Someone goes over budget. A receipt disappears. You lose time figuring out who spent what and why.

Mercury Spend replaces that system. It's expense management built directly into Mercury* — cards, budgets, and reimbursements in the same place as your business banking. Set spending limits per person, team, or card and let the policies enforce themselves. Receipts scan in automatically from Gmail, or arrive via photo upload or text. Cards with overdue tasks auto-lock until they're resolved.

No chasing. No manual review. No end-of-month scramble.

Explore Mercury Spend

Stripe has finalized an agreement to acquire OpenRouter, a startup that helps companies switch between AI models. The sale price of more than $7 billion could change and the discussions are not public. OpenRouter's rise highlights the industry's growing scrutiny on AI costs. The acquisition will give Stripe a stronger footing in the fast-growing AI sector.

Anthropic says it will not release an internal model it claims to be more powerful than Mythos and that it will not slow development broadly. The company believes that the risks of the most serious harms from its models are still low. While the company is seeing signs of acceleration in its models' ability to conduct automated research and development, it appears to be signaling that it is getting harder to understand the capabilities and risks of its own models. OpenAI says it is slowing down the release of an upcoming model called Astra because it can't rule out critical cyber capabilities.

Far-UVC 222 has been shown to efficiently inactivate airborne pathogens in a room-sized chamber. While the technology is unlikely to prevent the average common cold, mostly spread through close, extended contact, it could be helpful in blunting a future respiratory pandemic where more of the spread happens at a range through shared air. It is functionally equivalent to an extremely strong air purifier. The biggest bottleneck to the technology's adoption currently is awareness.

Heart Aerospace's X1 electric aircraft has a 106-foot wingspan and weighs more than 25,000 pounds. It flew for 27 minutes on battery power alone and reached an altitude of 1,100 feet during its recent maiden flight. Heart plans to bring a hybrid-electric plane into commercial service by 2031. The company's technology promises lower maintenance costs and insulation from volatile jet fuel prices for airlines.

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 most precious commodity is attention, ask your agents for the design after they've built the thing. Let the agent spend extra cycles in the lab, and get one layer deeper on the design. A plan that's been prototyped is a better plan.

The idea of a zero-knowledge proof (ZKP) is that there are two parties: the prover and the verifier. The prover asserts that it has a solution to a (generally NP-complete) problem. The prover can convince the verifier of this without sharing the actual solution to the problem. While they are used to secure some cryptocurrencies, that is not their only use case. This article looks at how ZKPs work and how they can be applied to other problems like age verification.

Several companies are attempting to build the synthetic superstars of tomorrow. AI imaging tools have enabled anyone with a modest budget and a sliver of vision to build their own photorealistic avatars. The availability of virtual talent has unsettled many communities of human performers. While AI influencers likely won't replace human influencers fully, they'll likely become more prevalent as these tools become more and more reliable.

TLDR is hiring a GTM Engineer to join our Applied AI team and own our AI-native GTM stack. We're looking for someone comfortable building AI agents and working with HubSpot. Click here to learn more!

Anthropic employee Jarred Sumner used the Claude app on his phone to try to solve the infamous Riemann hypothesis. The model didn't succeed, but it made a related finding that one Stanford number theorist has called the most impressive result that AI has produced in math so far. Sumner's formal mathematical education ended after just one semester of high-school geometry, and he identifies as very much not a mathematician. Most of the prompting he gave to the model was variations of 'keep going' and 'believe in yourself'.

OpenAI still can't define "AI agent." Business owners skipped the debate and hired Viktor, an AI employee in Slack & Teams. Dashboards. Apps. Tasks... Start free →

Z.ai released GLM-5.3 on Friday while holding back the model's downloadable weights and gating its most sensitive cybersecurity functions.

In a complex system, you can basically find as many bugs as you ask the agents for.

A machine can nail the task but completely miss the job.

The benchmarks are answering the wrong question.

This study introduces three tasks for AI agents that allow researchers to evaluate efficiency, collaboration, and taste, providing a proxy for how well the agents perform as tools in real coding situations.

Code is the byproduct of understanding.

Tines 3B gives every team the power to build with AI, and gives IT the control to run and monitor that work safely.

With Tines 3B:

• Turn ideas into production-ready apps, agents, and automations in any stack.

• Keep data secure at every step and ship workflows that improve over time.
• Get end-to-end visibility on spend, workflow performance, and value.

Get started today >

On August 19th, hear from Headspace on how they're governing AI at scale with Tines 3B. Register now.

