Nvidia CEO Jensen Huang had a romantic relationship with Alice Huang, a Chinese businesswoman whom US intelligence agencies have concluded was a spy for Beijing. The company has opened an investigation into Jensen Huang after the disclosure of the relationship. The board says that it has no reason to believe that either national security or Nvidia secrets were compromised. Alice Huang was likely recruited to gather information on Jensen Huang and potentially influence Nvidia to push for its most advanced chips to be made available to China.
Apple's software chief Craig Federighi is being pressed to pursue a more ambitious approach to AI as the company aims to transform Siri into an autonomous agent. Federighi has a reputation for being an AI skeptic. His leadership style has sometimes complicated product development and led to delays or less ambitious initial versions of devices, but he has demonstrated an ability to execute when others could not. Siri AI finally fulfills Apple's vision of a working voice assistant, but it is still far from the advanced tools offered by rivals.
Amazon has assembled a new team over the past several months to work on a robot product capable of navigating dynamic home environments. The robotics market is expected to surge in the coming years. Amazon redeployed most of the team behind its Astro robot earlier this year and hired an MIT roboticist to work on the new home robot project. It acquired Fauna Robotics, a New York-based startup with a 3.5-foot-tall humanoid robot, in March to create a new division called the Personal Robotics Group.
We are now in an age where longstanding open problems in science and math are dropping like flies. AI is producing proofs at a rate that is hard to keep up with. While many of these proofs have yet to be validated by humans, their Lean certificates provide the confidence needed for their validity. This is causing a stir within the research community.
Factory, Exa, and Bolt connected their coding agents to Sentry so production errors turn into fix PRs, not tickets. See how they did it, then use the recipes for Claude Code, Cursor, or Copilot to build your own. See how it works.
Large language models don't know how to design software that uses LLMs. This could be because most training code is pre-LLMs, or RLVR may just not be a good environment to write code that calls LLMs. Also, current AI agents might just not be good enough at software design for the task. LLMs are good and bad seemingly at random, but if you can predict where they will do well or badly ahead of time, you can get enormous value from them.
Programming is much more tiring with AI. This is a sign that you are diligently working on things that require human brainpower and judgment. AI is taking over the repetitive and obvious parts of the job, leaving developers to deal with higher-judgment tasks. It is no surprise that these would be more tiring.
AI is making it increasingly hard to tell good work from bad work and fakes. Humans rely on shortcuts based on signals to make decisions all the time. AI widely undermines these signals and lowers the incentive to produce good work. The speed and scale of the current signal collapse is unprecedented. We need signals to survive in the modern world, and the rapid collapse of many previously dependable signals will create large shifts in society.
The amount of software being created is going to go up much faster than the number of people paid to make it. AI will enable far more software of a kind nobody used to make. The developer jobs that grow will be the ones that the machines made faster. The demand for software can be bottomless and the work can still move somewhere else.
On 113 DeepSWE GitHub issues with the same open model, CodeAF solved nearly 4x Claude Code's and 2x OpenCode's, at up to 4.8x less per fix. A new open-source harness. Try on Github.
Developers now point coding agents like Claude, Cursor, and GitHub Copilot at your API before reading your docs. Score your docs with Agent Score to see where your docs stand.
The conventional way of working with AI has sequencing and context issues - this post looks at a way to work around this by organizing work like a priority queue.
Cognition analyzed five months of production data from Global 1000 engineering teams running Devin cloud agents and found that writing code accounted for 23% of the productive engineering hours. The other 77% came from planning, testing, security, and deployment. Learn why cost per session fell as usage grew, where enterprises point agent spend, and which use cases return the most engineering hours per session.Download the Benchmark Report >
Recent election results show that voters are overwhelmingly against the massive build-out of compute capacity that Silicon Valley is pushing. This could rein in AI's rapid growth and become a big problem for the tech companies developing the technology, and could have effects on the wider US economy. Stalling construction could result in dire consequences for tech companies that were already struggling to secure enough compute.
