Tune in and join us on the main stage for talks featuring Lin Qiao (Fireworks), Jensen Huang (NVIDIA), and Jay Parikh (Microsoft) as they explore the future of specialized intelligence.

Plus, get access to live and on-demand sessions featuring leaders from Meta, Salesforce, Cognition, Genspark, Factory, and more.

Topics include:

• Frontier Quality Without Frontier Costs
• Building Models That Win in Your Domain
• When the Stakes Demand Specialized Intelligence

Join the leaders shaping the future of specialized intelligence.

👉 Register for the Forge livestream

DoorDash will acquire Instacart for $17 billion. Instacart shareholders will receive $65 per share, 26% over the company's price on October 22. Combining the companies' technology and infrastructure will result in faster deliveries and more affordable prices. The deal is expected to close late next year or in 2028.

Reddit will start testing a logged-in-only experience in select non-US markets in Q4. The company's CEO says the login requirement will help Reddit retain new users. It will also provide more signals for ads targeting and help protect access to the site's data. The changes will make it significantly easier for Reddit to understand who uses Reddit and why, which will help the site manage things like AI scraping and evolving search habits.

Neuralink's Blindsight implant is now ready from an engineering standpoint. The device is designed to help blind people see. It is currently waiting on regulatory approval before it can be sold to customers. It has already received Breakthrough Device Designation from the US FDA, which will help speed review of the device.

Progress has started to accelerate in science in the last two years. AI can send hordes of superhumanly fast, well-read, and mathematically gifted researchers to accelerate scientific research. It is possible AI will enable new types of discovery and ways of understanding the natural world that no network of humans would ever be able to discover. Unless humans have actually somehow discovered most of the actual exploitable regularities in the Universe, we're about to enter a new golden age of science.

Services reach customers at the moment they're ready to act, with relevant, personalized experiences from day one and built-in payments to complete tasks in conversation. The opportunity to build is now.Explore the Alexa+ Partner Opportunity

Adding constraints to AI agents usually makes them less capable. Sometimes the solution is to create a template or a complete multi-stage workflow. The challenge is in creating the deterministic behavior while retaining the agent's freedom to use the best tools for the job. This post discusses a solution that uses a man-in-the-middle egress proxy that forces behavior by denying actions and forces reactions via automated messages.

Knowing what part of inference a project relies on for speed is important. Check thinking budgets, and read traces, as they often contain valuable insight into what an agent wants to do. Give agents the tools they want rather than what you think they need. Faster hardware still matters, sandboxing and microVMs add complexity, parallelization helps, and classic software performance optimization is still relevant.

Binance has suspended transactions after a hack that the exchange said totaled around $6.3 billion. Stablecoin providers Tether and Circle moved quickly to freeze some of the stolen assets. The heist appears to have been conducted by a North Korean hacking group. Binance's founder said in a post on social media that users' funds are safe.

The standard answer for what humans will do after the Singularity is that people 'will just hang out'. However, AI will destroy the processes by which people come together and build relationships. With unlimited entertainment, people might not even want to socialize with other people. The people in the AI utopia will have no skills, education, or accomplishments, so they'll likely not be worth talking to, and you probably wouldn't want to hang out with them.

Frontier Accelerate for Marketplace Premium offers a named engagement manager, hands-on technical consultations, plus $30K in Azure sponsorship to navigate publishing requirements and prepare your solution for market. Explore Premium and check eligibility now

Ramp's AI Spend Management lets you see every AI dollar in one place, understand where to save, and set limits to keep spend on track. Try it for free.

If you want to communicate with humans, your humanity has to be visible in your prose.

The Commanding Heights are scarce, controllable positions of economic power whose holders durably capture the gains of increasingly capable and diffused intelligence.

American officials have concluded that Alice Huang, a top executive at a Singapore-based data center company suspected of steering Nvidia chips to China, was a Chinese intelligence asset feeding information directly to Beijing.

This post discusses how to build an AI writing pipeline that turns rambling recordings into drafts.

AI models are improving so quickly that many people don't realise that AI agents can now help automate particular tasks that it previously couldn't do.

The General Hill-climbing System, combined with AI, is a problem-solving framework that works on pretty much anything you throw at it.