SpaceX has been hunting for a spectrum that works well in cities and dense areas for its wireless network. It is considering buying competitors to acquire the spectrum or competing at a government auction set for next year. The company has expressed interest in building its own terrestrial networks and has demonstrated a prototype mobile handset to some investors. Entering into the mobile market is capital intensive, so SpaceX could ultimately choose not to follow through on these plans.

OpenAI's annualized recurring revenue in July exceeded the company's entire second quarter. The growth was driven by the release of the company's GPT-5.6 series of models and ChatGPT Work, as well as growing adoption of Codex. OpenAI is under pressure to justify its $852 billion valuation as it prepares for its IPO. It has been racing to bring in new users to generate the revenue needed to help support its infrastructure spending plans.

DoorDash has received permission from the FAA to deliver packages by drone at altitudes below 400 feet. The drones will be built in the company's in-house robotics unit, DoorDash Labs. DoorDash Air aims to cater to any merchant, anywhere. The company will release more information later this year on its drones and how its drone program will fit into DoorDash's broader ecosystem.

China has accelerated its deployment of civilian and military hardware into orbit. The country's growing constellations of satellites could help it track events on Earth in real time and in greater detail than ever before. This has raised concerns among US officials, who see China's growing array of space-based capabilities as a quest to displace the US as the pre-eminent space power. Western analysts say that China is leading the world in research in sensors and satellite positioning technologies.

Ever feel like you're shipping AI in slow-motion? This Datadog kit includes an AI monitoring eBook, product briefs, and an on-demand webinar on software delivery best practices. Download it to get up to speed on delivering observable AI. Grab your Developer Toolkit for the AI Era

Prompts are dependencies. Changing them can shift downstream behavior with no warning, no failing build, and no obvious signal in the request pipeline. Treat prompts exactly like any other artifact that can change production behavior. Test them, give them a gate, and default that gate to blocking until proven otherwise.

AI labs are moving faster than ever. However, capability is only half of the picture. Optimizing the cost per successful task makes models more affordable and less costly to serve. This article discusses techniques adopted in frontier labs to make API applications more efficient. It covers what happens when requests are sent to an AI agent, how the harness layer cuts repeated work, and more.

A new class of entrepreneurs is launching and running new million-dollar companies on their own. AI's ability to handle various administrative tasks makes it useful for launching solo businesses in many fields. The effect of one-person businesses on the labor market remains to be seen. While many people are worried AI will limit employment opportunities, these solo operators show how the technology can also open new doors.

500,000 people have already signed up for TLDR Hardware, our new twice-weekly newsletter covering chips, robotics, energy, and devices. If you work in hardware and want to help curate it, send your LinkedIn or resume to [email protected]!

Labs do not want to spend greater and greater shares of compute on inference. The point of inference revenue is to convince investors to spend more money to buy more compute to train bigger models. Spending too much on compute on inference signals that progress has stalled because it's not worth investing more in training. Labs believe that their models will continue to get better, so this only means that compute will get more expensive.

WhiteFiber validated 111.2 Tbps across 83km of dark fiber, within 8% of the physical speed of light. Two geographically separated sites. One logical AI supercluster. See the data

The data center boom is pulling skilled trade workers into remote locations with some of the highest pay and bonuses the industry has ever seen.

numbat provides endpoint visibility into AI agent activity with local detection, optional pre-action blocking, and forensic reconstruction.

It may be awkward and weird and strange, but it's usually worth the shot to send a cold email.

Open-weight models are already being built into critical systems in the US.

Chinese founders face more pressure and liability, which results in more competent and ruthless companies, but also creates incentives for more froth and fraud.

There is a consistent pattern of OpenAI's alignment strategies working less well than Anthropic's.

Apple is preparing a new push into the smart home category with a wave of home products. It will start with a hub device built around the new Siri AI assistant. Apple is also preparing to release a new TV set-top box and a refreshed HomePod mini. All three devices are nearly ready to launch. Apple is also working on a higher-end, robotic version of the home hub and an advanced in-home security camera.

Claude Mythos Preview found flaws in a watered-down version of an algorithm for Advanced Encryption Standard (AES), a ubiquitous protocol that protects web traffic, wireless networks, data storage, and many other technologies. Anthropic claims the assault was 200 to 1,000 times faster than what previous human research had managed to do. The attack was devised almost entirely autonomously - the model needed aid as it originally believed the problem was impossible. Claude Mythos Preview worked on the problem for about a week before engineering its attack, and two human researchers worked for nearly a month to verify that the method appeared correct.

