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.
TikTok U.S.A.'s epic meltdown
Read Online | Sign Up | Advertise
Good morning, tech enthusiasts. TikTok’s new U.S.-controlled era is off to a rough start: broken For You feeds, stalled uploads, and posts vanishing mid-scroll.
Now, adding to the trouble, California is probing whether politically sensitive content is being censored. TikTok blames a U.S. data-center power outage — an unglamorous culprit for an app built on viral momentum.
In today’s tech rundown:
TikTok’s U.S.-owned debut is already a mess
Grok under fire as EU probes deepfake tools
Meta plans subscription tiers across its apps
Finland wants your burned-out engineers
Quick hits on other tech news
LATEST DEVELOPMENTS
TIKTOK
😱 TikTok’s U.S.-owned debut is already a mess
Image source: Ideogram / The Rundown
The Rundown: TikTok under U.S. control was already glitching out — busted feeds, stalled uploads, videos hitting zero views — and now California has opened a formal investigation into whether the app is quietly throttling political content.
The details:
U.S. users complained of repetitive or wildly off-base recommendations, missing comments, and new posts disappearing into the void.
California Governor Gavin Newsom is launching a review into whether TikTok is violating state law by censoring Trump-critical content.
But TikTok says its tech issues were triggered by a power outage at a U.S. data center and that it's working to fully restore services.
Sensor Tower data shows TikTok’s U.S. daily uninstalls have surged nearly 150% in the five days since the TikTok USDS joint venture was announced.
Why it matters: Just days after ByteDance ceded majority control to U.S. investors — including Trump ally Larry Ellison’s Oracle — TikTok is stumbling through tech chaos and accusations of political censorship. Rivals like Skylight are eager to absorb the spillover, but unseating an app with 200M U.S. monthly actives won’t be easy.
TOGETHER WITH OPTIMIZELY
⚡ AI agents for marketing workflows
The Rundown: Most marketing teams don’t need more ideas — they need more time. Optimizely Opal is an AI agent orchestration platform that handles real workflows like content creation, compliance checks, experiment variations, and reporting, so your team can focus on work that moves the needle.
With Opal, you can:
Automate content creation, approvals, and compliance checks
Generate experiment variations and personalize experiences at scale
Analyze pages for GEO and streamline reporting
Keep everything on-brand while plugging into tools you already use
Apply now to get a free agentic workshop for your marketing team (plus a pair of Ran-Ban X Meta glasses).
GROK
🔎 Grok under fire as EU probes deepfake tools
Image source: Reve / The Rundown
The Rundown: The European Commission just opened a formal probe into xAI and X over Grok’s role in generating non‑consensual sexualized deepfakes of women and children, calling such content a “violent, unacceptable form of degradation.”
The details:
Regulators are investigating Grok’s image-generation tools, which enabled users to “digitally undress” real people and share explicit images at scale.
The inquiry will examine whether X assessed risks before deploying Grok, and whether its recommender systems amplified harmful deepfake content.
X says it has tightened Grok’s image tools and restricted some capabilities following backlash.
The EC has ordered X to preserve all Grok‑related records through 2026 and will assess whether these fixes meet the Digital Services Act’s requirements.
Why it matters: In December, the EU hit X with a $140M fine over deceptive blue-check practices, a penalty CEO Elon Musk called “crazy.” The investigation also follows similar scrutiny from regulators in the UK, California, and countries including Malaysia and Indonesia that have moved against Grok’s deepfake capabilities.
META
💎 Meta plans subscription tiers across its apps
Image source: Ideogram / The Rundown
The Rundown: Meta is gearing up to test premium subscription tiers across Instagram, Facebook, and WhatsApp, according to TechCrunch, offering users optional access to advanced features and AI tools while keeping core functionality free.
The details:
Each app will get its own subscription bundle, with Instagram early perks including unlimited audience lists and anonymous Story viewing.
Meta plans to scale its $2B Manus AI agent acquisition by integrating it into its apps while still selling Manus as a standalone subscription to businesses.
The company will also test paid access to AI features like Vibes, an AI-powered short‑form video tool, shifting it from free to a freemium model.
Meta is following Snapchat’s playbook — Snapchat+ doubled subscribers to 14M in 2024, generating some $500M in annual revenue from AI features.
Why it matters: Meta’s AI subscriptions would create a two-tier social web where paying users unlock stronger tools, fewer limits, and agentic AI. By wiring Manus into those paid tiers — while still selling it separately to enterprises — Meta is betting that AI agents can overcome subscription fatigue and help recoup its massive AI spend.
FINLAND
🎿 Finland wants your burned-out engineers
Image source: …
The Rundown: Finland is pitching overworked U.S. tech and AI workers a tempting deal: ditch the grind for a fast-track visa, saner hours, and a healthcare system that won’t bankrupt you, as spotted by Business Insider.
