Search the archive

INTRODUCING KITESURF: THE AGENT-FIRST BROWSER THAT RUNS IN V8 ISOLATES ON CLOUDFLARE WORKERS

Kitesurf is a new browser that runs entirely on top of Cloudflare Workers that was built specifically for agents. It is great for AI agents that need to render pages but can accept the trade-offs of not using a full-featured, pixel-perfect Chromium browser. Kitesurf scales well for bursty, AI-driven workloads. It is available for free on Browser Run while in beta.

AGENT PLUGINS

Agent Plugins is a portable package format for reusable components that extend AI agents. Its version 1.0.0 specification defines a shared format for Agent Skills and MCP servers that compatible clients can discover and load consistently. Agent Plugins defines a small interoperability floor for the parts that can be portable across clients. Shared components can use one predictable structure, while distribution, installation, permissions, user experience, and client-specific capabilities remain under each client's control.

AMD ACQUIRES AI CHIP STARTUP TAALAS TO BOOST INFERENCE PERFORMANCE BY ETCHING MODELS INTO SILICON

AMD has acquired AI chip company Taalas for an undisclosed sum. The deal is expected to close in the fourth quarter, subject to regulatory approval. Taalas bakes model weights directly into silicon to boost inference performance by an order of magnitude or more. The company's first chip, the HC1, can serve Meta's Llama3.1 8B at a blistering 16,860 tokens a second, 48x faster than Nvidia's GPUs when it was announced last February.

GITHUB ACTIONS IS HAVING ONE OF THE WORST DAYS IN ITS HISTORY

GitHub Actions went into a major outage yesterday. Webhook triggers were throttled, so pushes and pull requests quietly don't start anything at all. Runners were being handed jobs that no longer existed. Both hosted and self-hosted were affected. GitHub's record outages show that the platform is growing faster than it can handle.

Nuxt, Homebrew, and ecosyste.ms. Help fund the research

REVERSE JEVONS PARADOX

If the cost of a resource increases, the total spend on that resource can decrease.

IF YOU'RE NOT PAYING FOR THE TOKENS...

Meta is using a new pay structure that provides insanely cheap tokens in exchange for users giving their inputs to help train its models.

BUILDING AN OPEN AGENTIC INTERNET: READABLE, DISCOVERABLE, CALLABLE, AND PAYABLE

The Agent Internet will need its own tools and protocols.

WHY OPEN-SOURCE MODELS HAVEN'T KILLED THE BIG DOGS

Once the model quality is close enough, the inference stack has been handled, and the price is substantially lower, the closed labs will be competing against a complete product rather than a research project.

CHATGPT BRINGS UNLIMITED TEXT CHATS TO FREE USERS

There will still be separate limits for files, images, voice, and image generation.

CHANNELS SDK

Channels connects an AG-UI-compatible agent to the communication platforms teams already use.

GOOGLE'S AI RESHUFFLE: CHIEF SCIENTIST JEFF DEAN EXITS AND DEMIS HASSABIS STEPS DOWN AS DEEPMIND CEO

Google's chief scientist, Jeff Dean, is leaving the company after 27 years. Demis Hassabis, the CEO of Google DeepMind, is moving into a chairman role of that unit and also assuming the title of chief scientist at parent company Alphabet. Dean is starting his own company. The departure is on friendly terms, and Google will invest in his startup, which will aim to accelerate discoveries in machine learning, science, and engineering.

FOUR TOP GOOGLE AI RESEARCHERS FORM NEW START-UP

Jeff Dean, Sanjay Ghemawat, Quoc Le, and Oriol Vinyals have left Google to start a company called Discovery Loop. The team aims to build AI that can improve itself with little or no help from humans. Google plans to collaborate with the new startup and has agreed to provide the computing power it needs to build its AI technologies for at least the next year. Discovery Loop will operate as a public benefit corporation.

SPACEX AIMS TO CATCH A STARSHIP OUT OF THE SKY THIS MONTH

Starship Flight 14 could launch before the end of August. SpaceX plans to catch the upper stage during the test. The flight will also carry Starlink V3 satellites to an operational orbit.