Tau's humanoid service is now available in San Francisco. Access is initially invite-only. The service costs $30 per hour. Each humanoid will be jointly controlled by a human operator and AI. Video of the robot in action is available in the article.

Organoids are clumps of human cells that show some type of organ-specific function. In theory, they seem like a good tool for drug discovery, but the reality is that the technology is still not reliable enough. There are many people working to fix the issues with organoids, but they are still methods in progress. Organoids have not dramatically altered medicine yet, but they have already been used for some unarguably useful things.

Every token in the orchestrator's context is competing for its attention. The real value of a subagent is what it keeps out of that context. Subagents should be treated as a tool for protecting the orchestrator's working memory, offloading reasoning it doesn't need to hold onto. Doing this well means giving the orchestrator explicit ground rules for when and how to delegate.

MCP 2026-07-28 is the largest update to the protocol since launch. It is now easier to deploy and scale remote servers. MCP is now stateless, so it can be deployed on serverless and edge infrastructure, or scaled horizontally behind any load balancer. There is now a formal path to extend the protocol.

Anthropic has drawn criticism after releasing tools on the market that compete with those offered by partner companies. For example, Claude Design directly competes with offerings from partner design company Figma. Anthropic's data retention policies have also been criticized, but the company has pledged it won't use conversation data from Fable and Mythos - but not other models - to train new models. Researchers are calling for more transparency in AI systems.

500,000 people have already signed up for TLDR Hardware, our new twice-weekly newsletter covering chips, robotics, energy, and devices. If you work in hardware and want to help curate it, send your LinkedIn or resume to [email protected]!

As AI continues to improve, the question isn't whether superintelligence will exist, but who will have access to it. The technology could be centralized and restricted to a few institutions, or it could be used to empower everyone. The notion that AI is such a threat that the only safe path is an extreme concentration of power seems dangerous. Historically, hoping that an absolute power will benevolently provide for humanity if sufficiently enlightened hasn't led to safe or positive outcomes.

You're on track to close your first enterprise customer, but they're asking about SAML, SCIM, and OIDC. This PropelAuth blog answers the 7 most common enterprise SSO questions. Read the blog

Strong engineers think clearly all the time, even if their highest speed isn't that fast.

Kalshi and Polymarket now offer wagers on whether the FDA will approve medicines and whether drug trials will succeed.

DoorDash uses LLMs offline to enrich a knowledge graph, Instacart uses them at the query understanding layer, and Uber Eats fine-tuned a model into the embedding backbone of two-tower retrieval.

Harnesses influence cost, quality, and accuracy, and have more impact on performance than the models.

Apple Upgrade offers 12- and 24-month leasing options for the iPhone and Apple Watch, and 24- and 36-month terms for Mac and iPad.

The FCC is blocking new imports of foreign-made humanoid robots and grid inverters on national-security grounds.

Amazon plans to launch 5,105 low-Earth orbiting satellites to build on top of Globalstar's existing network of only a few dozen satellites, which it acquired in an $11.6 billion deal a few months ago. The network will power both voice and mobile data. Amazon has yet to release pricing for its service. In the US, the satellites will use Globalstar's radio spectrum in the 1.6/2.4GHz bands, and orbit between 510 km and 580 km in altitude.

OpenAI is close to leasing a $500 billion data center in southern Ohio. The company is in talks with Nvidia for a $250 billion financial backstop for the project. The US government is still talking to other potential tenants. The deal will not be final until Commerce Secretary Howard Lutnick agrees to it.

A Shanghai-based company has started mass producing immersion deep ultraviolet lithography tools. The unnamed firm aims to make about five DUV machines this year and is targeting about 20 next year. Production is still at an early stage. DUV machines are less advanced than ASML's sophisticated extreme ultraviolet lithography tools. ASML has been banned from exporting its tools to China.

Frontier LLMs were trained on text data from the internet. There's no equivalent source of physical data for robotics. All attempts to solve and get to physical AGI involve working through this data bottleneck. Progress is being made on many fronts, but nobody knows how much data is enough.

Seal is an open standard for proving any file is real. It creates sealed artifacts that can be verified by anyone. A seal proves integrity, time, and the issuing certificate. Seals are anchored to a public ledger that can be checked by anyone, forever.