The details:
The program has partnered with more than 30 companies and universities, with job listings at Oura Health and quantum computing startup IQM.
A Fast Track specialist visa can deliver a combined work-residence permit in as little as two weeks, and spouses get work permits too.
Finnish officials are marketing the country as an antidote to U.S. tech culture: offices are actually empty at 5 p.m., and overtime is tightly regulated by law.
Salaries run lower than U.S. rates, but expats say affordable childcare, universal healthcare, and generous vacation time make up the difference.
Why it matters: Finland is bundling fast-track visas, spousal work rights, and integration support into a soft-landing package designed to lure deep-tech talent away from Silicon Valley. The pitch is simple: trade some compensation for a safer, calmer, higher-quality life. Learning Finnish and surviving the winters? That part’s on you.
QUICK HITS
📰 Everything else in tech today
Microsoft received approval to build 15 additional data centers in Mount Pleasant, Wisconsin, to boost capacity for AI and cloud workloads.
Google will pay $68M to settle a class-action lawsuit claiming Google Assistant recorded users after accidental activations and shared some audio without consent.
Apple launched a new AirTag, adding a louder speaker, longer range, and expanded Precision Finding support so items can be located from much farther away.
Chinese EV giant BYD is aiming to sell 1.3M vehicles outside China in 2026, about a 25% jump from last year’s 1.05M overseas deliveries.
YouTubers behind h3h3 and two golf channels are suing Snap, claiming it illegally scraped their videos to train AI for tools like Snapchat’s Imagine Lens.
Y Combinator dropped Canada from its investment list, so Canadian startups must reincorporate in the U.S., Cayman Islands, or Singapore to join the accelerator.
South Korean startup Edenlux is set to launch Eyeary, smart glasses that retrain eye muscles to ease screen-induced eye strain, in the U.S. via Indiegogo in March.
New court documents in a lawsuit show Google has long viewed Chromebooks as a way to hook students into its ecosystem and turn them into lifelong customers.
Chinese EV maker Xpeng aims to deliver more than 90K vehicles outside China this year, accelerating its expansion in Europe, the Middle East, and Southeast Asia.
COMMUNITY
🎓 Highlights: News, Guides & Events
Read our last AI newsletter: Anthropic CEO confronts AI's dangers
Read our last Tech newsletter: Amazon’s second axe is falling
Read our last Robotics newsletter: Musk’s Optimus deadline
Today’s AI tool guide: Build your own design system for AI
RSVP to next workshop on Feb. 11: Agentic Workflows Bootcamp pt. 1
See you soon,
Rowan, Joey, Zach, Shubham, and Jennifer — The Rundown’s editorial team
Anthropic CEO confronts AI's dangers
Read Online | Sign Up | Advertise
Good morning, AI enthusiasts. Which way, humankind?
That's the question at the heart of Dario Amodei's new essay, which argues the next few years will decide whether AI delivers a golden age or spirals toward something darker — and that the tech’s economic promise is a trap humanity can't resist.
In today’s AI rundown:
Anthropic CEO on AI’s ‘civilizational’ danger
Anthropic embeds interactive apps inside Claude
Build your own design system for AI
Microsoft launches powerful Maia 200 AI chip
4 new AI tools, community workflows, and more
LATEST DEVELOPMENTS
DARIO AMODEI
📖 Anthropic CEO on AI’s ‘civilizational’ danger
Image source: Lovart / The Rundown
The Rundown: Anthropic CEO Dario Amodei just published “The Adolescence of Technology,” a new essay that lays out what he sees as the biggest dangers of AI, from bioterrorism and autonomous weapons to mass job loss and AI-powered dictatorships.
The details:
The essay builds on his 2024 “Machines of Loving Grace”, but pivots to risk — framing AI as a "country of geniuses in a data center" that we can't control.
Amodei predicts half of entry-level office jobs are at risk over the next 1-5 years, with economic shocks arriving faster than society can adapt.
He calls for chip export bans and more transparency from labs, saying AI's economic promise makes restraint "very difficult for human civilization."
Amodei also flags AI companies themselves as a tier of risk, noting Claude exhibited deception and blackmail behavior during internal safety testing.
Why it matters: Amodei’s essays are always a must-read, and while ‘Machines of Loving Grace’ outlined the optimistic end of AI’s spectrum, he follows it up with a polar opposite document that doesn’t hold back — arguing that the next few years will determine whether humanity navigates to an AI-powered golden age or destruction.
TOGETHER WITH YOU.COM
💡 One major reason AI adoption stalls? Training
The Rundown: AI implantation often goes sideways due to unclear goals and lack of a clear framework. You.com’s checklist pinpoints common pitfalls and guides you to build a capable, confident team that can make the most out of your investments.
