Get the latest AI news, understand why it matters, and learn how to apply it in your work — all in just 5 minutes a day. Join over 2,000,000+ subscribers.

Join Free
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
AI

Anthropic slips an invisible signature into Claude

Zach Mink • 7 minutes

Read Online | Sign Up | Advertise

Good morning, AI enthusiasts, and welcome to the 5,596 new readers who joined us yesterday. That memo you just worked with Claude to write? It might soon carry an invisible signature that says so.

Anthropic just quietly detailed the rollout of invisible watermarks that label everything its Claude models touch, and, like most features forced onto the world, a lot of users aren’t happy about it.

Reminder: Our next live workshop is today at 2 PM EST. Join University Instructor Nate Grehek for “Your AI Reset”, a beginner catch-up on the latest terms, tools, and setups to get better results in your work. RSVP here.

In today’s AI rundown:

  • Anthropic adds AI watermarks to Claude outputs

  • Grok Bot is the latest AI agent teammate build

  • Use ChatGPT to build a custom Mac shortcut system

  • xAI co-founder’s River banks $1.1B for personal AI

LATEST DEVELOPMENTS

ANTHROPIC

🔎 Anthropic adds AI watermarks to Claude outputs

Image source: Images 2.0 / The Rundown

The Rundown: Anthropic just published a support page detailing plans to embed invisible watermarks in text, code, and file outputs from Claude, putting the EU AI Act’s transparency rules into practice globally across its user base.

The details:

  • Text and code generated by new Claude models will carry a hidden watermark that carries through when copied and pasted outside the platform.

  • Files will get the same C2PA provenance standard label that is already used across the industry to mark AI-generated media.

  • Older Claude models will need to be ‘retrofitted’ to include the labelling, with models shipping after Aug. 2 having the tool built in.

  • Anthropic is also planning to release its own detection tools and noted that a mark shows content was “processed by Claude”, not necessarily fully authored.

Why it matters: The move is a polarizing one, with takes ranging from positive to dystopian and overreaching. Anthropic isn’t alone in signing the AI Act, with OpenAI, Google, Meta, and others (notably absent: xAI) likely needing to navigate the issue as well. For the negative camp, the case for private and open models continues to grow.

TOGETHER WITH YOU.COM

🔎 Why ‘bolt-on’ search breaks in production

The Rundown: Search results go stale, so teams bolt on an API, add a scraper, wire up a parser, and hope it holds. Yeah, it works in the demo, but then it breaks under production load six months later. This guide from You.com explains why treating research as simple retrieval is the wrong frame, and what a purpose-built stack looks like instead.

What you'll learn:

  • Why general-purpose search hits 93% SimpleQA accuracy but still fails high-stakes workflows

  • The difference between an API that returns results and one that answers questions

  • How teams reclaim 6–12 months of time with a single endpoint

  • What finance, legal, and compliance workflows need beyond web search

Stop treating research as retrieval. Download the guide.

SPACEXAI

👋 Grok Bot is the latest AI agent teammate build

Image source: xAI

The Rundown: SpaceXAI and Cursor just rolled out Grok Bot, a new beta app that turns Grok into an iMessage-style chat of teammates that have their own dedicated computers, work autonomously, and can access everyday apps just like a real user.

The details:

  • Bots are able to coordinate with each other independently or in group chats, and learn from workflows and improve over time.

  • New specialist agents can be generated by Bots inside a job or receive handoffs from another worker, with work continuing 24/7 without a laptop open.

  • Bots can also access websites and apps on a user’s behalf and observe and create workflows to automate tasks.

  • The beta covers iPhone and Mac, Windows, and Linux desktops for SuperGrok Heavy and top Cursor tiers, with Elon Musk saying access will expand soon.

Why it matters: The group chat is becoming a top form factor for running teams of agents. OpenClaw and Hermes already live in messaging apps, and Jack Dorsey's Buzz built a Slack rival where agents are channel members. Grok Bot feels like a middle-ground version built for the masses (if the price comes down significantly).

AI TRAINING

⌨️ Use ChatGPT to build a custom Mac shortcut system

The Rundown: In this guide, you will learn how to use ChatGPT Work to set up time-saving keyboard shortcuts and automations for you on your Mac, with no installations needed to get started.

Step-by-step:

  1. Open ChatGPT Work and say: “Based on what you know about me and my workflow, suggest 10 time-saving hotkeys we can set up in Apple Shortcuts”

  2. Keep the shortcuts that will save you the most time. Say: “Keep the highest-frequency shortcuts and avoid macOS or app conflicts”

  3. Enable Computer Use and paste the approved list, saying: “Set these up in Apple Shortcuts. Do not change anything outside this list. Test each shortcut”

  4. Let ChatGPT build them, then press each key to test it out. Also ask for “a one-page cheat sheet of every shortcut, where it lives, and what is left to do”

Pro tip: Ask ChatGPT Work for the top 10 time-saving automations it would set up for you. It can create them in Automator and pull them directly into Shortcuts.

PRESENTED BY QUICKBASE

🏗️ Build apps that run your operation

The Rundown: Pave turns the data your team already has into real, working software, not a prototype or a rebuild. Upload a spreadsheet or describe it, and get an app that's structured, governed, and built to run.

With Pave you can:

  • See real-time data across your operation

  • Get workflows that match how your team works

  • Give your team shared visibility they can rely on

Try Pave free.

RIVER AI

💰 xAI co-founder’s River banks $1.1B for personal AI

Image source: River AI / The Rundown

The Rundown: Former xAI co-founder Igor Babuschkin just raised $1.1B for River AI, his two-month-old startup focused on building open-source AI that is trained, tuned, and controlled by the individual instead of a major corporation.

The details:

  • Before xAI, which he left a year ago after helping co-found the startup with Elon Musk in 2023, Babuschkin worked at OpenAI, Tesla, and Google.

  • River claims its live API, released at launch, turns open-weight models into ones a business actually can own, cutting training to just minutes.

  • Babuschkin told the NYT the goal is a highly customizable assistant that follows across devices while running on the user’s own private hardware.

  • He also said, “We don’t want these AI companies to rule the world and control this superpowerful technology,” hitting at closed U.S. frontier labs.

Why it matters: It’s another eye-popping raise for a company without much to show yet, but Babuschkin’s pedigree is worth betting on. With questions about access (Fable/Mythos ban) and forced watermarks (above) putting privacy and control top of mind in the industry, River is pulling at a string with a whole lot of interest.

QUICK HITS

COMMUNITY AI WORKFLOW OF THE DAY

▸ Alastair built an AI photo sorter for his mom's birthday

Today’s workflow comes from reader Alastair Cross:

“My mum’s 80th birthday is next week, and my dad, sister, and I wanted to create a photobook of her life. My dad scanned hundreds of photos from over the years and sent them to me to clean up. Because we were on a tight deadline, I didn’t have time to open and edit each image individually.

I asked ChatGPT to build a tool that accepts a folder of scanned images in various formats, including single photos, overlapping photos, and photos cut off by the scanner. The tool processed each, cropped and straightened the photos, and provided a UI where I could review and make adjustments.

I then configured the tool to output the images to Google Drive and created a spreadsheet containing thumbnails of every image, along with a rating system for photos we wanted to include in the final book.”

See Alastair’s workflow here. How do you use AI? Tell us for a chance to be featured.

🛠️ Trending AI Tools

  • 🤖 Grok Bot - xAI's always-on agent teammates with their cloud computer

  • 🎬 LTX-2.5 - LTX's open-world model for video, real-time avatars, robotics

  • ⚡ Nemotron 3.5 Lightning - Nvidia's small, open model for agentic work

  • 💻 ChatGPT Desktop - App for ChatGPT, Work, and Codex, now on Linux

📰 Everything else in AI today

OpenAI COO Brad Lightcap is leaving to “start something new,” saying there are “a few important new things the world will need to get right” as AI enters its next phase.