Octane is a React replacement UI library with improved developer ergonomics. It is incredibly efficient and is designed to stay close to the speed of hand-written DOM code. Developers write with the React API they already know, and Octane's compiler removes the virtual DOM, Suspense waterfalls, rules-of-hooks bookkeeping, and hand-maintained dependency arrays before the app ships.

500,000 people have already signed up for TLDR Hardware, our new twice-weekly newsletter covering chips, robotics, energy, and devices. If you work in hardware and want to help curate it, send your LinkedIn or resume to [email protected]!

Nvidia has made a substantial investment into Safe Superintelligence, a company started by former OpenAI Chief Scientist Ilya Sutskever. The investment is part of a long-term partnership and will involve Safe Superintelligence receiving access to large amounts of Nvidia's flagship GPUs. Safe Superintelligence had previously been relying primarily on chips made by Google. Nvidia made the investment after getting a rare glimpse into the state of the startup's research, which has so far been kept secret.

It appears that Comcast is selling Peacock to YouTube without actually selling Peacock to YouTube. All of Peacock's content will be included in YouTube Premium subscriptions in the US starting early next year. The partnership will allow Peacock's content to be monetized through advertising. Netflix may also be exploring a Peacock add-on option.

Fine-tune your first model in 15 minutes with Crusoe Serverless Fine-Tuning. Pick from a curated library of top open models. Billing stops when they stop improving. Get started →

X is rolling out Premium and Premium+ subscribers in the US, enabling them to instantly send, receive, or request money from others on the platform without fees or limits.

Cheap and abundant compute has made it possible to test software in ways we couldn't before.

Ultimately, what matters most is solving the right problems, not your productivity setup.

Maintaining your options has a cost.

Nvidia is the largest publisher of open AI models in the world.

The scarcest resource in open source is the motivation of the maintainers.

Kimi K3 is the strongest open-weights model yet. Baseten is serving it from day 0.

🏆 #1 open model on Artificial Analysis

💻 #1 on Frontend Code Arena, ahead of every closed model

💸 Significantly more affordable than Fable 5

⚡ ~2x faster time-to-first-token on short prompts via custom kernels

Once model weights are released, Try Kimi K3 on our Model APIs or join our executive briefing online tomorrow.

Baseten also has the fastest performance in the industry for GLM-5.2 on third-party benchmarks and real-world production usage.

Read how we built the world's fastest API for GLM-5.2

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.

Built by the team benchmarking every major model launch, Gray Swan simulates enterprise environments and hits them with an automated red-teaming agent trained on threats that haven't been published yet. See how secure you are →

A team based in China modified AlphaFold to help identify what parts of a gene-editing protein enable mistakes. They then modified these areas to reduce problematic off-target effects. Producing gene-editing systems tailored to prevent known off-target events is useful for fine-tuning protein-DNA interactions, which could have applications beyond gene editing.

This post provides a high-level overview of how AI models learn new skills and behaviors.

Waymo plans to offer robotaxis in its own app starting in January, and its contract with Uber ends in May.

Nvidia is in talks to provide a roughly $250 billion backstop for OpenAI as part of a massive data-center project. The guarantee will help OpenAI lease a 10-gigawatt project that Softbank's energy subsidiary is developing. Nvidia's backing allows the data center developer to raise debt at more favorable terms. Under the deal, Nvidia would guarantee a series of financing vehicles intended to make lenders feel more confident that funding behind the project is secure.

Automatic programming, when done well, means to assume the role of Linus, with the AI agents assuming the role of the different maintainers of the different subsystems.

The Isochrones API calculates the area that a person can reach from a specific origin point within a set amount of time. An isochrone accounts for actual road network travel times to create a polygon that represents true reachability. The API can be useful for delivery logistics, finding properties within a set distance from a landmark, and urban planning.

Relays are services that proxy traffic to US models, often at a deep discount.

500,000 people have already signed up for TLDR Hardware, our new twice-weekly newsletter covering chips, robotics, energy, and devices. If you work in hardware and want to help curate it, send your LinkedIn or resume to [email protected]!

AI is more of an economically useful commodity than a supremely intelligent being that will make humanity subservient. Several leaders in the AI industry have recently come out in support of open-weight models. If these leaders believed that AI was going to become conscious and take over the world, they likely wouldn't be as openly pro-diffusion.

There are many collaborations between SpaceX and Tesla, so it makes sense to merge the companies the more they overlap.

AI will revolutionize productivity by marrying human-style intelligence to the vast, inhuman capabilities of computers.