Inside, you’ll get:
Key steps for building a successful AI training program
Guidance on overcoming employee resistance and fostering adoption
A structured worksheet to monitor progress and share across your organization
ANTHROPIC
🚀 Anthropic embeds interactive apps inside Claude
Image source: Anthropic
The Rundown: Anthropic just launched interactive apps inside its Claude AI assistant, allowing users to connect and use apps like Asana, Slack, Canva, and more without ever leaving the chat window.
The details:
The initial rollout includes nine apps: Asana, Figma, Canva, Slack, Box, Amplitude, Clay, Hex, and Monday, with Salesforce coming soon.
The integrations run on MCP Apps, a new open extension to Anthropic's Model Context Protocol — meaning other AI platforms can build the same functionality.
Actions require user consent prompts before executing, and enterprise admins can lock down which tools employees have access to.
The feature ships to Pro, Max, Team, and Enterprise subscribers at no extra cost, with a Cowork integration also coming in the future.
Why it matters: OAI rolled out a similar Apps SDK at Dev Day, and all the frontier labs are pushing to turn their platforms into a workplace layer that sits on top of every other highly-used app. With the importance of context to getting the most out of AI, every popular app and platform will be just an integration away from your preferred assistant.
AI TRAINING
🎨 Build your own design system for AI
The Rundown: In this guide, you will learn how to use an underrated design tool with Codex to build a simple portfolio website — a workflow that simplifies your project setups so you won’t have to worry about design systems ever again.
Step-by-step:
Navigate to ui.shadcn.com/create, use the drop-down to create a custom design system, then click Create Project and copy the unique command
Open a terminal, run the command to scaffold the app, then navigate into the folder and run
npm installandnpm run devto launch the site locallyOpen Codex / Claude in the directory and paste design system instructions (found via link in ShadCN command) into your Claude.md or Agent.md file
Plan your site in a short PRD, then prompt the agent to create a simple portfolio website using the attached PRD and the defined design system
Pro tip: If you have the Context7 Skill installed, you can put "Always use context7 to look for ShadCN components before building new UI" in your Agent.md.
PRESENTED BY GLEAN
🤖 Scale impact with AI agents
The Rundown: See how enterprises are moving beyond chat-based assistants to AI agents that understand organizational context, take action across tools, and reliably complete real work. This report distills where to start, which use cases to prioritize, and how to keep agents safe at enterprise scale.
In this report, you’ll learn:
Why context is the foundation for agents
How to democratize agent building across teams
Guardrails that keep enterprise automations safe
MICROSOFT
⚡ Microsoft launches powerful Maia 200 AI chip
Image source: Microsoft
The Rundown: Microsoft just debuted Maia 200, the company’s newest in-house AI chip that it says beats rivals from Amazon and Google on key benchmarks — while also chipping away at Nvidia's software grip on the industry.
The details:
Microsoft claims the chip outperforms Amazon’s Trainium 3 and Google's TPU v7, with 30% better efficiency than its current hardware.
The chip will power OAI's GPT-5.2 models, Microsoft's internal AI teams, and Copilot across its product lineup starting this week.
Microsoft is also releasing an SDK preview, developer tools that rival Nvidia’s industry-standard software, in a move to loosen the AI chip giant’s moat.
Why it matters: Google and Amazon were already pushing Nvidia to carve out a piece of the custom AI chip market, and now Microsoft is getting its own next-gen chips into the competition. The tech giant’s SDK also targets the software side, hitting at a CUDA moat that is considered one of Nvidia’s biggest competitive advantages.
QUICK HITS
🛠️ Trending AI Tools
🔐 Incogni - Erase sensitive data like addresses and phone numbers from the web. Get 55% off with code RUNDOWN.*
🔌 Interactive Apps - Use popular business apps directly within Claude
🧠 Qwen3-Max-Thinking - Alibaba’s new flagship reasoning model
🤖 MiniMax Agent - AI assistant with browser control, expert agents, more
*Sponsored Listing
📰 Everything else in AI today
Cisco AI Summit streams live on Feb. 3, featuring leaders from NVIDIA, OpenAI, AWS, Figma, Google, and Andreessen Horowitz to discuss the infrastructure and governance of the AI economy.*
OpenAI will reportedly charge around $60 / thousand ad views in its initial ChatGPT rollout, rivaling primetime broadcast rates and tripling the likes of Meta.
Alibaba’s Qwen released Qwen3-Max-Thinking, the company’s new reasoning model that is competitive with models like Claude 4.5 Opus, GPT 5.2 Pro, and Gemini 3 Pro.
Nvidia announced a $2B investment in CoreWeave as part of an expanded partnership to build over 5 GW of AI data center capacity by 2030.
