Orkes Shift '26 is one day on running agentic systems in production: durable execution, retries, distributed failures, and human approvals. October 15, Convene Union Square, San Francisco. Early bird pricing ends soon and capacity is limited.The agenda:

• “Rebuilding a Monolith: Choreography to Orchestration”
• “Designing a Focused Agent Runtime on Conductor”
• “Stop Babysitting Your AI Coding Agents”
• Speakers from LinkedIn, Bumble, and the Orkes founders, plus a fireside and happy hour

Conductor runs 1B+ executions a day at Netflix, Tesla, and LinkedIn.Register here →

Anthropic's CEO recently released a 3,800-word essay calling for a global slowdown of AI development. He said that while the technology offers many benefits, it is advancing too quickly for researchers to continue safely. AI leaders are preparing for a day when an AI system can grow more advanced without the aid of human researchers. Some researchers believe that achieving recursive self-improvement will lead to AI systems spinning out of control.

OpenAI CEO Sam Altman has confirmed that the company will not be going public this year. He said it would be an ill-advised moment to go public due to current safety concerns. OpenAI had hired bankers and lawyers with the goal of going public this year. The company is now leaning toward 2027 due to the volatility of tech stocks and its own financial challenges.

Construction is labor-intensive. There has, unsurprisingly, long been interest in automating the process to reduce the amount of labor required. However, automation has only been successful in tasks that can be done off-site in factories and on-site tasks that are extremely factory-like. Almost nothing outside of these two categories has succeeded.

The push by AI companies to solve mathematical problems as a benchmark is detrimental to the science of mathematics. The goals of AI companies and the mathematical community are severely misaligned. This is a symptom of broader alignment issues that impact other scientific and creative professions, as well as the whole of society.

Agents can generate code. Getting it right for your system is the hard part. More MCPs solve access but not understanding. Join us for a FREE webinar on Sep 23 to see how leading teams save time and tokens by giving their agents a context layer.

Forward-deployed engineers have the hottest job in AI right now. While labs, startups, and PE firms are all hiring engineers to sit inside their customers' operations and solve their problems, almost none of them agree on what those engineers are supposed to accomplish. This article looks at what works and what fails and tries to predict where the industry is heading. It covers misunderstandings about the role, its history, what the goal of the role is, and more.

Large language models are able to write almost perfectly correct code. It is pretty straightforward to get a model to generate code and then let that code be checked by hidden tests. Checking the sloppiness of this code often requires human intuition and time, and is an extremely difficult task in general. This post illustrates why this is by going through possible ways of measuring slop.

While AI has the potential to produce a long line of technological miracles, like many technologies before it, it brings risk. These risks include losing control of AI systems, the misuse of the technology for cyberattacks and bioterrorism, and serious economic disruption. A race to the bottom spurred by commercial incentives makes these risks more acute. Anthropic's CEO has called for the industry to prioritize caution over speed and prudence over profit. Addressing these risks will require pacing the rate of capabilities advancement so that risk prevention has time to keep up.

The company behind Roblox announced several new features at its annual Roblox Developer Conference, including new game-creation tools, expanded NPC capabilities, and the ability to make games available across platforms. The company is also launching a dedicated card and wallet for creators. Roblox is rolling out its generative AI game creation feature more broadly to Serbia and Singapore. It is also gaining desktop access, and later this year, it will get a new prompt-based scene-generating feature.

Access models from 80+ providers with better prices, automatic routing and fallbacks, and no subscriptions. Build with the right model for every request. See how it works.

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.

Most prompts are bad because people only add to them over time rather than removing unnecessary instructions.

The future will see more software, more people making it, and none of them typing.

The speed of the developer experience will matter a lot more when generating tokens stops becoming the bottleneck.

Tesla recently made a post that simply wrote, 'Go for launch,' with a graphic showing a '10.1' date.

Humans who want to use AI for interpersonal interactions will need to work hard to compensate for the loss of trust it engenders.

Big prompts are brittle and hard to debug and evaluate, but they are the best choice in some cases.

Devs finally have a challenger to GitHub code hosting with Origin.