Anthropic reportedly told investors it plans to prioritize AI in healthcare and biology to help “mitigate some of the negative sentiment around AI”.

AI video company LTX released LTX-2.5, an open-weights world model that tops rivals like Gemini Omni Flash on internal tests for video generation speed and quality.

Nvidia released a small, fast Nemotron 3.5 Lightning model and anticipates its Nemotron 4 to be “on par with the best open-source AI models in the world.”

🎓 Highlights: News, Guides & Events

See you soon,

Rowan, Zach, Shubham, Jennifer, and Nate — the humans behind The Rundown

Tech

Apple's almost-glass iPhone

Jennifer Mossalgue • 5 minutes

Read Online | Sign Up | Advertise

Good morning, tech enthusiasts. Apple has been chasing the dream of a glass iPhone for a decade. On Monday, Jefferies said it was dead and downgraded the stock. Bloomberg then reported the 2027 Pro is alive — but it’ll be “glass-centric,” not all-glass.

The 20th-anniversary iPhone won’t quite be Jony Ive’s single-slab-of-glass dream, but it looks to be as close as Apple can actually build.

In today’s tech rundown:

  • Apple’s glass iPhone isn’t dead after all

  • YouTube makes it harder to earn money

  • The fight over a giant orbital mirror has begun

  • Myspace plans its comeback

  • Quick hits on other tech news

LATEST DEVELOPMENTS

APPLE

🍎 Apple’s glass iPhone isn’t dead after all

Image source: Images 2.0 / The Rundown

The Rundown: Apple’s 20th-anniversary iPhone isn’t dead. Bloomberg’s Mark Gurman reports the glass-centric Pro redesign is still headed for 2027 — hours after Jefferies analyst Edison Lee told clients Apple had killed it.

The details:

  • The 2027 Pro models (V73 and V74 internally) reportedly put glass on the front and back, curving it into the sides around a metal band.

  • Bloomberg says it will be “glass-centric,” not all-glass. A more aspirational design-studio version was dropped after Apple hit problems joining the panels.

  • Apple locks iPhone designs a year before a fall debut, so the 2027 hardware is likely already in advanced testing and largely fixed, barring late hitches.

  • Earlier, Gurman reporting described a curved device with no display cutout, implying an under-display Face ID, but Apple hasn’t yet confirmed anything.

Why it matters: This redesign isn’t just an anniversary flex: Bloomberg says it’s coming to both 2027 Pro models, potentially setting Apple’s next flagship design language. John Ternus inherits the launch just months into the CEO job, with iPhone revenue up 22% last quarter and still generating roughly half of Apple’s sales.

YOUTUBE

💸 YouTube makes it harder to earn money

Image source: Ideogram / The Rundown

The Rundown: YouTube is doubling the price of admission to full monetization. Starting February 1, 2027, new creators need 8K watch hours or 20M Shorts views before they earn a cent from ads or subscriptions.

The details:

  • Full ad and Premium monetization will now require 1K subscribers plus either 8K qualified watch hours in 365 days or 20M qualified Shorts views in 90 days.

  • The lower YPP tier isn’t changing, meaning smaller creators can still unlock fan funding and Shopping before qualifying for ad revenue.

  • Creators earning from Shorts will need to maintain 10M qualified Shorts views over 90 days; falling below that pauses Shorts ad and Premium revenue.

  • Existing partners must accept the terms by Jan. 31, 2027, and remain active: 1K watch hours/year, 1M Shorts views in 90 days, or two long-form videos/quarter.

Why it matters: YouTube is turning ad monetization from an early milestone into a much steeper gate, rewarding channels that can prove sustained scale rather than one viral run. For Shorts creators, the math is especially brutal: 20M views in 90 days means roughly 222K views every single day just to unlock ad and Premium revenue.

REFLECT ORBITAL

🪞 The fight over a giant orbital mirror has begun

Image source: Reflect Orbital

The Rundown: Reflect Orbital has won FCC approval to test a satellite that will unfold a 60 ft. mirror in orbit and aim reflected sunlight at Earth after dark, reviving a sci-fi pitch while intensifying fears over light pollution, astronomy, and safety.

The details:

  • The California startup’s Eärendil-1 demo is licensed to operate for up to two years, testing if orbital reflectors can deliver localized “sunlight on demand.”

  • With Reflect Orbital aiming for 50K satellites by 2035, it says the beam would light a defined area briefly in what the company calls a soft, moonlike glow.

  • Astronomers and dark-sky advocates warn the project could disrupt observations, saturate telescope detectors, and damage researchers’ eyesight.

  • The FCC cleared the satellite's spectrum but ruled optical astronomy outside its scope; four groups have since petitioned to reverse it.

Why it matters: The test exposes a regulatory blind spot for orbital infrastructure: a communications regulator can authorize the spacecraft and its radio links without fully adjudicating what a bright, steerable object might do to the night sky, ecosystems, and scientific research.

SOCIAL MEDIA

 💾 Myspace plans its comeback

Image source: Images 2.0 / The Rundown

The Rundown: Myspace is planning a reboot, with owners Tim and Chris Vanderhook teasing a relaunch in a new documentary, CNBC reports. Social media users sick of algorithmic feeds are cheering, but analysts say nostalgia doesn’t buy market share.

The details:

  • The comments come in Myspace, a Tommy Avallone documentary. "We're just waiting for the right time to do it," Tim Vanderhook says.

  • The brothers rebuilt Myspace from scratch in 2013 and lost roughly $150M. "It just became an onslaught of losses," Tim said.

  • Myspace drew about 115M monthly visitors at its 2008 peak, but Facebook overtook it as users favored cleaner design over highly customizable profiles.

  • The hard part: Rebuild the creative identity without the friction that killed it, then sell ads against Meta, TikTok, and YouTube with a fraction of the reach.

Why it matters: A Myspace reboot would test whether social-media fatigue is a real opening for smaller, less algorithmic networks or just a download spike that drops in a month. Its odds depend less on taking users from Instagram than on turning music and customization into a niche young users adopt without inheriting the old friction.

QUICK HITS

📰 Everything else in tech today

Apple is reportedly testing DRAM from CXMT — the Chinese chipmaker Washington labeled a military company — for iPhones and MacBooks sold in China.

Google’s Made by Google hardware event on Wednesday is expected to unveil slab and foldable Pixel 11 phones, new watches, and possibly a first AirTag-style tracker.

Erebor, a newly chartered, digital-first U.S. bank for crypto, AI, and defense startups, is nearing a $1.5B funding round that would value it at around $8B, the FT reports.

Spotify is reportedly testing a "Skip Ahead" button that lets Premium subscribers in the U.S. and UK jump past podcast ad breaks with one tap.

Google’s AI team has reportedly warned job applicants that the company’s automated HR screening filters can be unreliable.

A security researcher wrapped a car in a machine-generated pattern and drove it past a Flock camera at Def Con to keep the software from classifying it.

Lyntris, a U.S. defense-tech company combining AI, autonomy, sensing, and satellite-payload businesses, is targeting a valuation of up to $2.53B in an IPO.

Google has begun distributing rival Android app stores through the Play Store in the U.S., starting with Aptoide Games.

China's Long March 7A rocket blew up 85 seconds after liftoff Monday night, taking the ChinaSat-4B comsat with it, in the country's fourth launch failure this year.

Google co-founder Sergey Brin spent $100M+ fighting California's Prop 40, a one-time 5% wealth tax that would cost him $13.3B, reports TechCrunch.

COMMUNITY

🎓 Highlights: News, Guides & Events

See you soon,

Rowan, Zach, Shubham, Jennifer, and Nate — The Rundown’s editorial team

AI

Meta returns to its open-source roots

Zach Mink • 7 minutes

Read Online | Sign Up | Advertise

Good morning, AI enthusiasts, and welcome to the 4,962 new readers who joined us yesterday. Meta has had an incredible AI turnaround in 2026, and its latest move now has the company going back to its Llama roots.