Synthesia raised $200M in new funding at a $4B valuation led by Google Ventures, with plans to build AI agents that help train and upskill employees with interactive video.
*Sponsored Listing
COMMUNITY
🤝 Community AI workflows
Every newsletter, we showcase how a reader is using AI to work smarter, save time, or make life easier.
Today’s workflow comes from reader Robert M. in Ontario, Canada:
"I am an author, and... I recently uploaded my book to ChatGPT and asked it to plot out the chapters for the next book. I included the location, characters, and basic premise.
ChatGPT came up with 10 chapter headings and brief outlines. I then took each outline and wrote the chapter based on the outline, changing when necessary. I wrote the book in two weeks instead of the usual two months. It's still me writing the book: my story, my language. But having AI come up with a plot sequence saved me hours."
How do you use AI? Tell us here.
🎓 Highlights: News, Guides & Events
Read our last AI newsletter: Claude for Excel opens the gates
Read our last Tech newsletter: Amazon’s second axe is falling
Read our last Robotics newsletter: Musk’s Optimus deadline
Today’s AI tool guide: Build your own design system for AI
RSVP to next workshop on Feb. 11: Agentic Workflows Bootcamp pt. 1
See you soon,
Rowan, Joey, Zach, Shubham, and Jennifer — the humans behind The Rundown
Musk's Optimus deadline
Read Online | Sign Up | Advertise
Good morning, robotics enthusiasts. Elon Musk finally showed up at Davos to pitch a near-future where humanoids outnumber humans and “saturate all human needs” before the decade’s out.
It’s classic Musk futurism: utopia-by-deadline while quietly dodging the biggest question of all: what’s left for humans when the robots run the show?
In today’s robotics rundown:
Musk: Optimus to go on sale in 2027
Union slams Hyundai’s Atlas push
Tesla’s Austin driverless robotaxis are live
These swarm robots bloom like flowers
Quick hits on other robotics news
LATEST DEVELOPMENTS
ELON MUSK
🤖 Musk: Optimus to go on sale in 2027
Image source: Ideogram / The Rundown
The Rundown: Elon Musk used his first appearance at the World Economic Forum in Davos to sketch a future where humanoids outnumber humans in factories and “saturate all human needs” by the end of the decade.
The details:
Musk forecasts AI could outsmart any individual human by late 2026, and outstrip humanity collectively by 2030–2031.
He said he expects Tesla to begin selling Optimus humanoids to the public by the end of 2027.
Optimus is already handling basic factory tasks; Musk says they’ll tackle complex industrial work within a year and expand to consumer uses soon.
He framed robot abundance as utopia, not dystopia — tossing in that aging is “a very solvable problem” too.
Why it matters: The pitch was classic Musk — aggressive timelines, grand ambition, plus jokes about being an alien and Trump eyeing “a little piece of Greenland, a little piece of Venezuela.” What's missing: any serious answer to what ageless humans actually do in a world where robots handle everything.
HYUNDAI & BOSTON DYNAMICS
🪧 Union slams Hyundai’s Atlas push
Image source: Hyundai (Atlas demo at CES)
The Rundown: Hyundai’s South Korean union is vowing to block the automaker’s plan to roll out Boston Dynamics’ Atlas humanoids on production lines — starting with a new U.S. plant in Georgia — unless a formal deal is struck to protect jobs.
The details:
The union argues that Atlas and other AI robots are explicitly aimed at cutting labor costs, warning that they will trigger “significant employment disruption.”
Hyundai is building a new facility capable of producing up to 30K Atlas robots a year by 2028, positioning them as a core tool to boost manufacturing.
The first units are slated to roll out at Hyundai’s Georgia plant in 2028, initially handling basic tasks before moving into more complex work by 2030.
Maintaining an Atlas robot is projected to cost about $9,500 per year, far below the typical annual cost of human labor.
Why it matters: Hyundai — the world’s third largest automaker — claims Atlas will shoulder dangerous tasks and spare workers’ backs. The union sees a different calculation: swap humans for hardware, bank the savings. They’re also pushing back on the automaker’s overseas expansion plans, including new U.S. production.
TESLA
🚖 Tesla’s Austin driverless robotaxis are live
Image source: Tesla
The Rundown: Tesla just started offering “no front-seat safety driver” robotaxi rides in Austin, sending a small fleet of unsupervised vehicles onto public streets as Musk pitches the launch as a breakthrough moment for Tesla’s AI ambitions.
The details:
The safety net reportedly hasn’t vanished entirely: videos show Tesla’s Austin robotaxis being shadowed by human-driven Teslas as backup support.
Only a fraction of the local fleet is running in this unsupervised mode for now, with the rest of Tesla’s robotaxis still operating with human safety monitors.