Span surveyed more than 100 engineering leaders about how their teams use coding agents.

68% say the gap between their strongest and weakest engineers has widened since their teams adopted AI. Only 19% say effective practices become team or organization standards.

Help more engineers apply what works.

Span connects coding agent sessions to outcomes so you can identify effective practices, share them across teams, and measure progress.

Try Span free for one month →

The new MacBook Pro has an OLED touchscreen display. The display can achieve 2,000 nits of maximum brightness, and Apple claims it has a 10,000,000:1 contrast ratio. The new laptop is lighter than the previous model at just 3 pounds. The top-of-screen camera notch has been replaced by a Dynamic Island, and the Apple logo on the back of the lid glows like in older models.

Gemini 4 Carbon promises improvements in cybersecurity and complex software engineering. It is now available to selected cybersecurity defenders. A broader release will be available after Google improves guardrails on the model based on feedback. Carbon will launch at $2 per million input tokens and $10 per million output tokens, with a 95% discount for cached inputs. Prices will return to $4 per million input tokens and $20 per million output tokens when the introductory period ends.

Apeiron, a Delta-class suborbital spacecraft from Virgin Galactic, is being moved from Virgin Galactic's Spaceship Factory in Mesa, Arizona, to Spaceport America in New Mexico. After its arrival, aircraft engineers and technicians will complete work on the vehicle's avionics equipment and flight software. Apeiron will still require several ground tests and test flights before it can make a free flight. Virgin Galactic expects Apeiron's first commercial spaceflight to take place early in the second quarter of 2027.

The Grid is a factory-scale robot that can build circuit boards on demand. Every board takes its own route through the factory, which has machines that print solder paste, place parts, reflow, solder, and inspect. With The Grid, printing one unit of 100 designs is as easy as printing 100 units of one, and prototypes and production batches run on the same machines. It allows companies to manufacture huge numbers of low-volume boards, and could help bring manufacturing back to the US.

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.

The Slack Code API allows agents to start code channels, add relevant people and files, post updates as they work, and archive channels when the work is done. It enables teams to work together with agents on tasks with the same shared context. Users can see exactly what agents are doing and who approved it, and workspace controls and data boundaries are built in. Users can choose what agents and models they want to use to suit their needs.

Models have been able to create something that looks like a game for a while, but the games weren't fun. The latest generation of models exhibit meaningful improvements in game design, allowing them to create games with novel, compelling gameplay loops. The challenge now is coming up with concepts for games that wouldn't exist without AI assistance.

It is now possible to build a highly profitable business with significant scale without raising or hiring a ton. Founders can take advantage of new AI tools to generate financial freedom and build organizations to work on the things they want to, with the people they want, at a scale, ambition, and pace that works for them. This article explores how the founders of turbopuffer, a startup that powers the search infrastructure for many AI companies, built their company. It explores what life and work in the time of AI can look like, and shows how ambitious builders and founders can try it for themselves.

Google's advertising business failed to deliver on analysts' estimates in the third quarter. The company is still figuring out how to translate AI into faster sales in its core web-search business. While Alphabet reported 25% revenue for the quarter, revenue from search advertising decelerated. Its cloud business exceeded analyst expectations, drawing in $30.7 billion out of Alphabet's total of $127.6 billion in sales.

RemoFirst handles contracts, payroll, and benefits for international employees in 185+ countries. No local entity needed. Sign by Dec 31 and your first 3 months of hiring with $0 EOR fees.

Lawyers also need a convenient system of record.

Reddit's redesigned feed is designed to nudge users to engage with content beyond the communities they already follow.

The randomness used in PDF encryption is often generated very insecurely, allowing for the passwordless description of protected PDFs.

It can be difficult to tell whether you are overusing AI unless someone directly engages with the details of your work.

Demand from AI hyperscalers has quintupled prices for flash memory, something Apple previously paid pennies for.

GitHub has introduced new products and is rebuilding its infrastructure to become the most comprehensive software delivery system in the world.

Everyone's talking about the bottleneck shifting... our customers just want something stable as they scale, with actually good DX, that gets out of the way.