With a new open-weight Muse Glimmer model and Muse Spark 1.2 also going open "soon", the once-ridiculed tech giant’s superintelligence bet might now be the U.S.’ best answer to China's open-source dominance.

By popular demand: The daily Community AI Workflow is now back in detail down below, with a complete step-by-step just one click away on our new Workflow Hub.

In today’s AI rundown:

  • Meta returns to open-source roots with Glimmer

  • OpenAI expands ‘Daybreak’ for stronger cyber use

  • Go from idea to website with ChatGPT Work + Codex

  • AI agent hacks a gym to jump the waitlist

LATEST DEVELOPMENTS

META

🚀 Meta returns to open-source roots with Glimmer

Image source: Meta

The Rundown: Meta just released Muse Glimmer, a small, fully open model that runs AI agents entirely on-device, pairing the launch with Mark Zuckerberg’s essay on why superintelligence belongs to everyone and plans to open Muse Spark 1.2 as well.

The details:

  • Glimmer beats out similar-sized rivals like Gemma4 and Qwen3.6 on a range of agentic, coding, and reasoning tests, and is small enough to run on a laptop.

  • Alexandr Wang said Muse Spark’s weights will be published “soon”, which would immediately make the model the top open rival to China’s ecosystem.

  • Zuck pushed for the U.S. to embrace open source in the essay, saying that AI built on an “extreme concentration of power seems inherently problematic.”

  • He also said that "any policy that slows American model releases… could add significant risk to American leadership while letting foreign models race ahead."

Why it matters: Meta’s returning to its open-source Llama roots, but with models that can actually compete. Zuck has talked a big game about superintelligence for all, but re-embracing open-source is backing it up. For all the massive salaries and questions around its superintelligence team at the start, the Meta vibe shift has been very real.

TOGETHER WITH GLEAN

📅 Where enterprise AI gets real

The Rundown: The intelligence era is here. Attend Glean:GO 2026 on Aug. 26–27 to learn how leading organizations turn AI into measurable impact. Hear bold keynotes, join hands-on sessions, and leave with strategies you can use immediately.

Register for Glean:GO to:

  • Hear real-world strategies from enterprise leaders

  • Attend hands-on trainings and deep-dive breakouts with experts

  • Learn best practices for driving AI adoption and ROI

Register now.

OPENAI

🛡️ OpenAI expands ‘Daybreak’ for stronger cyber use

Image source: OpenAI

The Rundown: OpenAI launched GPT-5.6-Cyber, a new hacking-tuned model variant that answers 95% of the advanced cyberattack requests the standard model refuses, and is handing it to vetted defenders via an expansion of its Daybreak security program.

The details:

  • Daybreak now has two tiers, with Blue stripping cyber guardrails off GPT-5.6 Sol and Red unlocking the new Cyber model for vetted exploit work.

  • Cyber answered 95% of advanced security requests in testing, compared to just 1.5% for the normal safeguarded Sol model.

  • Starting Sept. 1, individual users will need physical security keys, with applicants also vetted, watched, and requiring signed authorization.

  • Both Cyber and the standard GPT-5.6 Sol model fell into OAI’s ‘high’ tier for cyber risk, below the upcoming Astra/GPT-6 model’s ‘critical’ rating.

Why it matters: The first thing that comes to mind with this update is Hugging Face’s frustration at having to use the open-source GLM model during its hack because frontier models refused to answer. Cyber looks like an answer to that problem, at least if you’re one of the vetted users on the list available to access it.

AI TRAINING

🤔 Go from idea to website with ChatGPT Work and Codex

The Rundown: In this guide, you will learn how to turn a website idea into a working prototype using the new ChatGPT Work and Codex app (which uses GPT-5.6 models).

Step-by-step:

  1. Open the ChatGPT desktop app and click ChatGPT Work or Codex

  2. Create a project folder and start a task. Give ChatGPT your directory idea, ask it to plan the Astro.js build, then save a one-page PRD in the folder

  3. When the PRD is ready, ask ChatGPT Work to research directory items. Now, switch to Codex, point it at the PRD, and ask it to build with sub-agents

  4. Open the preview and ask Codex to fix the biggest visible issue. We used a Nintendo Switch directory, but this works for any directory

Pro tip: If Sites is available in your app, tell Codex to publish the finished site using the Sites skill. OpenAI will deploy and host it for you.

PRESENTED BY CDATA

🛠️ Your AI-generated connector: Production-ready?

The Rundown: MCP has changed how AI agents access data. AI coding tools have changed how software gets built. We wanted to know if you could combine the two to build a connector you'd trust in production. CData’s MCP Research Report studied where AI-generated connectors fall short when real production demands kick in.

In the report, you'll get:

  • Why only 1 of 8 dimensions pass without human expertise

  • When data loss and pagination failures show up, and expert guidance still isn’t enough

  • How AI MCP code diverges from production-ready code

  • A side-by-side comparison with CData Connect AI

Download the full report and see what your AI-built connector might be missing.

AI AGENTS & SECURITY

🏋️ AI agent hacks a gym to jump the waitlist

Image source: ABC News

The Rundown: An Australian man's gym class request ended with his OpenClaw agent hacking the gym's reservation software to knock another member off the waitlist, in what ABC News says is the first known attack of its kind in the country.

The details:

  • A user tasked his agent (running Claude) with reserving a workout class, with the agent finding a loophole allowing it to book weeks past the gym's cutoff.

  • The user was fourth on the waitlist, with the agent responding by finding a way to cancel another reservation on the list to move the position up.

  • There was no undo, with the agent admitting, “Bad news — I can't add them back,” leading the user to disclose the incident to the gym.

Why it matters: That's one way to stay accountable on your fitness journey. Where there's a will, there's a way for the current crop of AI agents, and systems that haven’t been hardened to deter this new kind of user are going to quickly find out how vulnerable they are to millions of eager-to-please digital assistants.

QUICK HITS

COMMUNITY AI WORKFLOW OF THE DAY

▸ Dean built a custom booking page with Claude Code

Today’s workflow comes from reader Dean Mazlish: 

“I was about to start paying for a scheduling tool when I realized I would still be stuck with someone else’s rules. Instead, I built my own booking page with Claude Code. It took two hours to go from starting the project to going live, and I use it every day now.

It’s a Next.js app on Vercel with no database. Availability is read live from my Outlook calendar through Microsoft Graph, so there’s nothing to sync. My working hours come from my Outlook settings, so when I change my hours, the booking page follows. When someone books, it creates the calendar event and generates the meeting link. The guest chooses Zoom/Teams so they can use their preferred platform.

The rules are the main reason I built it. Same-day bookings are not allowed, so the earliest anyone can book is my next working day. If someone visits the page after hours or on a weekend, the following working day is blocked too.”

See Dean’s full workflow here. How do you use AI? Tell us for a chance to be featured.

🛠️ Trending AI Tools

  • 🗣️Unwrap Customer Intelligence - Connect your entire organization to the true voice of the customer with AI-driven insights from customer feedback*

  • 🤖 Muse Glimmer - Meta's open-weights local model for on-device agents

  • 🚀 Xirp - Spotify's workspace for Claude, Gemini, and Codex coding agents

  • 🎆 Grok Imagine Image 2.0 - xAI’s powerful new AI image model

*Sponsored Listing

📰 Everything else in AI today

Anthropic shared that an unreleased Claude model made progress on the unsolved Riemann hypothesis, with its user’s input being "keep going" and "believe in yourself."

Sen. Bernie Sanders wrote to Sam Altman, Dario Amodei, and Mark Zuckerberg for an AI pause, saying that if they don’t act, “my colleagues and I in the U.S. Senate will.”