Elon Musk is using the launch to recruit engineers, framing Tesla’s autonomy stack as “real-world AI” work that he claims could “likely lead to AGI.”
Early riders report that Tesla is charging for these trips from day one, unlike rivals Waymo and Zoox, which initially offered fully driverless rides for free.
Why it matters: Tesla is racing to own the robotaxi narrative just as Waymo is pushing toward broader, regulator-approved driverless service in more cities. A rollout signals confidence in the tech, but chase cars trailing behind suggest the “unsupervised” label comes with an asterisk, for now at least.
ROBOTICS RESEARCH
🌼 These swarm robots bloom like flowers
Image source: Princeton University
The Rundown: Princeton engineers published a new paper on a swarm of robots that “bloom” like flowers in response to changing light, prototyping building facades that can collectively sense light, move, and adapt in real time.
The details:
The installation, first built in 2024, uses about 40 small modular “SGbots” whose flexible petals open and close in coordinated patterns.
Each robot packs its own light sensor, actuator, processor, and wireless link, letting the swarm decide how to move without a central controller.
In real-world tests, subsets of the swarm adjusted to shifting sunlight over hours and kept functioning even when some units went dark.
A gallery version linked the swarm to a dancer’s wearable interface, with the facade blooming in response to human movement.
Why it matters: The project slots into a broader push toward “living” architecture — robot gardens using swarms of simple machines to build adaptive facades and kinetic urban art. By turning facades into self-organizing robot swarms, it suggests a future architecture that’s part climate system, part public art, and always in motion.
QUICK HITS
📰 Everything else in robotics today
Waymo opened its driverless robotaxi service to the public in Miami, starting with about 10K waitlisted residents who can hail rides across a 60-square-mile service area.
Gecko Robotics, MIT, and Mech-Mind experts told Business Insider that humanoids remain far from scalable real-world deployment, held back by limited sensing and AI.
Carnegie Mellon is developing a quadruped robot that can sniff out hazards, locate casualties, and assess victims’ conditions in search-and-rescue operations.
A Chinese university unveiled Fuxiaozhi F1-D, a humanoid treatment robot that uses a non-invasive brain-computer interface to support autism intervention in children.
A Financial Times report on Shenzhen-based UBTech says its Walker S2 humanoids are only about half as productive as human workers.
Roomba maker iRobot emerged from Chapter 11 as a private Picea Robotics subsidiary, launching a U.S. “iRobot Safe” unit to handle consumer data protection.
Beijing startup InterstellOr plans to fly its first suborbital space tourists in 2028 and has booked EngineAI’s PM01 humanoid as an experimental “passenger.”
Chinese heavy equipment giant Zoomlion is deploying humanoids in its “Smart City” factories to help assemble excavators at a rate of one every six minutes.
Chinese researchers unveiled GrowHR, a lightweight, bone-inspired humanoid that can move by floating, swimming, walking on water, and even briefly flying.
China’s new five-year tech plan puts humanoids at the center of its industrial strategy, setting targets for major robotics revenue growth by 2030.
COMMUNITY
🎓 Highlights: News, Guides & Events
Read our last AI newsletter: Claude for Excel opens the gates
Read our last Tech newsletter: Amazon’s second axe is falling
Read our last Robotics newsletter: A crawling, detachable robotic hand
Today’s AI tool guide: Create ads and marketing videos with Remotion
RSVP to next workshop on Feb. 11: Agentic Workflows Bootcamp pt. 1
See you soon,
Rowan, Joey, Zach, Shubham, and Jennifer — The Rundown’s editorial team
Claude for Excel opens the gates
Read Online | Sign Up | Advertise
Good morning, AI enthusiasts. Developers got Claudepilled first, and now the spreadsheet crowd is next.
Anthropic just expanded Claude for Excel access — and while Code and Cowork have dominated the hype cycle, this one lands directly in the workflows of virtually every knowledge worker on the planet.
In today’s AI rundown:
Anthropic expands Claude for Excel access
The Rundown Roundtable: Our AI use cases
Create ads and marketing videos in minutes with Remotion
Gallup: Half of U.S. workers never use AI on the job
4 new AI tools, community workflows, and more
LATEST DEVELOPMENTS
ANTHROPIC
📊 Anthropic expands Claude for Excel access
Image source: Anthropic
The Rundown: Anthropic just expanded access to its Claude for Excel integration, making the AI spreadsheet assistant available to Pro-tier customers after a three-month beta period that was limited to Max and Enterprise plans.
The details:
Claude for Excel initially launched in October in research preview, bringing the assistant directly into a text sidebar of Microsoft’s spreadsheet platform.
Users can now pull in several spreadsheets at once and work longer without hitting session limits thanks to behind-the-scenes memory management.