Buildkite is the pro-tool for CI orchestration, and when you're setup, you can forget we exist.

Programmable primitives you control. No compromises. Good enough for Cursor, OpenAI, Anthropic, Perplexity, Nvidia, Uber, Airbnb and more.

Build and test faster, more securely, more reliably, with the CI engineered for frontier scale, used by the teams setting the pace.

You deserve better CI. Start with our 30 day trial →

The iMac and the entry-level 14-inch MacBook Pro are both getting the M6 chip. The iMac will have two variants of the M6 chip, one with a 10-core CPU and a 10-core GPU, while the other will have 12 cores for each. The MacBook Pro will only have the 12-core option. Both computers will have the N1 chip for wireless connectivity, which means they will have Wi-Fi 7 and Bluetooth 6. Pre-orders will start on October 28, and they will start shipping on November 6.

Anthropic's prospectus claims the biggest risks to its business are risks the business poses to everyone else. All other human enterprise will be eaten by Claude, so you might as well own some of it. The company's governance structure is exactly what you would expect for a company worried about existential AI risks. However, it is also the same governance structure for a startup that wants to raise a lot of outside capital without losing control.

Researchers based in China have collaborated to create a tool that can take natural language prompts and use them to produce proteins with specified functions. This could enable anyone to design proteins - if they have a lab that can sort the good designs from the failures. The model's success rate is not great, but it can still get the job done. The tool likely won't remove the need for biologists, but it could help accelerate research.

Reflect Orbital is a US-based startup that plans to erect thousands of giant mirrors in space to reflect sunlight onto the Earth during night-time. The Federal Communications Commission has approved the startup's proposal to launch an initial demonstration satellite next year. The test will reflect sunlight in 3-mile-wide beams in strengths from moonlight up to the brightness of the Sun at noon. While the technology could be helpful for search-and-rescue operations, among other applications, critics say that it threatens a harmful change to the separation of night and day, with a huge number of unresolved questions and significant potential impacts.

Datadog's 2026 toolkit provides instant access to resources for building, shipping, and observing AI-powered applications with confidence. Product breakdowns reveal how to achieve end-to-end delivery metrics + faster, cleaner pipelines. An eBook explores how to monitor LLM applications, agents, and retrieval pipelines in production. An On-Demand Webinar covers best practices for modern software delivery. Get the toolkit.

Hallmark is an anti-AI-slop design skill for Claude Code, Cursor, Codex, and OpenCode. It contains fifty themes and 755 openly licensed photos. Hallmark can create custom pages from prompts or real products, and users can redesign and audit pages. Each generated page is self-contained HTML and CSS.

The coding interview was never popular, and AI is likely going to make it go away for good. Impact is now less bounded by years of accumulated knowledge. Fresh eyes have long been considered powerful, as newcomers often see things old-timers miss. Look for candidates with prototypes that exhibit taste, empathy, and principled thinking.

Amazon has released a memo to staff saying that token usage and code generation should be considered with care and judgment. The company has not instituted caps on employees' token usage, but it has acknowledged that metrics like token usage, the rate of adoption of AI tools, and code changes or reviews remain deeply flawed ways of measuring AI usage, and that they can lead to higher costs and too much complexity. The company is now looking at how many launches it is doing and how many successful changes it is making in production to measure how AI is improving engineer productivity.

Ivan Vendrov worked at Anthropic between October 2022 and May 2023. This post shares the most salient parts of his experience. Anthropic has since grown into a massive organization.

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.

While 79% of developers give AI agents cloud access, 97% cited concerns about the security implications. Get the report

Broadcom's CEO agreed to protect banks from losses in case Anthropic defaulted on the loans.

Large language models allow developers to write small, purpose-built software.

You need to know just enough when something is wrong, and be willing and able to dive deep when needed.

Some users are concerned about exposing conversations to audiences far beyond their original context, and that their conversations may be used to train AI models.

AI can change what someone will pay you to do a task.

A 1 gigawatt facility loses money in 98% of simulated outcomes at today's build costs and rental rates.

[Consulting knowledge layer…]

Your team solves the same problems in Slack every week, but the answers never reach the docs your AI reads. So when the next person asks, the agent serves last quarter's version, and the cycle repeats.