Major investors planning to participate in Anthropic's IPO, expected to launch the week before Thanksgiving, say the expected $1.5 trillion valuation will likely need to come down. The AI lab was expected to raise up to $100 billion, but is now predicted to raise around $80 billion, leaving it with a valuation closer to $1 trillion. Investors have revolted against AI bets this week as the midterm elections have put several AI skeptics in power, potentially slowing data center development in the US. Anthropic has avoided committing to specific plans for its IPO to maintain optionality.
Blue Origin has announced two significant upgrades to its New Glenn rocket that reduce the cost of manufacturing the vehicle and expand its capacity for payload delivery. The company aims to eventually launch the vehicle more than 1,000 times a year. It has started test firing a new engine that will eventually replace the BE-4 engine that powers the first stage of New Glenn. The company is also developing a hydrogen-oxygen third stage for the rocket to give it an extra kick.
A SpudCell is an artificial cell created from lifeless ingredients that can grow, copy its DNA, divide in two, and start its cell cycle all over again. It is still far from a self-sustaining cell that any scientist could manufacture from prepared ingredients, but the technology could eventually make it easier to produce medicines, foods, and other items that could not otherwise be produced with ordinary biology. The researchers working on SpudCell hope that it will eventually be a platform for biotechnology that will allow a simple cell to be engineered in scores of useful ways. SpudCell could potentially be misused, so its researchers are taking security seriously from the start - its DNA and proteins are engineered so that it can only survive on an exotic diet of chemicals not found in nature.
Your agents have MCP access to Slack, Jira and GitHub. Your engineers still babysit them. Read why connectors alone fall short, and how using a real context engine can make the same coding task 83% faster and 60% cheaper.
*sshole is a vulnerability in Linux that allows attackers to log in via SSH without knowing the password. The vulnerability exists in systems using OpenSSH with musl without pluggable authentication modules. Despite the prerequisites, the vulnerabilities likely threaten a large number of systems. It is still unclear where the vulnerability resides and what other packages may be vulnerable.
GitHub's new Git storage architecture is capable of handling thousands of writes per second with reads that scale on demand. It can support 5,000 concurrent agents working on a repository while serving read traffic with consistently low latency. The new architecture is designed to support agentic development for hundreds of millions of developers. This post walks through the requirements behind the design of the new architecture and explores two of its components.
At the heart of Dario Amodei's worldview are two interwoven philosophies: rationalism, which aims to apply dispassionate thinking to the world, and effective altruism, which seeks to maximize social good with mathematical precision. These schools of thought form the foundation for Anthropic. As the company has grown, Amodei's views have become increasingly prominent. This article takes a look at how Anthropic's chief executive's thinking on AI was shaped.
A 'Hackintosh' is a non-Apple computer running macOS. The Hackintosh community started fading away after Apple released its first ARM-based Mac and discontinued support for x86 versions of macOS. However, the rise of AI has revived activity in the space, and there has been an explosion in development, activity, and success in the past few weeks. This post lists a few recent developments, including macOS 27 running on a Snapdragon laptop, iOS 17 running on the Nintendo Switch OLED, and macOS 27.2 Beta on an ARM SBC RK3588 with Mali drivers.
When an agent can't use your docs, developers don't get an error. They reach for a competitor. Find out where you stand across Cursor, Claude, and ChatGPT. Score your docs.
Tech stack complexity is out of control, and the last thing you need is a bigger ‘stack'. Microsoft Azure gives you a unified view of applications, cloud resources, and AI—so you can ship more software with less open tabs. Simplify with Azure
This essay looks at what it takes to build a physical platform capable of performing biological experiments, measuring their outcomes reliably, and generating evidence at the scale AI requires.
The average time from disclosure to attack has been falling for some time, but AI has changed who can operate inside that window, and how fast known bugs get weaponized.