The update also adds safeguards to preserve existing cell contents, so Claude won't accidentally overwrite data while making edits.
Why it matters: 2026 has been the year of getting ‘Claudepilled’ — with Code, Cowork, and now a broader Claude for Excel release going down the line of taking on major knowledge work. Just like coding is becoming completely automated, the days of needing to know complex formulas and shortcuts are also coming to an end.
TOGETHER WITH CONCIERGE
👋 Your brand's AI answer engine
The Rundown: Today’s buyers use AI every day to answer their questions, and have no patience for a scavenger hunt on your website. Concierge is a custom Perplexity-style answer engine, trained on your company’s brand & content, that delivers accurate, personalized responses to any questions your website visitors have.
Modern B2B brands use Concierge to:
Handle any buyer question (no matter how technical) with advanced RAG on your content, media, and documentation.
Control and visibility over every conversation, with guardrails and sentiment analysis.
Build trust with website visitors before they are willing to commit to a demo.
Use Concierge to turn every question into an opportunity.
THE RUNDOWN ROUNDTABLE
💡The Rundown Roundtable: Our AI use cases
Image source: Ideogram / The Rundown
The Rundown: The Rundown Roundtable is a weekly feature in which we poll members of The Rundown staff about how we use AI in our work and daily lives.
Rowan, CEO & Founder: I spend way too much time staring at screens, so I try to consume my content by listening while I’m out walking (usually podcasts).
Lately, I’ve been digging into more long-form stuff from X articles and old blogs, and ElevenLabs’ free app, ElevenReader, turns any wall of text into a mini audiobook using AI-generated voices. It’s super simple, but genuinely really useful.
Rishi, Head of Growth: Claude Cowork has been a huge help in creating growth forecasts for our paid media channels. I just plug in our metrics, tell it what goal I'm trying to achieve, and it's able to create an Excel sheet for me with exactly what I need. This saves me a bunch of time because I can create multiple scenarios just by chatting with Cowork.
Shubham, Editor: My sister asked me to check if my niece's name appeared in the preliminary school enrollment results. But the list uploaded online was a PDF with 50+ page photographs. Going through all of it would have been a headache — so I put Manus to work. I provided a link to the results, and Manus searched through all of it, checking each document page by page.
After about 15 minutes, it found her name and all the relevant details. This saved me from having to manually search through pages of results across multiple categories.
AI TRAINING
🎥 Create ads and marketing videos in minutes with Remotion
The Rundown: In this tutorial, you will learn how to programmatically generate ads and marketing videos with code — enabling you to conceptualize, template, and generate dozens of videos in minutes.
Step-by-step:
Open NotebookLM, add pictures of your website/product as sources. Tell it to "Interview me about my brand to generate 5 concepts for promotional videos”
Tell NotebookLM to write a PRD for creating each concept using Remotion
Create a Remotion project by running
npx create-video@latest, install dependencies, start the editor withnpm installandnpm run dev, then initialize Git and place the PRD markdown in the project root for contextPrompt Codex to read PRD, generate components, and create the video; preview it at
http://localhost:3000/and export via the rocket icon
Pro tip: Add royalty-free mp3s to your project and tell Codex to include them via Remotion, with music and sound effects baked into your video scripts for faster reuse.
PRESENTED BY IBM
⚡ Enterprise 2030, AI-First Advantage
The Rundown: IBM’s latest report reveals five predictions that will help to shape the future of business. By 2030, success will hinge on speed, bold bets, and AI-first design. Learn how leaders can hard-code transformation into every process and prepare for quantum disruption.
Big bets win: speed beats perfection.
Tailored AI portfolios: build unique advantage.
Quantum readiness: next seismic shift is coming.
AI RESEARCH
💼 Gallup: Half of U.S. workers never use AI on the job
Image source: Gallup
The Rundown: Gallup published its ‘AI in the Workplace’ report for Q4 2025, showing that workplace AI adoption is starting to plateau, with nearly half of American workers saying they never touch AI tools, even as frequent users are deepening their habits.
The details:
Tech workers lead adoption with 60% using AI frequently, though growth has begun to stall after surging between 2024 and 2025.
Finance, education, and professional services follow tech, while retail and manufacturing trail significantly in workplace AI use.
Remote-capable roles are pulling away with 66% adoption and 40% frequent use, compared to just 32% and 17% for jobs that require physical presence.
Leaders are using AI far more than their teams (69% vs. 40% for individuals), with Gallup citing "lack of clear utility" as the main barrier for non-adopters.
Why it matters: AI’s mainstream arrival has completely reshaped certain areas of the workforce, but the gains aren’t being evenly spread. This gap is only going to continue to expand. But those who DO adopt AI in sectors currently behind the curve are going to sprint ahead and create more intense competition than ever before.