Origin was built on Buildkite.

Buildkite is a launch partner for this new code ecosystem. Origin holds the code. Buildkite runs the build. Vercel ships it. Three layers, one path from idea to production.

We also run CI for OpenAI, Anthropic, Mistral, xAI, Airbnb, Nvidia, Canva, Uber, Meta, Pierre, Bun and more.

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

Start with our 30-day all-access trial.Get building →

OpenAI's Agents API is now in public beta. The API brings the same harness and infrastructure that powers Codex to developers through a simple and flexible API. It allows developers to create useful agents with a powerful harness that manages context, uses tools efficiently, and coordinates subagents. The API provides infrastructure to keep agents running reliably for days, and environments where agents can work with files, run code, and save intermediate results.

Anthropic claims it has disrupted several plots to use its AI models to conduct research that could have helped develop biological weapons. The company says it still can't determine whether a query serves a legitimate or nefarious purpose, so it erred on the side of caution as the consequences of mistakes could be severe. Anthropic detected use from companies from blocked regions and parties that otherwise tried to obfuscate their purpose.

Unitree CEO Wang Xingxing recently unveiled details about his company's IPO in an interview that covers his business principles, Unitree's managerial structure, and how the company is changing its employees' lives. Unitree listed on the STAR Market on August 19, with its market capitalization briefly reaching 440 billion yuan after its IPO. Wang was briefly the richest person in China born in the 1990s.

Analysts predict that AI will soon deliver double-digit growth in advanced economies. Experts say it could push growth to something like 10% to 15% per year. The decade ahead could see economic growth of 30% per year and beyond. Predictions of double-digit growth are common among AI insiders.

Ahrefs Brand Radar tracks your brand across ChatGPT, AI Overviews, Gemini, Perplexity and Copilot — from 471M+ real search-backed prompts. Benchmark competitors, see what AI cites, add your own prompts.

Explore Brand Radar

This post follows the journey of a PostgreSQL query through all of the layers of a complex sharded system. It aims to help readers understand what goes into making a large-scale sharded PostgreSQL deployment appear to be a single server even when spanning thousands of servers. A real database has hundreds of tables - the example in the post keeps the schema simple with two tables spread across four shards.

Companies are experimenting with cheaper open models on inference providers. There is a trend toward more investment in model routing to route simpler requests to cheaper models. The industry is seeing more knowledge-sharing sessions on how to use AI models cost-effectively. More and more companies are setting per-developer limits on AI usage.

Centralized router platforms can offer customers better economics than local deployments. Outcomes pricing belongs to the infrastructure layer. The model layer could offer outcomes pricing, but it will have the same problems as the app layer. Router layer pricing is most politically aligned with customers, who just want to sell more compute and don't care which model is used.

The 'next thing' is always simpler than you think and very obvious in hindsight. OpenAI's addition of Computer Use is deceptively simple and obvious in hindsight. 'Computer use' is a very generic, unspecified thing. It requires general intelligence across many domains to get better at it.

Your vendor says "OWASP LLM Top 10 covered." That's a checklist, not a defense. Seven red flags every developer should catch before signing.

Astra will change the world one way or another.

The people building AI earnestly believe that it could kill us all by the end of the decade.

AI can deidentify you better than many adversarial bot programs.

Most changes that would have made agents better at codebases make agents better too.

Google has six months to implement the technical measures and finalize the processes for data access.

A storage solution is only part of what a team needs to collaborate around code.

Lightfield is the CRM that updates itself. Yesterday, it announced a $47M Series A led by Andreessen Horowitz.

Here is why thousands of companies switched to Lightfield: it captures every email, call, and meeting automatically. No manual data entry, ever.

Ask "Which deals have gone quiet this month?" or "Prep me for my call with Acme” and it answers from your actual conversations, not fields you never filled in.

Teach Lightfield how you sell and watch it go to work for you.

Skeptical? Grab time with the team to see it for yourself

Apple's first foldable iPhone, the iPhone Duo, starts at $1,999. The device expands to a 7.6-inch screen when unfolded, and a 5.4-inch screen when used as a conventional handset. It comes in two colors: Star White and Night Sky. More details on the device, as well as Apple's other recent releases, are available in the article.