Elon Musk wants to combine SpaceX's Starlink satellite broadband network with mobile antennas, fiber connections, and other ground infrastructure, so that regular smartphones can seamlessly tap into signals from both Earth and space.
The only thing evolving faster than AI agents are the startups building them. But as you race to scale your agentic systems, you need to retain deep programmatic control over agent logic, tools, and the environments they operate in.
This new guide from Google Cloud offers a technical deep-dive into code-first agent development, revealing how it's possible to stay in control and innovate fast. You'll learn how to:
🚀 Accelerate development with ADK and Antigravity.
🤖 Deploy to Agent Runtime using Agents CLI.
📈 Improve reliability and performance with AgentOps.
Apple quietly lowered the daily spending cap on AI models for its employees last month. Some engineers had their budgets cut from $300 to $200, and the company is planning further cuts for some engineering teams that may potentially reduce their daily allowances to as little as $100. Groups outside of engineering may have their limits reduced to $25 from $100. Employees say that few workers were using their full $300 daily allowances, making the previous limits unnecessarily generous.
OpenAI is rolling out Images 2.6 globally in ChatGPT and Codex. The new model generates images 50% faster than the previous iteration, and it can also create GIFs. Users can adjust the speed of GIFs, edit individual frames, and share them. GPT-Image-2.6 Flare and Sunburst are now available through the API.
SpaceX has launched a commercial page for Starfall, a disk-shaped reentry capsule that can carry up to 1,000 kg of payload. The capsule can supply 1 kW of power and connect to the internet through Starlink, allowing researchers to pull data from experiments while in orbit instead of waiting for splashdown. It can be used for manufacturing, research, or testing and is aimed at biomedical customers, pharmaceutical companies, and materials science research. Starfall's timeline will depend on Starship - SpaceX is preparing for Flight 15, where it aims to catch both stages with the launch tower.
Varda Space Industries' autonomous factories carry miniature laboratories in which materials can be processed without astronauts. Once the process is complete, a capsule separates from the spacecraft and returns to Earth. The company currently has three spacecraft simultaneously in orbit. Manufacturing in space has several advantages, one of the most important being microgravity. Varda plans to launch seven missions in 2027 and 10 in 2028.
A Chrome full chain used to take six to nine months. XBOW's AI agents did it in six days for $40,000. Google paid the first $250,000 full-chain award. Every stage, and what it means for you.
Design engineering requires being deeply involved in the technical craft and material of software. They help make decisions about functionality and architecture in code while still being a designer. Design engineers are uniquely positioned to make convincing prototypes of future theories. They are the perfect people to help figure out the shape of the future.
The new version of X.com's web app was built by five engineers working with AI coding agents in under six months. It loads pages significantly faster than the old web app over slow connections, and it already serves logged-out users and the pages that search engines crawl. This post covers why X.com was rebuilt, how the new web app is being rolled out, and how the old and new apps compare today. It also looks at the performance work completed since logged-out traffic moved over.
Voice is a natural modality that is underutilized. It is great for rambling at coding agents and for capturing context. There is a lot of ambient context that exists in conversations with a ton of downstream value. This post presents some use cases for voice as a way to interact with AI.
Nearly every relevant local ballot measure opposing data centers won or is winning, with the exception of North Dakota. The constraint on AI dispersion is becoming political. AI will likely be a scarcity for much longer than expected. This post looks at what the implications might be if Americans don't change their minds about data centers.
Ads get ignored on social media but not in TLDR! Reach developers, PMs, marketers, founders and other tech leaders where they actually pay attention. Learn more about sponsorship opportunities.
Join Kelsey Hightower and a panel of industry experts on November 9 to explore findings from the report and learn how to move beyond AI adoption to operationalization. Register today.
Sequoia Capital is closing its London office, shutting down its Arc accelerator and its Scouts program, merging its early-stage and growth teams into one, flattening its partnership, and killing the Monday partner meeting.