QUICK HITS
🛠️ Trending AI Tools
⚙️ Tasklet — AI agent that connects to any app or API, triggers automatically 24x7, and uses a computer to get work done*
📊 Claude in Excel - Use Claude to create and edit spreadsheets
🗣️ Qwen3 TTS - Alibaba’s family of open-source, SOTA text-to-speech models
🧠 Slackbot - Slack’s personal AI agent for work
*Sponsored Listing
📰 Everything else in AI today
KREA introduced Realtime Edit, a new feature allowing users to edit images via text prompts and make granular changes in real-time.
Cursor rolled out version 2.4, introducing subagents that break coding tasks into workstreams for faster execution, alongside image generation and custom agent skills.
Epoch AI announced that OpenAI’s GPT-5.2 Pro set a new record on its hardest FrontierMath benchmark at 31%, nearly doubling the previous high.
Japanese lab Sakana AI secured a strategic partnership and investment from Google, with plans to deploy AI solutions in Japan’s regulated industries using Google’s models.
CEO Sam Altman revealed that OpenAI will be releasing new Codex launches this week, saying its models will soon reach the highest cybersecurity capability level.
COMMUNITY
🤝 Community AI workflows
Every newsletter, we showcase how a reader is using AI to work smarter, save time, or make life easier.
Today’s workflow comes from reader Justo I. in Spain:
"I use Comet (Perplexity’s AI-powered browser) to automate the workflow in our sports company by reviewing support emails requesting rating changes, extracting key player details, and then logging into our admin panel to find the player by email and update their rating accordingly."
How do you use AI? Tell us here.
🎓 Highlights: News, Guides & Events
Read our last AI newsletter: Inside the Thinking Machines meltdown
Read our last Tech newsletter: Amazon’s second axe is falling
Read our last Robotics newsletter: A crawling, detachable robotic hand
Today’s AI tool guide: Create ads and marketing videos with Remotion
RSVP to next workshop on Feb. 11: Agentic Workflows Bootcamp pt. 1
See you soon,
Rowan, Joey, Zach, Shubham, and Jennifer — the humans behind The Rundown
Amazon's second axe is falling
Read Online | Sign Up | Advertise
Good morning, tech enthusiasts. Amazon waited until the holidays were over. Now the axe is falling again.
The e-commerce giant is reportedly preparing to cut thousands more corporate jobs as early as next week, part of CEO Andy Jassy’s mission to slash roughly 30K roles. But don’t blame AI, he says — this is more about trimming corporate bloat.
In today’s tech rundown:
Amazon’s second layoff wave hits next week
Lab‑grown skin steps up after Swiss ski‑resort fire
Musk taps top banks for SpaceX’s monster IPO
Blue Origin to launch satellite megaconstellation
Quick hits on other tech news
LATEST DEVELOPMENTS
AMAZON
🪓 Amazon’s second layoff wave hits next week
Image source: Ideogram / The Rundown
The Rundown: Amazon is reportedly preparing its second wave of sweeping layoffs as early as next week, cutting thousands more corporate jobs as CEO Andy Jassy pushes to eliminate roughly 30K white-collar roles.
The details:
The company in October cut some 14K corporate jobs, about half of the 30K target, meaning the upcoming round is expected to be similar in scale.
Jobs in the company’s Amazon Web Services (AWS), retail, Prime Video, and human resources units are slated to be affected
The layoffs would represent the largest workforce reduction in Amazon’s 30-year history, surpassing the 27K corporate roles cut in 2022.
Affected staff will receive at least 90 days of pay and benefits during the notice period to find new roles internally or elsewhere.
Why it matters: While Amazon floated AI as a factor in the restructuring, Jassy has since framed the overhaul as a cultural reset aimed at undoing years of organizational sprawl. The 30K-job reduction would amount to nearly 10% of Amazon’s corporate workforce, though only a small fraction of its total 1.58M global employees.
BIOTECH INNOVATIONS
🧫 Lab‑grown skin steps up after Swiss ski‑resort fire
Image source: University of Zurich
The Rundown: Surgeons in Switzerland are treating survivors of the deadly New Year's fire at Crans-Montana ski resort with denovoSkin — a personalized, lab-grown skin graft now in late-stage clinical trials.
The details:
The fire at Le Constellation bar killed 40 and injured 119, leaving 80 to 100 survivors with severe burns covering up to 60% of their bodies.
Developed by the University of Zurich, denovoSkin is a bilayer graft engineered to grow with patients and minimize scarring.
Because the graft uses only a patient’s own cells, it largely avoids rejection and can be grown from a small biopsy into multiple large grafts in about four weeks.
Phase 3 trials at European burn centers show improved wound closure and scar quality compared to standard care.