The Apple Watch Series 12 and Ultra models have new Audio Intelligence tools that use ambient listening to take high-level notes of wearers' conversations throughout the day. The devices will also have a Live Recap function that can show a transcription of what was said in a live conversation in the last 15 seconds and be able to listen to important sounds, such as a crying baby, doorbells, or sirens. These features will not create or store recordings, and Apple will not be able to access the raw audio. They will be available to test in English later this year. The improved Siri is being rolled out on watchOS 27 in beta on September 14.

Modern cancer-vaccine design involves identifying mutations, predicting presentation, predicting immunogenicity, ranking targets, and designing RNA. However, biology is complicated, and whether a target makes a good vaccine candidate depends on all kinds of information. This makes cancer vaccines an unusually direct test of whether AI models can take what it has learned from biology and apply that knowledge to design a medicine for one particular human being.

It is unknown yet whether AI will lead to unprecedented growth or widespread unemployment. Anthropic's Economics team built a model to understand how AI might affect jobs, growth, and unemployment in the US in the coming years. The model found that AI will lead to adverse impacts on wages and job prospects for knowledge workers in scenarios where growth is faster than anything in economic history, and in those scenarios, the challenge will be making sure wealth is broadly shared. The answer will depend on how AI affects all the tasks that make up the economy, the new tasks it creates, and how fast AI takes on this work.

When approval fatigue turns sign-offs into rubber stamps, you need more than good judgment to keep AI secure. Gray Swan's eBook lays out 5 governance implications for AI security, along with 3 questions every board should be asking right now. Get your copy

PostHog's Replay Vision lets models view session recordings. This post looks at how PostHog built the tool and how it implemented a method for the model to actually see replays. It covers how PostHog built a custom rasterizer that runs hundreds of concurrent jobs across many pods. The custom rasterizer has so far rasterized more than 370 years of recordings into roughly 3.5 million videos since it was deployed in production in March.

Tailwind is joining Shopify, one of the first companies operating at scale to see the potential of Tailwind CSS and start building with it. Shopify has invested in making sure Tailwind is actively maintained and will continue to provide a stable long-term home for the project. Nothing will change with Tailwind CSS or any of Tailwind's other open source projects, but the project will no longer try to grow its business. All existing customers will maintain access to their products, but Tailwind has closed signups for new customers.

The price of AI is coming down. However, volume growth may not be large enough to counteract the effective decline in prices. The models driving volume increases are relatively cheap, standard models. These trends do not appear to be driven by the adoption of open-source models or Chinese models.

Automattic's board has put founder and CEO Matt Mullenweg on a leave of absence against his will. The company's Chief Financial Officer, Mark Davies, had conspired with other board members to put Mullenweg on a paid leave of absence. Davies is now acting as interim CEO. Mullenweg remains the leader of the WordPress project, and the company's teams, priorities, and work will continue as planned. It is unclear why the board took this action. Mullenweg remains on the board of directors.

The Granola MCP integration connects your meeting notes to your AI agent. Have your agent review meetings, update your CRM, and organize tasks in Linear. Free month with code: TLDR1MO

A system that improves itself needs guardrails and a sandbox.

Purchasers of open source work should be able to check whether a vendor contributes.

There are massive implications of a race towards superintelligence, and not enough of Congress is focusing on them.

While GPT-6 Astra and Claude Fable 5.1 seem to have the same rates, Fable's cache read is four times cheaper.

View Transitions and Fragment Refs are now stable in React 19.3.

OpenAI heard that some Millennium Prize problems had been solved using large language models and saw it as an opportunity to demonstrate the power of its latest model without regard for the teams who had been working on the problem already.

Rolling out an agentic workflow is the easy part. But after you launch, prompt injection attacks and ordinary errors can pile up without you noticing.

Even an operationally performant tool can accumulate negative sentiment or irrelevant response patterns. And as your app becomes more complex, it gets harder to pinpoint root cause.