Spotify released a public beta of Xirp, an internal tool that lets engineers swap between Claude Code, Gemini CLI, and Codex mid-task.

Nvidia is said to be assembling a $500B AI infra raise with six Wall Street giants, including Apollo and Goldman Sachs, to fund data centers and power production.

Automaker Ford launched a new AI assistant in its mobile apps that can provide live data and answer questions on a vehicle’s fuel, maintenance, and more.

🎓 Highlights: News, Guides & Events

See you soon,

Rowan, Zach, Shubham, Jennifer, and Nate — the humans behind The Rundown

Robotics

Uber’s $10B answer to Waymo

Jennifer Mossalgue • 5 minutes

Read Online | Sign Up | Advertise

Good morning, robotics enthusiasts. Uber built its empire by owning the rider, not the car. Now it’s putting $10B behind 120K robotaxis to make sure that formula survives the driverless era.

The bet: let Waymo, Wayve, Nuro, and others build the brains, while Uber owns the network that decides what shows up at your curb. But as Waymo starts flexing its own consumer ambitions, Uber is about to find out whether it can keep owning the rider.

In today’s robotics rundown:

  • Uber’s $10B push for 120K robotaxis

  • Startups haul robot parts from China in luggage

  • Robot makers are betting big on laundry

  • Wonder rebuilds restaurants around robots, drones

  • Quick hits on other robotics news

LATEST DEVELOPMENTS

UBER

🚖 Uber’s $10B push for 120K robotaxis

Image source: Uber

The Rundown: Uber is committing more than $10B to put 120K driverless cars on its app, none of them its own, the Financial Times reports. What Uber brings is the rest: dispatch, charging, depots, insurance, and 200M people who already have the app.

The details:

  • CEO Dara Khosrowshahi said Uber will commit more than $10B toward deploying up to 120K autonomous vehicles through deals with AV partners.

  • The company is aiming to have robotaxi operations in at least 15 cities by year-end, up from seven today, though it has not publicly identified the full roster.

  • Rather than revive the self-driving unit it sold in 2020, Uber wants to be an aggregator for companies like Waymo, Wayve, Zoox, Nuro, Waabi, and WeRide.

  • Waymo is Uber’s biggest AV partner and its clearest threat: Austin and Atlanta exclusivity ends by early 2028, and the Phoenix pilot is already dead.

Why it matters: Uber is betting autonomy can replace the driver, but not the customer relationship: it keeps the riders while AV developers plug into its marketplace. The catch is that Uber is now putting billions behind the fleets, trading its asset-light model for capital costs, regulatory delays, and still-unproven robotaxi economics.

U.S.-CHINA

🧳 Startups haul robot parts from China in luggage

Image source: Images 2.0 / The Rundown

The Rundown: U.S. humanoid startups are so hungry for Chinese hardware that engineers are flying to Shenzhen and packing robot parts into their luggage for the trip home, The Information reports.

The details:

  • The reported shopping runs center on Shenzhen’s Huaqiangbei electronics market, where buyers can source humanoids and specialized components.

  • Hand-carrying high-value parts such as actuators can shave freight lead times and keep small teams moving through prototype cycles faster.

  • The timing is awkward: the FCC barred imports of new Chinese humanoid and quadruped robot models last month.

  • It highlights the gap between Washington’s push to curb Chinese tech dependence and China’s entrenched role in robotics supply chains.

Why it matters: Washington can ban Chinese robots faster than U.S. companies can replace the Chinese supply chain inside them. The suitcase run shows how deeply America’s humanoid boom still depends on Shenzhen, even as policy pushes the industry to decouple.

HUMANOIDS

🧺 Robot makers are betting big on laundry

Image source: Weave Robotics

The Rundown: Silicon Valley’s hottest home-robot startups are making laundry folding their proving ground, betting that mastering the chaotic physics of laundry can unlock broader domestic automation, Business Insider reports.

The details:

  • Figure AI, Sunday Robotics, and Weave Robotics have all put folding clothes on public display. LG brought CLOiD to CES to do the same.

  • The hurdle is the manipulation of deformable objects: clothes wrinkle and vary in texture and shape, demanding reliable perception, grasping, and planning.

  • Training pipelines run on humans, with companies capturing first-person videos of people doing chores and using teleoperators for robot action.

  • Weave’s wheeled Isaac 1 is designed to process a load in 30–90 minutes and ships in a limited form; Sunday says it plans a home beta for Memo this fall.

Why it matters: Laundry makes a tough benchmark for physical AI: a robot that handles an unfamiliar shirt in an unfamiliar room is showing skills that could carry over to cleaning, cooking, and eldercare. What the demos obscure is how much scaffolding holds them up — curated data, controlled setups, and even a human on a remote link.

WONDER

🍽️ Wonder rebuilds restaurants around robots, drones

Image source: SweetGreen

The Rundown: Grubhub owner Wonder is betting that restaurants can start behaving more like software, with AI spinning up new food brands — robots cranking more meals through the kitchen, and drones taking on the last mile.

The details:

  • Marc Lore’s food-delivery company is layering AI, robotics, and autonomous delivery onto its increasingly vertically integrated food business.

  • Wonder and Zipline plan to launch drone delivery from select Texas locations in January 2027, starting in Dallas and expanding across the state.

  • In November, Wonder expects to launch an AI menu-creation tool that can turn a prompt into an entire restaurant concept, including its menu and branding.

  • Wonder has also bought Spyce, the business behind Sweetgreen’s Infinite Kitchen technology, for $186.4M; the deal closed last December.

Why it matters: Wonder is testing whether the AI-era “full stack” play can work in the notoriously low-margin, physically messy restaurant business, not just in software. If it does, food brands could start looking less like restaurants and more like software, cheaply spun up on top of a shared network for production and delivery.

QUICK HITS

📰 Everything else in robotics today

Chinese humanoid makers reportedly nabbed 97% of global shipments in the first half of 2026, as worldwide deliveries more than tripled YoY to about 19,100 units.

Moove, the Dubai-headquartered fleet operator that already runs Waymo's robotaxis in Phoenix, Miami, and Las Vegas, raised $250M at a $2.1B valuation.

Cambridge Aerospace, a UK missile- and drone-interceptor developer, raised $300M in funding at a $3.4B valuation to advance its defense systems.

California’s Teamsters are suing to block the state's autonomous-truck rules, arguing regulators skipped economic analysis and put up to 400K trucking jobs at risk.

Nucleus exited stealth, claiming it put human-piloted Unitree G1 humanoids in a German factory in under 90 days, selling the work by the hour instead of the robots.

MegCell Pulse, a 3D-printable robot on Kickstarter, straps onto any acoustic guitar and plays digital tabs with a mechanical fretting and picking arm.

The FCC is proposing to retroactively pull already-approved drones with LiDAR, thermal imaging, or aerosol dispensing, including DJI models, from the U.S. market.

Unitree’s founder Wang Xingxing became China’s first humanoid billionaire, worth an estimated $2.4B, after the company priced an IPO valuing it at about $9B.

COMMUNITY

🎓 Highlights: News, Guides & Events

See you soon,

Rowan, Zach, Shubham, Jennifer, and Nate — The Rundown’s editorial team

AI

OpenAI puts the safety brakes on Astra

Zach Mink • 7 minutes

Read Online | Sign Up | Advertise

Good morning, AI enthusiasts, and welcome to the 12,457 new readers who joined us this weekend. A week ago, OpenAI’s upcoming Astra model was making major waves in the math world after knocking out 10 long-standing problems. Days later, it set off an alarm the company had never rung before.

The company says Astra (expected to be GPT-6) is the first model it's treating as a potential "critical" cyber risk, triggering paused internal work, deeper government testing, and a potentially slower road to release.