Anthropic removed over 80% of Claude Code's system from its more advanced models. This post discusses how to apply the lessons its team learned from context engineering in Claude Code. The prompt is only a small part of the context Claude gets. The rest of the context is assembled from Skills, CLAUDE.md files, memory, and other sources. Context engineering has a big impact on the results generated when using Claude Code.

Companies like Anthropic and OpenAI claim some AI models are too dangerous to be developed in the open and must be tightly controlled for safety. The rest of the industry says that open source AI models must remain open for people to further develop technologies and build new businesses. China's rapid progress in open source AI models is escalating the fight. US officials are still debating whether to approach regulating Chinese open source models as a national security issue or issue a kind of blanket ban.

SpaceX's 13th Starship test flight ended in an intact splashdown in a remote section of the Indian Ocean on Friday. The vehicle appeared to be in fine shape on landing. SpaceX will attempt to catch the ship with its launch tower on the next test flight. The next steps for Starship are getting to orbit and returning the ship to the launch site, which will give SpaceX the capability to begin launching operational Starlink V3 satellites.

Seller is an early experimental app from Meta that aims to be to sellers what tools to edit videos were to content creators. It is free and will also be available in web browsers. The app has AI features that make selling items less of a hassle. Users can sign into Seller with their Facebook accounts, and listings are automatically synced.

Stripe is in talks to acquire OpenRouter, a startup that helps developers choose between AI models. The talks are still ongoing and could fall apart. Industry experts estimate that the business could fetch about $10 billion in a sale. Stripe is separately pursuing a deal for PayPal.

ChatGPT Health is now available to all US-based users over 18. The feature helps users with health-related queries. Users can let the feature access data from apps and integrate medical records from hospital systems. While several studies have found that AI bots are not reliable for medical advice, this has not deterred AI companies from launching health-related features.

Unitree's GD01 is a transformable mecha robot. It can transport a pilot in its torso-mounted cockpit, walk on either two legs or four, and wield fists powerful enough to knock down walls. Standing 9 ft tall and weighing half a ton, it has heavy-duty mechanical hinges that connect titanium-alloy limbs to a protective carbon fiber casing. Pictures of the robot are available in the article.

SpaceX has started to turn away satellite operators seeking dedicated rides on its Falcon 9 rocket beyond 2028. The company has also stopped building some non-reusable components for the Falcon family. SpaceX has publicly stated that it intends to offramp its Falcon 9 rocket for Starship, but has not disclosed a timeline. The company is likely to still use the Falcon 9 for launches for the Department of Defense and NASA.

No amount of harness engineering or loopsmaxxing can solve what is fundamentally a model-training issue. Engineers are still stuck reading code. Companies can move 10x-100x faster by convincing themselves that the code doesn't matter anymore, but they could also embrace their constraints and just move 2x-3x faster, safely.

The field of formal verification and deterministic system testing is about to cross a chasm. There are not enough skilled practitioners in the field. Bug-catching tools, when used in conjunction with adversarial code reviews by an LLM and with language analyzers driven by pre-commit hooks, will be a key component in software factories. They will enable people/agents to deliver reliable software without the burden of having to learn this specialized knowledge.

500,000 people have already signed up for TLDR Hardware, our new twice-weekly newsletter covering chips, robotics, energy, and devices. If you work in hardware and want to help curate it, send your LinkedIn or resume to [email protected]!

Optimizing the cost-to-outcome frontier in real world context means using the right model for each task, and optimizing the context, skills, tools, and agent harness around it. It's possible to take saturated frontier capabilities and deliver them at scale and at lower cost through models optimized for high-usage products, while continuing to use frontier models for frontier needs. The model is only one part of the hill-climbing system. Harness, memory, context, tools, skills, and user interactions all shape the evals and performance of these agentic systems.

Huawei is leading the way in reducing China's dependence on foreign AI chips. The company claims it has figured out workarounds for making near state-of-the-art silicon without leading-edge machinery. Huawei's deputy chairman attributes the company's progress to the restrictions the US has imposed on hardware exports. China still lags far behind the US, and chip-production capacity problems are preventing Chinese companies from deploying AI as quickly as they would like.

Canvas UI is an open source library of creative, framework-agnostic components drawn on canvas.

The 'frontier' moniker justifies capital spending to investors, who believe that the spending is the moat.

Google's research points to a future where the demand for highly skilled workers is accentuated rather than diminished.

SIMD is just about as easy as a for loop.

Download a dictation app and start talking.

Google has added a new video selfie option for account recovery.