Guru fixes the source before anyone asks:

✅ Turns resolved threads and tickets into drafts your team approves.

🗣️ Verifies/unverifies content on rules written in plain language.

🤖 Keeps stale content out of every agent's context.

🧠 Serves the verified layer to any AI tool over MCP.

Teams have seen: 80% less token use, 16% fewer repeat questions, and 96% of searches end with the AI answer.

Stop answering the same question twice.

DoorDash is looking to acquire rival Instacart in a deal that would value Instacart at around $17 billion. The deal could be finalized as soon as this week, but there is no guarantee it will come together. Instacart's core business involves online grocery shopping and convenience home delivery from retailers. The acquisition will add to DoorDash's food delivery services from retailers such as Kroger, Safeway, and Aldi.

Anthropic unveiled its IPO paperwork yesterday, putting it on track to go public as soon as the week before Thanksgiving. The 400-page prospectus tells a story of how the company could usher in a new era of progress that will transform how people code, conduct research, build financial models, and automate routine tasks. It also warned that AI could lead to the development of new bioweapons, destabilize the geopolitical order, and evolve beyond the ability of humans to understand and control. Anthropic plans to trade on the NASDAQ under the ticker 'ANT', with a valuation expected to land between $1.5 trillion and $2 trillion.

OpenAI has announced three new discoveries in mathematics, computer science, and physics. For the math result, which is about harmonic analysis, the AI lab will release the AI output, so researchers are free to organize their own efforts to explain the result. OpenAI will form an external group of collaborators to explain the computer science result, which is about complexity theory, and will make the AI output publicly available. The lab will also form an external group of collaborators for the physics result, which is about quantum field theory, but defer to them on the release of the materials.

Solar power will likely provide 10% of the world's power by the end of the year. Solar electricity is the fastest-growing source of new electricity in history as it has become unbelievably cheap. China is still the world's biggest solar power producer by far, but the installations in the US are skyrocketing. Even with tariffs, solar is still one of the cheapest sources of new power in the US, and it is getting built regardless of political lines.

Review the new AI accelerator offering, co-engineered with NVIDIA. The A5X integrates directly into Google Cloud's AI Hypercomputer, combining compute, storage, networking, and software into one cohesive system. Backed by the Google Virgo Network, it scales to nearly a million GPUs across multi-site deployments. Watch the video.

The Counting House pattern adds AI to systems that handle large volumes of human work, notably customer service departments. It can be used to control and improve costs, quality, and human ramp-up time using comparatively cheap AI models. A Counting House uses classic data science and engineering to put lower-tier models on hardened rails to support human workflows. It allows teams to measure and report how well models do against their human counterparts.

Claude writes about 90% of the code in Claude Code. This post explains how the team reviews the code at the volume its agents produce. The team moved most of the reading in the review stage to machines, optimized for human time and attention, and invested heavily in making it fast and easy to undo mistakes. This resulted in merged pull requests to Claude Code's codebase more than doubling while the share that had to be reverted fell by more than 40% between June and September.

Microsoft is extending voluntary buyout offers to 60,000 employees in Canada, India, the UK, and the US. The program is time-limited. It provides eligible employees with financial support and career planning services. Microsoft expects less than 30% of employees will accept the latest buyouts. The voluntary retirement program is part of a broader shift by Microsoft to alter its performance system and how it awards bonuses and stock options.

Blogging began to die when Google started aggregating the internet. Social media made it much easier to create and share content, but this caused the amount of content to increase by orders of magnitude. Algorithms started to matter, and then these sites started becoming personal entertainment sites, tuned for engagement. The latest series of security vulnerabilities unveiled by AI will further shrink the public internet.

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.

Keeping a high-quality bar at 1,000 PRs a week is hard because developers must know that each change is correct, safe, and worth merging without a human reading every line.

Work on JIT has stopped, but there might be time for an alternative PEP to be submitted before the council's deadline.

Anthropic's S-1 describes an AI company with an enviable business model, explosive growth, and increasing efficiency.

turbopuffer v3 will be the first and only search engine to adopt fully vectorized execution, an architecture more commonly associated with analytical databases, when it goes to production later this year.