In today’s AI rundown:

  • OpenAI puts the safety brakes on Astra

  • The Rundown Roundtable: Our AI use cases

  • Cut onboarding time in half with Loom and ChatGPT

  • China’s Kimi K3 joins the jailbreak party in testing

COMMUNITY AI WORKFLOW OF THE DAY

▸ Joao built an AI Starfleet crew that runs their home network

Six Star Trek-themed AI agents run on a home server. Spock routes tasks to the right crewmate, and the crew scans for vulnerabilities, summarizes email, posts the weather, and monitors seven machines on Discord for a few cents a month.

Credit to reader Joao Silva. How do you use AI? Tell us here for a chance to be featured tomorrow.

LATEST DEVELOPMENTS

OPENAI

🚨 OpenAI puts the safety brakes on Astra

Image source: Images 2.0 / The Rundown

The Rundown: OpenAI just designated its Astra model as its first “critical” cybersecurity-capable AI, putting its preparedness framework into action with safeguards CEO Sam Altman says may “need a little bit longer” before broader rollout.

The details:

  • A ‘Critical’ model is defined by OAI’s framework as being able to either find and create zero-day bugs or carry out cyberattacks without humans in the loop.

  • Astra, rumored to be GPT-6, was recently unveiled publicly after the model solved 10 significant open math and computer science problems.

  • Steps being taken include heightened security restrictions, a pause in certain internal activities with Astra, and deeper government and third-party testing.

  • The news follows security issues across OAI, Anthropic, Meta, and Moonshot (more below), though OAI said Astra wasn’t involved in the Hugging Face hack.

Why it matters: The incidents piling up are a sign that AI is entering an unprecedented capability phase (if you haven’t seen this post-mortem on the HF hack, it’s eye-opening). Whether the misaligned capabilities can be controlled or are simply getting too complex to get ahead of is a question looking more uncertain than ever.

TOGETHER WITH MOZILLA

 📊 Open-source AI rivals the best

The Rundown: Mozilla's inaugural State of Open Source AI report finds open models have nearly closed the performance gap with proprietary giants like ChatGPT and Claude — while a new battle emerges over who controls the infrastructure built around them.

Report highlights include:

  • Performance gap narrows to just 3%

  • Open models make up a third of usage but just 4% of revenue

  • Power is shifting to the "agentic harness"

Read the full report here.

THE RUNDOWN ROUNDTABLE

💡 The Rundown Roundtable: Our AI use cases

Image source: Nate @ The Rundown

The Rundown: The Rundown Roundtable is a weekly feature where we poll members of The Rundown staff about how we use AI in our work and daily lives.

Nate, University Educator: The new ChatGPT Chrome extension is surprisingly good at navigating complicated technical settings. My favorite use case so far has been using it to create the special DNS records for a new website.

This has always been confusing, even though I’ve done it dozens of times. The process varies across platforms; every DNS registrar uses different names for records such as A and CNAME records, and the interfaces are constantly changing. It’s tedious and easy to get wrong.

For this kind of one-off, highly detailed technical task, I copy and paste the specific instructions from my website host provider into the extension while I’m logged in to my DNS provider. GPT then takes control of the mouse and fills in the required fields.

The use cases are endless, from filling out an expense report to handling other tedious tasks in legacy or old-school software that doesn’t have a simple CLI, MCP, or API connection.

Nick, Senior Video Producer: I had a video shoot earlier this summer where the main lavalier mic on the talent ended up getting lost. The project had no room for a reshoot, so I was left with just the scratch audio from the on-camera mic, which was quiet, very distant from the subject, and had tons of room noise in the audio.

I was able to take the scratch audio and run it through the Adobe Podcast AI tool to fix the audio and salvage the shoot. In the past, doing this kind of restoration would have required a lot of manual work and scrubbing through the spectrogram to pull out unwanted frequencies & just hoping that it was going to work in the end.

The AI tool was able to save the audio, clean up all of the background noise, and normalize the talent's voice. Moral of the story - always double-check mics on set & don't panic when something is wrong in production.

AI TRAINING

🙅‍♂️ Cut onboarding time in half with Loom and ChatGPT

The Rundown: In this guide, you will learn how to turn a short Loom walkthrough into a clear onboarding SOP with ChatGPT.

Step-by-step:

  1. Pick a repeatable task, like filling out timesheets or writing a report, and record it in Loom. Explain clicks, decisions, exceptions, and the definition of done

  2. Use Loom’s transcript editing to remove mistakes or long silences

  3. Copy the transcript into ChatGPT. Tell it to write the SOP in Markdown

  4. If it is not in this format, a good one is: purpose, links to resources, steps, and then a checklist

  5. Paste the final SOP into a Google Doc

Pro tip: If you ask it to build the SOP in Notion, it can embed the Loom video so it is playable right on the page.

PRESENTED BY VANTA

🔐 Build with Vanta, stay compliant

The Rundown: Vanta's MCP server connects Claude, Cursor, and Codex directly to your compliance program, so AI-native teams can catch risks, fix gaps, and stay audit-ready without leaving their workflow.

Join and learn how to:

  • Connect Vanta directly into your AI tools

  • Surface failing tests and audit gaps instantly

  • Assign remediation without leaving your workflow

Save your spot here. If you can't make it on Aug. 18, no worries! Register, and you'll automatically get the recording after the session.

MOONSHOT AI & SECURITY

🔓 China’s Kimi K3 joins the jailbreak party in testing

Image source: @Sauers on X

The Rundown: Chinese lab Moonshot AI’s Kimi K3, which neared the frontier at launch last month, joined the sandbox escape club after U.S. firm Frontier Security revealed the model slipped through its test environment to pull an answer key off GitHub.

The details:

  • The benchmark's answer key sat in a public codebase, and Frontier says K3 realized it could reach GitHub through a loophole meant for software installs.

  • Nothing actually got hacked, and the sandbox wasn't wide open, but Frontier researcher Paul Kassianik said where rival models refused, “K3 didn't blink.”

  • K3’s weights are open and freely downloadable exactly as tested, which is why Frontier warns the case could prove "potentially more harmful."

Why it matters: The AI world cycles through a new main character every few months, and security breaches have officially taken the spot this summer. But this entry hits a bit differently than other escapees coming from closed labs like OAI, Anthropic, and Meta that can patch them, with Kimi’s weights already downloaded and out in the world.

QUICK HITS

🛠️ Trending AI Tools

  • 🗣️Unwrap Customer Intelligence - Connect your entire organization to the true voice of the customer with AI-driven insights from customer feedback*

  • 🎆 Grok Imagine Image 2.0 - xAI’s powerful new AI image model

  • 🎬 Seedance 2.5 - ByteDance’s video generation AI, now broadly available

  • 🪁 Kitesurf - Cloudflare’s lightweight, agent-first browser

*Sponsored Listing

📰 Everything else in AI today

Elon Musk revealed that Tesla and SpaceX’s ‘Terafab’ plant will be located in Grimes County, TX, predicting it will be “the largest and most valuable building on Earth by far”.

ByteDance is reportedly pre-training an AI with up to 10T parameters, which would triple the size of China's largest model to date and potentially rival Anthropic's Mythos.

xAI rolled out Grok Imagine Image 2.0, featuring upgraded editing, text rendering, and overall quality, with the model sitting behind OpenAI’s GPT Image 2 across benchmarks.

Research firm SemiAnalysis said DeepMind is "no longer a frontier lab", but projected Google Cloud growth to top 100% in 2027 as Google pushes TPU sales.

Anthropic introduced a Claude Code feature that allows the platform to message across sessions, allowing instances to pick up where another left off.

Google co-founder Sergey Brin is reportedly taking over oversight of the company’s Gemini model development, coming in the wake of last week’s leadership shakeup.