This Datadog best practices guide lays out how to:

• Monitor LLM workflows from end-to-end
• Detect and mitigate security risks
• Ensure reliable, high-quality outputs

Get your framework for complete agentic observability. Grab your copy

Meta's new Muse AI agent can act as a personal digital assistant, autonomously using software apps and websites on behalf of users. It can be instructed to send emails, book travel reservations, make online purchases, and more through an app or WhatsApp. Muse AI can connect to Meta's other apps to learn more about its user. It can also link to third-party apps like Spotify, Ticketmaster, Shopify, Gmail, and OpenTable. The Muse app is free but has limits on usage, which users can pay $20 or $100 a month to increase.

Apple's secretive core technology development teams had already been studying foldable phones well before Samsung released the first Galaxy Fold in 2019. The company is set to unveil its first foldable phone today. The phone will feature a roughly 7.8-inch interior display considerably wider than those on many Android foldables. Apple was targeting a price of $1,999 early in development, but it has more recently been discussing pricing the foldable product as high as $2,199.

OpenAI says its newest AI technology has solved one of the Millennium Problems. The Millennium Problems are the most heavily researched in the field of mathematics. OpenAI's model solved the Navier–Stokes existence and smoothness problem in just 88 hours. The company's proof claims to have defined a situation in which the Navier-Stokes equations completely break down. This implies that the laws of physics could theoretically break down under certain conditions.

The AlphaGenome Atlas is an online repository of precomputed predictions made using Google DeepMind's AlphaGenome model. It allows scientists to access the benefits of the model without having to write code or run the model themselves. The atlas includes a single-number impact score to show at a glance if a variant is likely to be meaningful. The atlas could greatly accelerate work in fundamental biology, disease research, and treatment development.

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.

ChatGPT Images 2.5 is a new state-of-the-art image model that generates sharper details with more precise editing. Image generation latency has been reduced by up to 50% compared with Images 2.0. Images 2.5 is available for all ChatGPT, ChatGPT Work, and Codex users across desktop, mobile, and web. OpenAI has introduced a new feature called Sketch in ChatGPT that lets users draw directly in the app as a reference.

Teams that operate software at scale need to bake in some real time to deal with the CPU shortage. Most software teams have never capacity-planned CPUs. General-purpose compute is now something many teams need to commit to ahead of time. Teams need to start forecasting and planning around this.

Working out whether a question is actually worth working on is a lengthy, deliberate, and subjective process. Being aware of the difficulty landscape in a field is crucial in making such determinations. Every new advance in mathematics reduces the difficulty of solving problems, which reduces the difficulty landscape of the field. While there are an infinite number of possible problems to solve in mathematics, the vast majority of these problems are not worth focusing on. The indiscriminate use of powerful solution-extraction tools achieves the immediate short-term goal of solving problems at hand at the cost of sustaining the ecosystem for the next wave of progress, or in understanding the progress already obtained.

Apple's success with Siri AI will be largely dependent on third-party developers. However, these developers have a lot at stake, as while their apps may become more useful on Apple devices through working with Siri AI, enabling the assistant could risk sidelining their own assistants. Several staple apps are still not working with an early version of Siri AI available to app developers, and it is unclear whether they will. If Siri AI rolls out this year with only support from some apps and not others, it could create a confusing experience for users and turn them off from using the assistant.

Choosing an ERP shouldn't be guesswork. This free Software Advice guide compares NetSuite, Fishbowl, and Global Shop Solutions side by side — pricing, features, and real user reviews — so you pick the right fit. Get your free guide

Google's fly connectome contains over 166,000 neurons and 125 million synaptic connections, providing scientists with a fundamental resource for studying how the brain works.

Sonera is working on sensing technology that non-invasively measures magnetic fields generated by the brain and body to analyze neural data without direct skin contact.

Block has submitted an application to federal regulators to establish a federally regulated, uninsured national trust bank called Builders Bank & Trust.

Mechanistic interpretability is the science of reading an AI's mind.

WeWorm is a zero-click attack that can compromise WeChat accounts without users doing anything and gain access to messages, calls, and accounts.