Instagram would rather keep fans on its own platform than lose them to YouTube.

Nvidia already charges $4,500 per GPU for a one-year subscription to its AI enterprise platform - it is unclear whether the two prices cover the same products or licensing terms.

AI inference is getting expensive. But does it have to be? SambaNova powers the world's largest AI models with blazing-fast inference and industry-best economics. Get the performance you need without the price tag you don't. See what's possible.

Apple's iPhone Duo is a miniature tablet that can be folded up into a phone. Its main target audience is people who want to have a tablet everywhere they go. Commuters and business travelers who spend a lot of time on trains, buses, and planes will probably like the device. Everyone else may feel frustrated with what is otherwise an average smartphone.

High-frequency traders are now racing for compute. They are investing in the sector as it secures them access to greater processing capacity and faster chips. These firms primarily profit from tiny price differences as they move rapidly in and out of positions. As they increasingly rely on machine learning, their success is directly tied to their computing power. While the risks are significant, most high-speed traders see few options if they want to avoid becoming obsolete.

Researchers have developed an implantable neural chip that tracks communication across brain regions. It can record brain activity and extract biomarkers that describe the strength of brain waves and synchrony between signals from different brain regions. These signals can be decoded using a machine-learning algorithm to distinguish between cognitive and emotional states. The capability to extract brain-network biomarkers to trigger precisely timed stimulation has many applications across next-generation neurotechnology.

SpaceX has been granted access to restricted space tracking data by the US military. The access raises new national security questions as well as questions about whether the Trump administration is attempting to hand over traffic control in Earth's orbit to a donor. Elon Musk's Doge program fired many of the federal government's space traffic control staff in 2025, and the administration has twice attempted to financially gut the Traffic Coordination System for Space department. The moves have been widely viewed in the industry as an attempt by Musk to privatize space traffic control.

On November 5th, the Sentry team will be hosting a live workshop on session replays. You'll learn how to create custom React hooks, get user feedback with replays attached, and which filters and tags you should use to find the sessions you're looking for. Register free

AI can now one-shot many projects, allowing almost anyone to create software. However, it still struggles with many harder tasks. People who say that everything just works or that software is solved aren't working on the hard problems. They need to be more ambitious about the software they're building as AI models become more capable.

Context engineering is hard in unique ways because AI models aren't people. It helps to understand how models are built, optimized, and trained. Their unique attributes and construction create a unique set of failure modes to be reckoned with. This means context engineering will forever be a challenge with AI, no matter how good it gets.

This post tells the story of what it was like at Salesforce after it acquired Slack. The cultural gaps between the two companies were evident. Slack had a tight-knit culture where humans would help each other, whereas Salesforce was a much larger company with systems and processes for things to get done. Slack's cognitive overhead was nothing compared to the overbearing weight of the Salesforce monolith.

TSMC Arizona's first plant has been producing chips for two years. Its rate of output in Phoenix is comparable to that in Taiwan. TSMC sends engineers to Arizona on contracts of up to three years, with families moving from Taiwan to the site. Some of the families that moved in the first wave now want to build lives in the US, as they want to give their children a chance to integrate, for their kids' education, or because they prefer the American lifestyle, and there are legal pathways for doing so. This article looks at the impact the factory had on Phoenix, the employees who moved to America to work there, and the chip manufacturing industry in the US.

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.

We are trying really hard to make agents productive while we're not around, but that requires excellent planning.

The Jev hype cycle is a prime example of the manipulative marketing that we are all too familiar with at this point.

Cap 0.6.1 is the first version that doesn't have a web browser inside it.

Most people will no longer need to be significant producers of intelligence in order to become significant consumers of it.

It's strange that OpenAI and Anthropic cut their primary revenue sources at a time when they are losing billions and are planning to IPO soon.

We have to direct our attention more consciously as the future hinges on the choices we make now.

Your follow-ups are only as good as your meeting notes. Wispr Flow Notetaker captures the meeting, then takes a second pass over the audio so names and terms come out right. Less cleanup, better notes.

Paying for an AI meeting notetaker? If your plan qualifies, switch to Wispr Flow and we cover the time left on it with Wispr credit, plus 3 bonus months. Up to 12 months free. Voice dictation in every app is included.