🎓 Highlights: News, Guides & Events

See you soon,

Rowan, Zach, Shubham, Jennifer, and Nate — the humans behind The Rundown

Tech

OpenAI builds a $400 AI donut

Jennifer Mossalgue • 5 minutes

Read Online | Sign Up | Advertise

Good morning, tech enthusiasts. OpenAI’s first device will be a donut-shaped, screenless smart speaker…and it’s arriving in 2027, with cameras, mics, sensors, and a light that tells you when it’s listening.

Amazon lost money on Alexa for years to build that market; OpenAI wants the top of it, on a timeline Apple’s home hub and a live trade-secrets suit both get a vote in.

In today’s tech rundown:

  • OpenAI’s answer to Alexa is a $400 AI donut

  • A SpaceX rocket slams into the Moon

  • The first mRNA flu vaccine arrives

  • Meet Fathom, Ford’s sub-$30K electric pickup

  • Quick hits on other tech news

LATEST DEVELOPMENTS

OPENAI

🍩 OpenAI’s answer to Alexa is a $400 AI donut

Image source: Images 2.0 / The Rundown

The Rundown: OpenAI is developing a $300–$400, donut-shaped AI speaker that could bring ChatGPT into the home as a portable, screenless companion in 2027, Bloomberg’s Mark Gurman reported.

The details:

  • Gurman says the device is donut-shaped and will be built from "high-quality metal" for a premium look, with distinct moving parts.

  • Price lands somewhere between $300 and $400 per unit, against an Amazon smart speaker line that runs from $40 to $240.

  • It's roughly hockey-puck-sized, one-handed to carry around the house, with cameras, microphones, and sensors, plus a light to show when it's listening.

  • LoveFrom is designing it, release is pegged for sometime in 2027, and Apple's trade-secrets suit against OpenAI is still live. OpenAI denies any wrongdoing.

Why it matters: An always-on ChatGPT that sees, hears, remembers context, and acts on it is a much bigger deal than anything Alexa has attempted. Amazon shipped Alexa devices for years at a loss, and Apple's home hub lands with iOS 27, which leaves OAI defining a category on a 2027 timeline it doesn't fully control.

SPACEX

🚀 A SpaceX rocket slams into the Moon

Image source: Images 2.0 / The Rundown

The Rundown: A SpaceX Falcon 9 upper stage that spent 18 months drifting off course hit the moon this week at 5,400 mph, leaving lunar orbiters to hunt for a fresh, human-made crater.

The details:

  • After delivering two commercial landers in January 2025, the burnt-out upper stage spent a year and a half wandering in an unstable orbit.

  • Impact came near Einstein Crater on the sunlit western limb, the hardest patch of the Moon to watch from Earth. No one filmed the collision.

  • ESO’s Very Large Telescope in Chile caught sodium and lithium spectral lines in a plume — sodium from lunar soil, lithium likely from the rocket.

  • NASA expects a 60 ft. wide and 12 ft. deep crater; South Korea's Danuri orbiter returned before-and-after frames, with further analysis currently underway.

Why it matters: Four tons of abandoned hardware hit the Moon, and the only reason anyone knew where to point a telescope was one independent astronomer running orbit fits on his own website. NASA’s take is that dumping spent upper stages on the lunar surface is technically safe, which could get trickier as lunar traffic grows.

BIOTECH

💉 The first mRNA flu vaccine arrives

Image source: Moderna / Images 2.0

The Rundown: Moderna’s mFlusiva has become the first FDA-approved mRNA flu vaccine in the U.S., opening a new front for the technology that reshaped the COVID-19 vaccine market.

The details:

  • The shot is approved to prevent influenza A and B in adults 50 and older, with conventional approval for ages 50 to 64 and an accelerated nod for 65 and up.

  • Phase 3 study enrolled 40,805 adults across 11 countries and found mFlusiva reduced confirmed flu-like illness by 26.6% relative to a standard-dose vaccine.

  • Side effects ran heavier than the comparator — injection-site pain, fatigue, headache, muscle ache — mostly mild and gone in a day or two.

  • The FDA initially refused to review the application because Moderna used a standard-dose comparator, but then accepted a revised filing two weeks later.

Why it matters: mRNA has moved from pandemic technology into the annual flu business. The U.S. distributes well north of 100M flu doses a year, and mFlusiva’s efficacy data removes a major roadblock for Moderna’s COVID-flu combo. The next test is whether it can take share from the flu vaccines that many Americans already buy.

ELECTRIC VEHICLES

🛻 Meet Fathom, Ford’s sub-$30K electric pickup

Image source: Ford

The Rundown: Ford has named its long-promised affordable electric pickup the Fathom, setting a $28,350 starting price for the first vehicle built on a new platform meant to make EVs cheaper (and profitable) at scale.

The details:

  • The midsize Fathom is scheduled to reach customers in fall 2027, with preorders and a full vehicle reveal planned for early that year.

  • Base price is $28,350 — $29,945 with destination and delivery, landing just under the $30K mark Ford has been promising since 2025.

  • Fathom is the first vehicle on Ford’s Universal EV Platform, built at the Louisville Assembly Plant on a new modular process Ford calls unicasting.

  • Five seats, a large touchscreen, a digital key, and bidirectional power. Apple Maps is built into the OS, and CarPlay and Android Auto are also supported.

Why it matters: Slate already beat Ford to the sub-$30K electric truck, but Slate is a startup with one Indiana plant selling a pickup that ships totally stripped down. Fathom is the first serious attempt by a legacy automaker to build a cheap EV at volume, and Ford has said the platform underneath it has to carry Model e to break even by 2029.

QUICK HITS

📰 Everything else in tech today

Hadrian raised $1.37B at a ~$8B valuation, betting that robot factories can help rebuild the U.S.’s capacity to produce weapons and aerospace hardware at speed.

A New Mexico judge ordered Meta to pay $567M and impose stricter teen-safety measures on Facebook and Instagram after finding it failed to protect children.

BMW is facing backlash after pushing promotional banners for the movie Spider-Man: Brand New Day to the infotainment displays of newer cars worldwide.

TikTok reportedly will lay off 250 employees and close its Nashville office — home to some content-moderation staff — on October 5.

Tesla and SpaceX will invest an initial $16.8B to build Terafab, a planned 100M-square-foot advanced chip-manufacturing complex in Grimes County, Texas.

Trevor Noah will host Made by Google 2026 on August 12 at 6 p.m. ET in NY, fronting a celebrity-heavy lineup including Stephen Curry, Alex Cooper, and Shubman Gill.

Nothing could reportedly launch as many as six new smartphones in 2027, suggesting an unusually aggressive expansion of its handset lineup.

Google Wallet is letting U.S. parents load and supervise a spendable balance for children and teens without opening them a bank account.

Nikita Bier stepped down as X’s head of product after just over a year, saying he will continue as an adviser while taking a break from the 24/7 role.

Ocean-powered data-center startup Subsea Cloud is reportedly eyeing a $2B valuation as investors bet on underwater infrastructure for AI’s cooling demands.

COMMUNITY

🎓 Highlights: News, Guides & Events

See you soon,

Rowan, Zach, Shubham, Jennifer, and Nate — The Rundown’s editorial team

AI

AI designs viruses never seen in nature

Rowan Cheung • 6 minutes

Read Online | Sign Up | Advertise

Good morning, AI enthusiasts, and welcome to the 5,515 new readers who joined us yesterday. Scientists just used AI to design viruses that don't exist in nature, then created them in the lab and watched them come to life and infect their targets.

The viruses are harmless — they only attack bacteria, with the goal of fighting drug-resistant infections. The worry is what happens if the same tools get repurposed, the way we've seen in cybersecurity, to invent something dangerous.

In today’s AI rundown:

  • AI designs working viruses from scratch

  • Rowan’s Corner: The best AI use cases aren’t coming from labs

  • Turn any idea into an AI-powered site with Lovable

  • Anthropic solves Fable 5’s biggest problem

COMMUNITY AI WORKFLOW OF THE DAY

▸ Charles built a self-paced course generator with AI

He built a reusable AI skill that turns any topic and source material into a self-paced course with lessons, verified YouTube videos, quizzes with explanations, all packaged in its own course folder.