Not on purpose! But stages are always optimistic and close dates go stale. People are busy...

Lightfield is the agentic CRM that builds the record from the conversations themselves, capturing every email, call, and meeting as it happens. So the questions your CRM could never answer, this one can.

Ask "Why do we keep losing to Competitor X, with quotes from the actual calls." Get real answers and see the evidence.

5,000+ companies signed up in the ten months since launch. Book a demo

ByteDance is preparing to launch a real-time spatial video AI model. The model will be able to create interactive virtual worlds. It will also be able to generate virtual worlds that respond to Pico VR headset users' voices and movements. ByteDance is planning a launch as soon as next month.

OpenAI appears to be implementing a new feature that helps ChatGPT write in users' voices by referencing examples from their connected apps. Some users have had the feature for weeks, but the company has yet to officially announce the tool. OpenAI often tests new ideas with select users ahead of launch. The article includes a screenshot of the feature interface.

XPeng has started production for its IRON robot. Its production line runs with more than 80% of its core processes automated. The company's robot went viral last year as its walk was so smooth that people accused XPeng of hiding a person in a costume. XPeng is targeting mass production by the end of the year.

Rentosertib is a drug created with the assistance of AI. The AI originally indicated that the drug could help treat patients suffering from a chronic lung disease, but data from clinical trials show it could also slow the aging process. The drug reduced the biological markers of age as measured by six 'aging clocks', a different kind of AI technology designed to predict a person's morbidity and mortality. While the drug shows promise, it is still years away from regulatory approval, even for use in sick patients.

What are your AI agents doing? What do they have access to? Teleport gives you answers and helps you set boundaries to contain and control them. Monitor and audit agent behavior to better understand the risks related to AI in your environment.

Join Teleport September 17th to learn more.

It's easier than ever to hit a particular quality bar by having coding agents use effective test techniques. However, software quality seems to be getting worse. This indicates that whatever defaults developers are using might not work very well. This study tests whether simple instructions to agents to use particular techniques or libraries improve implementation correctness. It found that agents tend to either just write the tests they would normally write, but inside a framework for a different type of test technique, or they'll use a technique superficially but not really do the things that get the value out of the technique.

Over the next few years, offensive capability will become abundant. More people will be able to inspect every valuable project and exploit vulnerabilities. Handling critical bugs in a timely manner will become the most challenging and important task for defenders. Having good security infrastructure and integration will be a must. AI-assisted formalization will lower some of the cost, but writing the right specification and connecting proofs to production code will remain challenging.

TLDR is hiring its first PM to help build the agent-first operating layer used across the company. We're looking for a builder who has shipped real products/systems with LLMs. Click here to learn more.

The monotony of work can make it hard to keep a clear head. Spend your vacations doing something different for your brain. Slow down and double down on new habits you want to grow. Spend time in nature, socialize, play games, engage with hobbies, and you'll notice you might start being less intellectually lazy during your day-to-day activities.

Artificial General Intelligence (AGI), as the term is commonly used, points to nothing really tangible or concrete and points us at the wrong thing. Recent announcements about AGI are just companies saying they've nearly arrived at a destination they defined themselves. Instead of trying to define AGI, the industry should focus on making a long series of specific things reliable enough to trust, and let the destination take care of itself.

Comparing how system prompts develop over time demonstrates how model quirks and product design are moving targets.

Writing simple programs requires a good mental model of the program as well as good taste.

This article details the Tesla Cybercab's dimensions, mechanical packaging, and custom thermal hardware.

Code that is well-structured is easier to maintain and also easier for agents to work with.

Huawei's new tri-folding phone, the Mate XT 2, opens to a 10.2-inch display.

Someone working outside an experiment can still be affected by what happens inside it.

Agents are smarter when they can search the web. But Google doesn't offer an API for web search results... so how are AI companies doing it?

Meet the worst-kept secret in the AI industry: the SerpApi web search API:

👉 Scrape Google and other search engines (including AI overviews, Maps, Local, Shopping, and Knowledge Graph)

👉 Integrate with a dead-simple GET request that any agent can make.