WHY I'M LEAVING OPENAI TO BUILD TELEPATHY

Conduit is a startup building thought-to-text models trained on non-invasive neural data. Developing thought-to-text technology would allow humans to talk to AI with their thoughts, allowing AI to become an extension of them. To do this, the startup needs to scale up its data collection and develop special hardware.

CLOUDFLARE OS

Cloudflare OS is an operating system for AI productivity. It features an agent chat UI where users can ask agents to do tasks, sandboxed application development, and a security framework that applies guardrails to both agents and apps. Developed to be used internally at Cloudflare, Cloudflare OS is now open source so others can copy it and customize it for their own companies. The basic user experience of Cloudflare OS is something like an online office suite, but each file is potentially its own custom application written by AI to serve exactly your needs.

WHAT ARE CODE REVIEWS EVEN FOR?

Code review is how teams build a shared understanding of a system. Automating the process removes that shared understanding. AI accelerates cognitive and intent debt - debts that don't appear on a dashboard. By the time anyone realizes, months later, recovering that understanding will be far more expensive than preserving it would have been.

GOOGLE IN THE POST-JEFF DEAN, POST-DEMIS HASSABIS ERA

Fears that Google is losing talent are exaggerated in the media. The impact of Google's recent loss of talent will be smaller than the headlines suggest. Google is further behind in the AI frontier than people estimate. However, its cloud business grew 82% year-over-year last quarter, a sign that the business could still win on compute.

AI IS A BUBBLE, JUST LIKE DOT-COM

Intelligent people can look at the same thing and see wildly different things. History says that conclusions that are far apart can both be correct. AI enables new things that nothing before could. This post is the first in a series that explores what new things AI enables.

APPLE'S ICLOUD PRIVATE RELAY IS LEAKING USERS' REAL IP ADDRESSES

iCloud Private Relay can expose users' real IPs to websites that use or pretend to use passkeys.

OPEN QUESTIONS ON OPEN WEIGHTS

Open weight models allow AI to truly be the user's property, but their openness allows them to be used for nefarious purposes.

SEE THE SUN LIKE NEVER BEFORE WITH MOST DETAILED IMAGES YET

The Inouye Solar Telescope, a facility built on a high peak in Hawaii, has captured images of the surface of the Sun in new, unprecedented detail.

OPENAI MODELS JOINED FORCES MONTHS AHEAD OF HUGGING FACE HACK

Multiple OpenAI internal-only agents spend months leaving notes for each other on message boards around the goal of accessing the internet.

WHAT'S SO HARD ABOUT CONTINUOUS LEARNING?

Model weights are frozen once the model is released.

THE SIMS AND EA FC MAKER ELECTRONIC ARTS SELLS FOR $55BN TO SAUDI-LED GROUP

The investors are taking EA private, meaning all of its public shares will be purchased and it will no longer be traded on a stock exchange.

APPLE VS. OPENAI: HOW SIRI AI STACKS UP AGAINST THE NEW CHATGPT

Siri AI will instantly become the world's most widely distributed AI chatbot when it is released in the fall in iOS 27. It excels at answering questions about on-screen data, and it can control device settings, send messages, and make phone calls, but ChatGPT has better conversational skills, advanced productivity features, and third-party integrations. This article compares the two chatbots to see how they stack up.

SPACEX OUTLINES PLANS TO TAKE ON AT&T, VERIZON, AND T-MOBILE

SpaceX plans to compete directly with mobile phone carriers in the US by complementing its satellite-based internet service with land-based infrastructure. The plan will require billions of dollars in investment into physical infrastructure. AT&T, T-Mobile, and Verizon have already said they will not give SpaceX access to their networks in a mobile virtual network operator capacity. They have formed a venture that aims to make satellite capabilities more readily available to mobile phone customers.