Enter your email and 3 months start today. Send your receipt to claim the rest.

Claim your free months | See if your plan qualifies

Three OpenAI safety researchers were fired after security officers discovered Signal chats with disappearing messages turned on on their personal devices. OpenAI claimed the researchers were fired for violating clear policies for handling sensitive information and not for raising safety concerns or speaking out. The researchers had spent recent months at the company pushing for more industry collaboration on safety. They claim their firings will serve as an excuse for OpenAI to close a channel to independent third-party insight.

The Alexa Tablet lineup offers a more open approach to software than Amazon's previous Fire brand that lets consumers use the devices like any other Android tablet. Users can download apps from the Google Play Store and access popular services like Gmail and YouTube. Amazon previously restricted access to its own app marketplace. The 12-inch Alexa Tablet Pro is well constructed and has a smooth 12-hertz display, making it competitive with Apple's iPads.

SpaceX is in talks with Crown Castle Inc., one of the largest owners of cellular towers in the US, as part of its plans to secure a minimum of 80,000 cell sites across the country. Crown Castle is likely also in talks with Amazon and other satellite-based telecom providers regarding leasing deals. Elon Musk has openly discussed creating a standalone wireless carrier from scratch. SpaceX recently acquired low-band spectrum and has received approval to launch 15,000 satellites to provide mobile service directly to phones.

Moderna and Merck's personalized mRNA vaccines against melanoma demonstrate how mRNA technology has the potential to fundamentally challenge how cancer is treated. The vaccine's success offers a new paradigm for cancer therapies. The companies are already engaging with regulatory authorities to get the vaccine approved.

More context, better output from Cursor, Claude or ChatGPT. Wispr Flow lets you say it instead: explain the bug and the edge cases out loud, and it types clean, formatted text in any app. 4x faster than typing, on Mac, Windows, iPhone and Android. Try Wispr Flow free

We need ways to prove that code is free from bugs. Formal software verification, which could provide this guarantee, has always required years of expert effort, but LLMs can now produce formal proofs at large scale. This enables new software development workflows that include proven security guarantees. Formal verification will likely become an essential part of secure software development as model capabilities continue to improve.

Anthropic employs around 3,500 staff, but its coding output is closer to that of a company with 35,000 to 70,000 staff. Big tech companies like Google and Meta moved off Git a long time ago because it did not scale for them. Both built their own custom solutions, just like Anthropic did with Scion. This post provides some details about how Anthropic's custom source control system works.

Mathematics seems to be a field where the outputs are celebrated less as examples of subjective expression and more as advancements in knowledge and understanding. A mathematician's work may be informed by their experience, but that's not what gives their contributions value. Mathematical knowledge remains objectively true whether it is observed by humans or sea sponges or computer programs.

Ondřej Surý is one of the volunteers who helps maintain and automate patches for PHP. He alone is responsible for most of the world's distribution of PHP. His work, along with just a handful of others, has allowed countless websites and organizations to fix the recently revealed PHP vulnerabilities quickly. He's become arguably the unsung hero of the entire internet, despite the fact that the only compensation he receives for his work is from donations. This article features an interview with Surý where he discusses the unfolding AI-cybersecurity crisis, how the internet's infrastructure is maintained by small crews of volunteers, and how it feels to save the web.

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.

If your plan qualifies, Wispr Flow covers the time left on it with credit, plus 3 bonus months. Meeting notes and dictation, one subscription. Claim your free months

Most programs can't be meaningfully sped up by rewriting them in C or Rust.

The US has become increasingly reliant on foreign nations for dogs that play an important role in the US military and law enforcement.

Supervisory engineering work is a different job from software engineering and so requires a different system around it.

Finance companies in the AI era need to be built around continuous decisions, portable assets, and delegated distribution.

We are the ones who get to decide what to do with all of the productivity that AI creates.

JPEG XL is the end-game universal image format because it has excellent compression efficiency, a universal feature set, wide color depth, gamut, and native HDR, lossless JPEG transcoding, high-speed parallel software encoding and decoding, and an open-source, permissive license.