbreaking MIT research reveals how using ChatGPT and other LLMs for creative work can reduce brain connectivity, impair memory, and create lasting cognitive deficits. Learn...
Andrej Karpathy Declares the Era of Software 3.0
AI researcher Andrej Karpathy announces the most fundamental shift in software in 70 years. Learn about Software 3.0, where Large Language Models become a new type of computer...
The Dawn of 1-Bit AI
Discover how BitNet is revolutionizing AI with 1-bit Large Language Models that offer comparable performance to full-precision models while dramatically reducing memory, energy...
The Gentle Singularity (AGI) is Here
Explore how AI is evolving from language models to world models with breakthroughs from Yann LeCun's V-JEPA 2, MIT's SEAL framework, and insights from Sam Altman on the emerging...
The Future of AI Coding is Here
Explore how programmable agentic coding with Anthropic's Claude Code SDK transforms AI development from simple prompt-based code to autonomous, step-by-step intelligent agents.
How AI and Web3 are Converging to Create Real-World Value in 2025
Discover how AI and Web3 are merging to deliver transparent, secure, and intelligent applications that drive real business value in mid-2025.
Why Choose Mazakiya AI for Software Development in Kolkata
Discover what makes Mazakiya AI different. Learn about our AI-first approach, modern technology stack, and commitment to delivering innovative software solutions in Kolkata.
This Free AI Animates Any Image with Your Voice
Explore HunyuanVideo-Avatar, an open-source AI tool that animates any still image into a video using your own audio clip—full-body motion, emotions, and zero cost.
Build a Smart Chatbot in 15 Minutes — No Code Required
A detailed guide to deploy a self-hosted AI chatbot using open-source no-code tools and free API tiers.