You can explain a bug out loud in seconds. Typing it into Cursor or Claude takes long enough that you cut the context, and the answer gets worse.
Wispr Flow lets you just talk. Ramble, pause, change your mind mid-sentence, and it types clean, formatted text with the filler words gone. It works anywhere you can type, no plugins required, 4x faster than typing.
It learns your vocabulary, from teammates' names to company jargon, on Mac, Windows, iPhone and Android.
OpenAI added 200 million users over just the past two months. Its growth is accelerating. ChatGPT's search volume is approximately 20% of Google's, and its users create 1.5 billion images with it per week. This post contains an edited transcript of an interview with ChatGPT consumer product lead Adam Fry, where he discusses OpenAI's plans to merge ChatGPT's Chat and Work modes, image generation, interactive interfaces, advertising, education, and public unease about AI.
Developers can now enable their apps to bring content and actions directly into Siri AI by using App Intents. Skills are available for more complex requests that require Siri to combine several actions into a single experience. Siri AI can now surface any entity in the Spotlight semantic index. Submissions to the App Store for apps with the new capabilities will open up in the spring.
BrainCo, a Chinese neurotech company, is developing a wearable device that suppresses appetite by stimulating the brain. It aims to bring the device to market within the next two to three years. The device will launch first in China before potentially expanding to the US. BrainCo currently markets brain-computer interface-enabled headsets to address neurological disorders such as autism, ADHD, and insomnia. Its devices can collect, decode, and stimulate neural signals.
The fringe movement of biohackers and people taking a more hands-on approach to longevity has moved into the mainstream thanks to the COVID-19 pandemic and AI chatbots. More and more people are now taking their health into their own hands and spending time, money, and energy on health and longevity. This post looks at how people view health and explores how the people at the cutting edge of health and technology approach their own health. Some of the biggest opportunities in longevity will go to the companies that help people take action and manage their health over time.
Wispr Flow Notetaker uses your dictionary and calendar to catch uncommon names and industry jargon, then pulls out decisions and next steps. Ask anything across your meetings and get linked to the source. Free on Mac and Windows. Try Notetaker free
AI-native startups can ship even faster than fast-moving pre-AI startups because they risk nothing by moving too fast. The old constraints of software engineering have disappeared, allowing startups to move faster, but also break things. The new constraints are too much context switching, losing shared context, tougher pacing, less focus on design, and AI plan limits. Many existing technologies and processes will crack under the change.
Interpolative software is code that is solidly within the training distribution of models. It can often be specified with just a few sentences of prompting. Novel software is something that has fundamentally not existed before. Writing novel software with agents is challenging, as they are not trained on the task. Not a lot of software is novel, which is fine, as there is still a ton of value in the interpolative stuff.
The economics of angel investing have shifted. Many investors have realized it's potentially more profitable to raise a fund instead of using their own money. Technology startups now move much faster, and running around grabbing small checks from angels is less palatable than grabbing a pre-seed fund to fund a whole initial round. Also, when angels try to fund a round, they are often pushed out by the valuation game.
Google will turn off AI tools in its software for K-12 students unless school administrators explicitly choose to make them available. However, schools that choose to restrict AI entirely will also lose access to Google search. Google had earlier turned on AI features as the default option and is claiming that the reversal is in response to feedback from educators and parents. The company's decision to use this strategy shows how much power it has over the US education system.
Ads get ignored on social media but not in TLDR! Reach developers, PMs, marketers, founders and other tech leaders where they actually pay attention. Learn more about sponsorship opportunities.
Wispr Flow Notetaker connects to Claude, ChatGPT, Gemini and any tool that speaks MCP. No copy-paste required. Free on Mac and Windows. Get Notetaker free
This post looks at how Namespace builds custom Mac racks, how it starts macOS instances from scratch in seconds, and how to deliver new Xcode versions mere hours after Apple releases them.
Your agent found the right page. Now it needs to read it without wading through messy HTML. SerpApi's Extract API turns public webpages into clean Markdown.