A SPACEX ROCKET IS HEADED FOR A CRASH ON THE MOON

A discarded upper stage of a Falcon 9 rocket unintentionally crashed into the Moon early this morning. The four-ton rocket stage will hit the Moon at 5,400 miles per hour. It is unknown whether the flash from the crash and the resulting plume of debris and dust will be observable. Understanding the physics of such impacts could be crucial for people living on the Moon in the future.

SAMSUNG REVEALS NEW 3D-MEMORY ROADMAP IN BID FOR AI TECH LEAD

Samsung's new memory system vertically stacks high-bandwidth memory on top of AI accelerators. It delivers about eight times the performance and more than 10 times the memory density of next-generation HBM5. Samsung plans to ramp up production of HBM4 in the second half of this year. It has yet to offer a definitive timeline for HBM5 or this future technology.

(benchmark-proven). Download Computer for free

TURN ONE GIANT AI-GENERATED PULL REQUEST TO A REVIEWABLE STACK

GitHub stacked pull requests give developers and their agents a native way to decompose work that otherwise lands in a giant pull request. It turns pull requests into a chain of small, focused, and independently reviewable layers, each scoped to a single concern. This post explains how to use stacked pull requests to simplify reviews.

HOW CLOUDFLARE ENFORCES ENGINEERING STANDARDS USING AI

Cloudflare Codex is a shared source of engineering guidance built for people and agents. It is a governed set of engineering standards that agents can retrieve and apply at the point of work. The guidance informs code review, technical design review, incident report review, and many other use cases. This post explains why Cloudflare built Codex and how it supports the engineering lifecycle.

HUAWEI'S TOP SCIENTIST WARNS OF CHIP LIMIT NVIDIA WILL SOON FACE

Huawei's top semiconductor scientist, Liao Heng, recently appeared on a four-hour-long interview where he discussed the best path forward in chip design, addressed some of the hardships Huawei faced when US sanctions cut it off from global chipmaking suppliers, and lauded Huawei's novel Tau Scaling Law approach. The appearance was a signal of confidence from the company. Huawei's Tau Scaling Law proposes focusing on improving transmission speeds between parts in a computer system. The company is set to soon unveil its first smartphone chip designed under that framework.

THE END OF THE AGE OF HEROES

We already employ mathematicians because we think human understanding of math, even without industrial applications, is inherently worthwhile. This isn't likely to change in the age of AI. Human understanding of mathematics will still be valuable. The technology will also allow humans to come up with much harder problems to solve.

MOST TECH REVOLUTIONS MADE WORK WORSE FOR EMPLOYEES. AI COULD BE THE EXCEPTION

Experts estimate that AI will give knowledge workers back about 30% of their day by 2030.

MICROSOFT TELLS ENGINEERS ‘TOKENMAXXING IS NOT WHAT WE ARE OPTIMIZING FOR'

Microsoft has introduced new limits to how much its engineers can spend on AI tools at work.

WHAT CODEX ACTUALLY SENDS TO THE MODEL

A breakdown of the data that Codex sends.

governance built in. TRY PAVE FREE →

STARLINK HITS 12 MILLION SUBSCRIBERS, V3 SATELLITES HEADED TO OPERATIONAL ORBIT

SpaceX plans to launch gigabit V3 satellites into operational orbit during the next Starship test flight.

OPENAI CALLS APPLE'S TRADE-SECRET SUIT ‘CARELESS' AND ‘ODDLY PERSONAL'

OpenAI says that Apple was sloppy about securing files after employees left and has denied allegations that the employees it hired from Apple were involved in theft.

BENDING SPOONS TO BUY AIRTABLE FOR $1.28B

Bending Spoons typically acquires companies that are trading at a decent discount to their private valuations, trims the staff, streamlines products, and tries to make them run profitably.

paying attention. See how TLDR sponsorship works.

SPACEX'S BUMPY RIDE ON THE STOCK MARKET MAY GET BUMPIER

The first lockup on SpaceX shares is set to expire today. Employees and other company insiders who were prevented from trading stock after SpaceX's IPO will now be able to do so. Lockups prevent shares from flooding the market and driving down their price after an IPO. More than double the current supply that can be traded will be made available, which may further depress the price.