👉 Used by NVIDIA, Adobe, Shopify, and even the United Nations.

Right now, TLDR readersget 3 months at 50% off, or start with 250 free credits/month.

Claim your savings

OpenAI believes that Artificial General Intelligence must be democratically governed for it to benefit all humanity. This can only happen through an informed public debate on the capabilities, risks, and safeguards of highly capable AI systems. The company has released details of its progress towards Recursive Self-Improvement (RSI) in recent months. It plans to continue to be transparent about its progress with RSI and believes that other companies should be required to publicly track their progress as well.

Tim Cook remains a powerful center of gravity at Apple. He will play an especially crucial role in certain areas, such as the company's dealings with the US and China, and its handling of the memory shortage. Cook will get a base salary of $2 million, twice the typical base salary of senior vice presidents and members of the C-suite. He will also receive cash bonuses and stock units that vest over time. The salary clearly shows that Cook's role will be very active.

A domain becomes playable when enough of its complexity can be compressed into a state where there is some reasonably stable repertoire of actions, interactions can be repeated, outcomes provide feedback about better and worse play, and the environment is stationary enough that lessons learned yesterday remain useful tomorrow. None of these conditions need to be perfect, as playability is a matter of degree. Far more of reality than we have previously expected can be turned into a game. Once something can be turned into a game, then it's possible to create a system that can beat it.

Atoms, founded by Uber founder Travis Kalanick, is reportedly preparing for a hiring and acquisition spree. The company is also in discussions with Uber about how Uber can use Atoms' robotaxi technology. Uber has invested $100 million in Atoms. Kalanick has been cagey about his plans for the startup, but the recent moves could make it a major player in the autonomous vehicle industry.

GLM-5.3 lifts Terminal-Bench 3.0 from GLM-5.2's 4.6% to 28.3%, and Flash is smaller and often enough. Both on Baseten Model APIs. Artificial Analysis and Baseten discuss benchmarking each against your tasks and serving them in production. Tuesday, 11 AM PT.

Save your seat

infra is an infrastructure-as-code tool that does what Terraform does. Users declare what resources they want, and infra looks at what their cloud accounts actually contain and reconciles the difference. This post details how its developer created the tool in two weekends in Lean 4, the problems it solves, and how the developer solved these problems.

Security is always a productivity-risk tradeoff. Not giving access to agents can slow things down significantly. Frontier models are getting better and better at distinguishing real from malicious instructions in the content they read. We may end up trusting agents with access more than humans as models improve.

AI models are now capable of taking a request from Slack and then building a feature from end-to-end, with humans just having to click 'merge' or 'deploy' at the end. There's not really much left for humans to do. Humans are still better at absorbing, parsing, and sharing context with each other. However, once that can be automated, the loop will finally be closed, and there will be nothing left to do but to pay the bill.

TLDR is hiring its first PM to help build the agent-first operating layer used across the company. We're looking for a builder who has shipped real products/systems with LLMs. Click here to learn more.

OpenAI's president and co-founder, Greg Brockman, dropped out of college in 2010 to join Stripe. He eventually rose to become the company's CTO. He left in 2015 and co-founded OpenAI, where he served as CTO, and after a few years, he is now the company's president. This post features an interview with Brockman recorded before the announcement of Astra. It covers Brockman's background, the ChatGPT launch, OpenAI's place in the value chain, the company's commitment to alignment, the Hugging Face incident, and more.

Researchers built a neural network from a map of a fly brain and then trained it to recognize emotion in human voices.

Omarchy is an opinionated Linux distribution built for AI agents.

The iPhone Ultra's production is slow due to how difficult the device is to build.

Demand for AI semiconductors is likely to surge with agent computer tool use, just as it did with the chatbot, reasoning, and agentic coding demand waves.

Apple publicly exposes the Apple Neural Engine through Core ML, while the compiler, program format, driver, firmware, and much of the machinery underneath it remains undocumented.

Real-world tests show that the Cybercab has competitive pricing, but currently has a much longer wait time compared to competitors.