👉 Extract any page with a simple GET request and get token-efficient output your agent can use right away.
👉 Scrape Google, then extract the top results for deeper research.
👉 Need a whole site?Map Mode finds a domain's URLs, and Crawl Mode fetches connected pages.
Apple is reportedly working on an AI-driven home robot potentially named the HomeView Ultra. It plans to launch the robotic product with a possible price tag of $1,000 or more by 2027. The earliest prototypes of the home robot first came out in 2018. The robot, which will support FaceTime calls, will be able to autonomously roam around and switch lights on whenever a person enters the room.
SpaceX has held talks with EchoStar to potentially acquire Boost Mobile outright or gain access to its cell network, which is leased from AT&T and T-Mobile. Starlink recently confirmed plans to become a major mobile carrier in the US. The deal will need approval from AT&T and T-Mobile, both of which have already turned down SpaceX for access to their cell towers. SpaceX last year acquired $2.6 billion worth of spectrum licenses from EchoStar to improve its Starlink network's 5G connections.
Chinese space officials have confirmed the country's plans to land astronauts on the Moon within three years. The landings will involve a lunar-orbit rendezvous similar to what NASA did with the Apollo program. China and the US are racing to land astronauts on the lunar surface, as it would be a major geopolitical win. Both countries are also building capabilities to establish a long-term presence at the Moon's South Pole.
Nearly all roadmaps point to a utility-scale quantum computer being developed around 2030. Despite this, there is little confidence or clarity on how these machines will actually be useful. The consensus is now that they'll first be used for breaking cryptography. The second most concrete answer is to use them to generate simulation data for AI labs' science models. This post looks at how else quantum computing could be applied and where all the revenue will come from.
OpenAI's monorepo's size and demands increased at a staggering pace over the last year, necessitating a revamp of the AI lab's source control system. Its technical staff built Pando, a new backend for Git repositories optimized for concurrent pushes to many branches and to maximize read throughput. The team implemented the backend without interrupting any engineering work or changing how the team uses Git. Pando cut median push latency in production by nearly 5x, and it now delivers nearly a petabyte of Git data per day.
Agents are collapsing their entire tool catalog into one 'exec' tool. This allows models to write short scripts that chain, combine, and await actions inside isolated sandboxes. Models can now reliably combine and chain dependent actions in a single pass. In the previous standard tool-calling model, agents could only perform single actions, but the exec tool allows them to perform multiple actions.
People open source projects for various reasons. Some do it to build a company or product, others do it performatively, and some ship code just for the love of the game. Open source was never free to copy, and the code was never the moat. AI can remake a product, but it can't provide the rest of the value a company generates. The technology is helping maintainers ship and fix problems faster.
Organizations are losing attention as everyone within is doing something they think is useful, but all that work draws on the same limited capacity to understand, review, and decide. Just because work exists doesn't mean it's the right thing to do. Output is not proof of value. A working implementation only means that something can be built, not that it should be - some changes have negative value.
Fern Docs sites are now prebuilt as static pages. On customer sites like Cash App, average mobile Lighthouse scores rose from 60 to 98. Make your docs fast.
Ads get ignored on social media but not in TLDR! Reach developers, PMs, marketers, founders and other tech leaders where they actually pay attention. Learn more about sponsorship opportunities.
Marvell designs custom AI chips and other technology that connects processors and moves data around data centers, and Lumentum has technology that uses light to move data between chips and servers.
Soon, almost all researchers will engage in AI-assisted replication of their own work before submitting to a journal, which will improve the quality of research going forward.
Reasoning agents that run 24/7 place new demands on infrastructure. On November 9, join Google Chief Technologist Amin Vahdat and fellow product leaders to see how Google is rising to the challenge. The exclusive broadcast will cover:
The next horizon of AI compute: The latest advancements in TPUs supporting agentic workloads with high throughput and ultra-low latency.