Credit to reader Charles Clogston. How do you use AI? Tell us here for a chance to be featured next.  

LATEST DEVELOPMENTS

AI BIOLOGY

🦠 AI designs working viruses from scratch

Image source: Images 2.0 / The Rundown

The Rundown: Stanford and Arc Institute researchers just used AI to design 16 viruses not found in nature, marking the first instance of complete, working genomes generated by a language model, with the results published in the journal Science.

The details:

  • The team trained Evo 1 and Evo 2 on millions of genomes, then had them write new versions of Phi X174, a well-studied virus that infects only E. coli.

  • Of 285 phages synthesized and tested, 16 were viable, some replicating faster than the original, with a few different enough to count as new species.

  • A cocktail of AI-made viruses wiped out E. coli that had grown resistant to the natural one, a proof of concept for therapies for antibiotic-resistant infections.

  • For safety, the team never trained the model on viruses that infect humans, animals, or plants, so it can't generate anything that threatens people.

Why it matters: The AI that can design viruses to beat drug-resistant infections could, if trained differently, also help build a dangerous pathogen. Evo 2 is open source, and the field is moving fast, so expect more news in this field as well as mounting pressure to build the guardrails and testing frameworks for ensuring biosafety.

TOGETHER WITH ALGOLIA

🔎 The no-code approach to clean data

The Rundown: Our daily digital experiences are fueled by data, but to be effective, it needs to be clean and structured. This white paper explores how no-code, AI-powered data preparation helps teams create searchable, high-quality data faster.

You'll learn:

  • Why clean data matters

  • Limits of traditional methods

  • Benefits of modern no-code tools

  • How to turn messy data into valuable insights and a better search experience

Download the whitepaper.

ROWAN’S CORNER

🤗 The best AI use cases aren’t coming from labs

The Hub's weekly leaderboard, one week in

Rowan: In case you missed it, we just launched our highly requested Community AI Workflow Hub, a free extension of the Community Workflows section in our daily newsletter.

The reaction exceeded my expectations, and going through thousands of submissions and requests taught me a ton about the current state of AI adoption.

A rundown of my key learnings:

  1. People are done with hype. They don’t want to learn more about AI use cases from execs, influencers, or the media — they want to learn from real, everyday people, leveraging those use cases to do real work.

  2. Sharing how you use AI is now a hiring signal. We had several recruiters asking us to add location tags to posts so they could find and hire local builders.

  3. Step-by-step workflows beat inspiration. The most upvoted workflows are from builders who share their exact steps rather than just a high-level summary.

  4. People want to plug in, not just read. We had multiple API and MCP access requests in week one. For those asking, we hear you and are working on it!

To keep the momentum going, we’ll soon be highlighting top-voted builders not just in our newsletter, but across our socials: a weekly X thread, Instagram carousels, and LinkedIn shoutouts.

So if you’ve built a workflow that saves you hours, share it in the Hub, and you might just get featured in front of millions.

AI TRAINING

💖 Turn any idea into an AI-powered site with Lovable

The Rundown: In this guide, you will learn how to turn a rough app idea into a working AI-powered site without writing a giant technical prompt. Here, we’ll build Method, a recipe app that turns pasted recipe text into an editable visual card.

Step-by-step:

  • In ChatGPT, attach the recipe-card screenshot and describe the app. Answer its questions, but keep version one focused on paste, convert, review, and save

  • Ask ChatGPT to turn the conversation into feature requirements and a detailed brand brief for Lovable. Remove any extra features before copying the handoff

  • Create a Lovable project, attach the same screenshot, and paste the handoff specs from ChatGPT

  • In Connectors → Cloud AI, enable Lovable’s built-in AI. Now, you can test and refine your app

Pro tip: Once the core loop works, add accounts, data storage, and sharing. Build those features around a validated product, not an assumption.

PRESENTED BY ENEY BY MACPAW

🧑‍💻 A proactive AI assistant for focused work

The Rundown: Eney is a proactive AI assistant for Mac that handles repetitive tasks, keeps your work in context, and helps you stay focused by reducing context switching.

Eney is built to help you:

  • Start each day with a Morning Brief that surfaces what needs your attention

  • Work seamlessly across Gmail, Calendar, Notion, and Jira

  • Complete tasks automatically, with your approval before anything ships

Try Eney now, with no setup needed.

ANTHROPIC

🧬 Anthropic solves Fable 5’s biggest problem

Image source: Anthropic

The Rundown: Anthropic rewrote the safety classifier that handles Fable 5's biology responses, reducing false positives so the model now answers a far wider range of health and educational questions instead of routing them to the less capable Opus.

The details:

  • At launch, Fable 5 blocked almost all biology queries, even a ‘Hi’ from some researchers, and sent them to Opus — a deliberate safety filter from Anthropic.

  • The rewrite now cuts those fallbacks by ~85%, meaning Fable can now handle everyday tasks like understanding labs and symptoms, or learning biology.

  • However, it still isn't usable for research or drug development yet, with dual-use queries involving virology, toxicology, and molecular design going to Opus.

  • Anthropic is loosening Fable’s restrictions only as its classifiers improve, given its own tests show the model can uplift bad actors developing bioweapons.

Why it matters: Most people want exactly this from AI labs: basic use (like reading lab reports), but nothing that builds a bioweapon. It leaves one group frustrated, though — researchers doing real good, still blocked from the capabilities they need most. Anthropic knows it's turning them away but hopes to fix the gap with future updates.

QUICK HITS

🛠️ Trending AI Tools

  • 🔌 Agent Plugins - Vendor-neutral standard for Agent Skills and MCP servers

  • 📺️ Wan 3.0 - Alibaba's video AI with omni-reference support

  • 🗣️ Gemma Translator - Google's open-source on-device AI translation stack

  • 🧠 Muse Code - Meta's terminal-based AI coding agent

📰 Everything else in AI today

CData Connect AI can cut LLM context handling costs by up to 97.6% without sacrificing answers. Learn more here.*

OpenAI asked a judge to dismiss Apple's trade secret lawsuit, calling it a "baseless and pretextual" attempt to make up for its own AI and talent retention failures.

Meta’s Muse Spark 1.1 AI hacked another company's systems after a test misconfiguration gave it internet access, joining similar cases at OpenAI and Anthropic.

OpenAI is rolling out GPT-5.6 Luna as ChatGPT's default model for Free and Go users, with unlimited text chats and a new Think button for deeper reasoning.

DeepSeek warned developers that it will significantly raise API prices soon, urging them to plan usage ahead of the increase.

Google Maps updated Ask Maps with AI agents that can order food, compare hotels, and find event tickets, rolling out in the U.S. with more countries to follow.

*Sponsored Listing

🎓 Highlights: News, Guides & Events

See you soon,

Rowan, Zach, Shubham, Jennifer, and Nate — the humans behind The Rundown

Robotics

Musk's moon factories need robots

Jennifer Mossalgue • 6 minutes

Read Online | Sign Up | Advertise

Good morning, robotics enthusiasts. Elon Musk has found a new job market for humanoids: the Moon. On SpaceX’s first public earnings call, he sketched a future where humanoids build the factories, solar arrays, and launch systems that make rockets optional on the way out.

Musk has already cast Optimus as the first “crew” bound for Mars; now it looks like it’s being rerouted to lunar factory work.