AI Hypercomputer architecture: How co-designed compute, storage, networking, and open software drive efficiency across every layer.
Production scaling strategies: Approaches to shortening deployment cycles, cutting infrastructure costs, and scaling open-weight and proprietary models.
Project Suncatcher preview: A first look at Google's moonshot - taking compute into space.
Cisco has given Padmanabhan Krishnamurthy, a 27-year-old ex-startup founder who joined the company through an acquisition, the job of refactoring 98% of the company's legacy code within 24 months. Krishnamurthy has four weeks to produce a proof of concept. He has been given a budget of unlimited tokens to achieve the goal. Cisco is looking to hire more employees that fit Krishnamurthy's mould.
OpenAI's latest funding round received tepid interest from investors. The deal values the AI lab at $1.25 trillion, less than the $1.4 trillion valuation it had earlier hoped to achieve. Investors appear to have grown pickier about which companies they want to back and on what terms. The industry recently experienced a brutal selloff tied to disappointing earnings results from giants, including Google, Amazon, and Intel.
NASA is no longer an indispensable customer for SpaceX. SpaceX is now increasingly disinterested in the US government for most contracts outside of data services. The company is still fulfilling its existing contracts with the government, and at a lower cost and generally higher reliability, but it is no longer signing many large new contracts for spaceflight and services in the traditional launch and in-space flight domains. It sees the vast bulk of its future revenue coming from AI.
Meta's Max robot is designed to help out around people's homes. The prototype was built with the help of a Chinese robotics company in roughly three months. Powered by Nvidia's Jetson Thor chips, the robot weighs roughly 106 pounds and is about 1.5 meters tall, and has a display that can show facial expressions, a built-in speaker, and hands designed to pick up small objects. Meta is already working on a second prototype with updates to the design of the robot's torso and hands, as well as a projector and additional sensors to help it handle stairs better.
Ads get ignored on social media but not in TLDR! Reach developers, PMs, marketers, founders and other tech leaders where they actually pay attention. Learn more about sponsorship opportunities.
Postgres 19 most likely won't be released until November and could possibly be postponed further. Several large features were reverted quite late in the beta period. AI has made it easier to identify bugs, and this has resulted in a lot more bug reports that the PostgreSQL committers have to work on. This has taken the focus off development of new features, as the project cares about quality.
With product engineering, figuring out what the requirements are is still most of the job. This post presents an interview model where LLMs can be used in interviews to roleplay as customers, and engineers talk to the models to extract requirements. The approach rewards candidates who think of edge cases, approach things sideways, and think about whole systems at once. It's relatively cheap and can be used in conjunction with a standard human interview.
Digital minimalism is about being intentional about how you access digital technology. It is about making a conscious choice rather than something that you get sucked into by default. We need social media, the internet, and the web less than before due to AI, and there's so much you can now do purely locally and offline. The flaws of the previous paradigm are now reaching their apex, making digital minimalism more and more necessary.
TLDR is hiring a senior engineer to build the agent-first operating layer used across the company. We're looking for a builder who has taken an LLM agent to production and still owns it. Click here to learn more.
A string of unusual transactions started to cause alarm at Binance on Thursday afternoon. Binance's top executives put together an informal war room with about 100 participants in response and worked through the night to navigate one of the most damaging crises in Binance's history. The company was forced to halt deposits and withdrawals as more than $6 billion was drained from the exchange. This article details how the hack unfolded based on interviews with people with knowledge of the events and analyses by crypto forensics firms.
Shopify's Fall '26 Edition brings store connectors to nine AI agents. Manage products, inventory, and orders from the chat, or turn an idea into a storefront preview. Try it now.
This post looks at how society could give everyone a stake in productivity improvements in a way that still keeps the natural self-regulation of the free market.
The fight to regulate AI and reform American science runs through the Office of Science and Technology Policy, an underutilized office run by Michael Kratsios, who had previously worked with Peter Thiel.