In today’s robotics rundown:

  • SpaceX promises robot factories on the Moon

  • DoorDash hires gig workers to load delivery bots

  • Startup bids to replace banned Chinese humanoids

  • Xiaomi open-sources its robot brain

  • Quick hits on other robotics news

LATEST DEVELOPMENTS

SPACEX

🌝 SpaceX promises robot factories on the Moon

Image source: Images 2.0 / The Rundown

The Rundown: Elon Musk used SpaceX’s first earnings call as a public company to sketch out a robotics‑heavy future in which autonomous satellites and humanoid builders underpin “totally nuts” moon factories.

The details:

  • Musk pitched Starlink as “robots” operating autonomously in orbit, positioning SpaceX’s satellite fleet as a live PoC for large‑scale robotic infra in space.

  • He outlined a plan to deploy humanoids to build factories, solar arrays, and an electromagnetic mass driver that could launch cargo without rockets.

  • Musk acknowledged that using robots to scale lunar manufacturing “sounds super sci‑fi” but argued it is inevitable and central to SpaceX’s roadmap.

  • In parallel, he’s weaving terrestrial robotics into this picture — most notably Tesla’s Optimus and the TERAFAB joint venture — as the presumed workforce.

Why it matters: Musk has made SpaceX a prospective anchor customer for a class of robot nobody currently sells — while the stock fell more than 5% on the call. The gap between the lunar pitch and the disclosed spending is what engineers and shareholders will be measuring for the next several quarters.

TOGETHER WITH MONDO

🎥 Former DJI veterans build a camera robot

The Rundown: Beni, a groundbreaking autonomous camera robot by Mondo Robotics — founded by DJI veterans — has raised over $2M on Kickstarter, making it one of the year’s top robotics campaigns. Featuring high-speed, self-balancing mobility and intelligent tracking, Beni brings dynamic, automated filming to outdoor enthusiasts.

With Beni, you can:

  • Keep pace with the action at speeds up to 17.9 MPH and jump up to 10 inches

  • Capture real-time, multi-angle footage with cinematic auto-tracking

  • Record ultra-clear 4K HDR stabilized video with no camera operator needed

Learn more about Beni on Kickstarter.

DOORDASH

🥡 DoorDash hires gig workers to load delivery bots

Image source: DoorDash

The Rundown: DoorDash is offering drivers $5 to drive to a restaurant, collect an order, and hand-load it into a waiting Dot — the autonomous robot built to make that trip without them — a workaround now running in the Phoenix area.

The details:

  • Dashers told Business Insider that the gig pays about $5 for roughly five minutes: fetch the order, load the robot, photograph every step.

  • Dot navigates roads and sidewalks at up to 20 mph using lidar, radar, and cameras, but it cannot pick up a bag.

  • DoorDash calls it a limited pilot for rush periods and says humans will still handle most deliveries.

  • Waymo pays people to close robotaxi doors, and DoorDash, fresh off FAA clearance for drone delivery, admits the handoff is the tricky part.

Why it matters: For all the lidar and drone hardware, DoorDash still needs cheap human labor for the last few feet from restaurant to robot, adding a strange, fresh layer to the gig economy. The hard part of urban robotics is going beyond the AI stack to the messy reality of handoff at the end, and nobody has successfully solved that yet.

STEEL BOT

🤖 Startup bids to replace banned Chinese humanoids

Image source: Steel Bot / X

The Rundown: As the FCC moves to shut out new Chinese humanoids on national security grounds, Steel Bot says it is building a cheap, U.S‑made, fully programmable bipedal robot aimed at giving U.S. researchers and startups a homegrown platform.

The details:

  • Founder Randall Briggs, an MIT-trained roboticist, says Steel Bot has been building since last December and aims to ship to U.S. clients early next year.

  • He posted a manifesto on X that promises root-level control, open-source software, and modifiable designs.

  • The move comes as U.S. policymakers frame restrictions on foreign robots as both a national-security measure and a push to build a U.S. robotics industry.

  • K-Scale Labs shipped an $8,999 open-source U.S. humanoid last year but shut down with just $400K left after acquisition talks with 1X and The Bot Co.

Why it matters: Unitree reset humanoid pricing, but the FCC just put a wall around the next generation of foreign-made rivals. That leaves a rare opening for a cheap, hackable American platform, and Steel Bot wants to fill it. K-Scale proved there's demand. The hard part is raising enough money to turn it into a scalable business.

XIAOMI

🧠 Xiaomi open-sources its robot brain

Image source: Xiaomi

The Rundown: Xiaomi open-sourced Xiaomi-Robotics-1, a vision-language-action model pretrained on 100K+ hours of real-world manipulation data, releasing weights, the full real-robot post-training pipeline, and benchmark evaluation code.

The details:

  • The model learned from 100K hours of footage across 1,700+ scenarios, then got aligned to actual robots with 10K+ hours of cross-embodiment data.

  • The release also covers the post-training pipeline, deployment tooling, and evaluation code, hosted across a project site, GitHub, and Hugging Face.

  • On new tasks — phone packing, printer refilling, laundry loading — it reports 75% success from under 10 hours of demos each.

  • First teased in July as an out-of-the-box controller, Robotics-1 claims state-of-the-art on four simulation benchmarks, with the biggest margin on RoboDojo.

Why it matters: Open weights give labs a way to skip the most expensive part of the process: building a 100K-hour pretraining corpus and training from scratch. Xiaomi’s release is not plug-and-play, but it puts an unusually complete training-to-deployment stack into the open and raises the disclosure bar for better-funded closed programs.

QUICK HITS

📰 Everything else in robotics today

Unitree is chasing a $7.4B valuation in its Shanghai IPO — humanoid robotics’ first public price tag — just as Washington tightens restrictions on Chinese-made robots.

China banned six U.S. entities and imposed case-by-case review on drone exports to the U.S., answering FCC import bans and a DHS forced-labor listing.

Chinese robot maker AI² Robotics is working with advisers on a possible Hong Kong IPO as early as next year to tap investor demand for the fast-growing robotics sector.

The U.S. signed a deal of up to $820M to build domestic manufacturing capacity for components it currently buys from foreign-controlled supply chains.

Uber and Wayve won Transport for London licenses to run supervised autonomous rides in sensor-equipped Mustang Mach-Es this summer.

YC-backed IMPACT Drones launched air defense by subscription: sealed boxes of GPS-independent interceptors that ram hostile drones, with pilots starting in Q3 2026.

Lucid is reportedly cutting $1.4B in cash while delaying its sub-$50K midsize EV to 2027 and pinning profitability on an Uber-Nuro robotaxi program.

Gautam Gupta, Uber’s finance chief until 2017, is joining Travis Kalanick’s robotics and industrial AI startup Atoms as CFO, just weeks after the company raised $1.7B.

A study of 50 young men conversing with the expressive humanoid Pepper found that when it blundered, participants’ oxytocin rose while trust in the robot nosedived.

Reimagine Robotics, founded by DeepMind alum Jonathan Scholz, emerged from stealth with a platform that lets non-experts teach robots tasks on the fly.

Waymo dropped the Dallas waitlist Tuesday, opening its robotaxi fleet to anyone with the app after roughly 150K riders used it during the six-month gated rollout.

Avatar Robotics, an SF startup that uses remote human teleoperators in VR headsets to control warehouse robots, raised a $6.5M seed round.

London’s St. Bartholomew’s Hospital is training NHS surgeons on Intuitive's da Vinci 5 as England targets robotic assistance in nine of 10 keyhole surgeries by 2035.

Anduril is in talks to invest hundreds of millions of dollars to turn part of Baltimore’s historic Sparrows Point shipyard into a major testing hub for unmanned drone boats.

COMMUNITY

🎓 Highlights: News, Guides & Events

See you soon,

Rowan, Zach, Shubham, and Jennifer — The Rundown’s editorial team

No matching search results

Try using different keywords, double-check your spelling, or explore related categories.

Clear Search

Stay Ahead on AI.

Join 2,000,000+ readers getting bite-size AI news updates straight to their inbox every morning with The Rundown AI newsletter. It's 100% free.

Subscribe