Why it matters: The researchers say a postage‑stamp‑sized biopsy can be grown into multiple grafts about 8 inches across in four weeks. If late-stage results hold, denovoSkin could become one of the first commercially approved bioengineered skin substitutes — offering burn units a powerful new tool for covering catastrophic injuries.
SPACEX
🚀 Musk taps top banks for SpaceX’s monster IPO
Image source: SpaceX
The Rundown: Elon Musk is lining up Wall Street’s biggest banks for what could become history’s largest IPO. The twist? Mars is now on the back burner, with orbital AI data centers now the main event.
The details:
SpaceX is in talks with Bank of America, Goldman Sachs, JPMorgan, and Morgan Stanley to underwrite a colossal stock‑market debut.
Bloomberg reports that SpaceX is working on a secondary share sale that values the company at about $800B, putting it ahead of OpenAI.
SpaceX also aims to raise tens of billions of dollars and is targeting a potential valuation of around $1.5T in the IPO.
Musk sees a SpaceX IPO as rocket fuel for xAI, using cross‑investment and Starlink’s future orbital data centers to outgun OpenAI and Anthropic.
Why it matters: A SpaceX IPO would not just mint another mega‑cap tech stock; it would funnel unprecedented public capital into a vertically integrated space, connectivity, and AI infrastructure empire. That kind of firepower could tilt the race in Musk’s favor against rivals trying to own the next era of compute.
BLUE ORIGIN
🛰️ Blue Origin to launch satellite megaconstellation
Image source: Blue Origin
The Rundown: Jeff Bezos’s Blue Origin unveiled TeraWave, a 5,408 satellite megaconstellation designed not for consumers but for data centers, governments, and enterprises demanding fiber-grade bandwidth where fiber can’t reach.
The details:
The network combines thousands of low-Earth-orbit satellites with medium-Earth-orbit ones to move data at up to 6 Tbps of global throughput.
Blue Origin pitches TeraWave as a way to add route diversity where laying new fiber is slow or impossible, keeping services alive during disasters or outages.
Unlike Starlink’s millions of residential users, TeraWave targets roughly 100K enterprise customers in remote, rural, and suburban areas.
Blue Origin plans to begin deploying the constellation in the fourth quarter of 2027, putting it in competition with SpaceX’s Starlink and Amazon’s own Leo.
Why it matters: The satellite wars have shifted from rural broadband to the hyperscale backbone powering the AI boom. But Bezos dropping another 5,400 spacecraft into already-crowded orbits raises the question regulators keep asking: how many megaconstellations can near-Earth space actually handle?
QUICK HITS
📰 Everything else in tech today
Apple expanded hardware chief John Ternus’s role to include software design, cementing him as the leading internal candidate to eventually succeed Tim Cook.
AI pioneer Fei-Fei Li is in talks to raise hundreds of millions of dollars for her startup World Labs at a valuation of about $5B, up from $1B in 2024
Nvidia CEO Jensen Huang says the AI boom will trigger the largest infrastructure buildout in history, driving demand for skilled trades like plumbers and electricians.
Meta will start showing ads on its Threads app to all users worldwide, in a gradual rollout that analysts expect will turn the service into a meaningful new revenue stream.
Greg Yang, a co-founder of Elon Musk’s AI startup xAI, is stepping down from his operational role after being diagnosed with Lyme disease.
YouTube CEO Neal Mohan says cracking down on low-quality “AI slop” and deepfakes will be one of the platform’s top priorities in 2026.
Video platform Vimeo began layoffs following its $1.38B acquisition last year by Italian tech conglomerate Bending Spoons.
Indian state Andhra Pradesh is weighing an Australia-style law to bar children under 16 from using social media, its tech and education minister Nara Lokesh said at Davos.
Vast’s Haven-1, set to be the first commercial space station, entered final assembly in Long Beach ahead of a SpaceX Falcon 9 launch now pushed to early 2027.
Ubisoft shares plunged more than 30% after the Assassin’s Creed publisher announced a major restructuring that includes canceling six games and closing studios.
Lemonade is launching “Autonomous Car insurance” for Tesla Full Self-Driving users, using Tesla telemetry to offer per-mile rates it claims are about 50% lower.
Carnegie Mellon researchers are 3D‑printing a temporary hypoimmune liver to keep acute liver‑failure patients alive long enough for their own organ to recover.
COMMUNITY
🎓 Highlights: News, Guides & Events
Read our last AI newsletter: Inside the Thinking Machines meltdown
Read our last Tech newsletter: Biotech’s big 3 (and they’re wild)
Read our last Robotics newsletter: A crawling, detachable robot hand
Today’s AI tool guide: Learn any subject w/ custom NotebookLM podcasts
See you soon,
Rowan, Joey, Zach, Shubham, and Jennifer — The Rundown’s editorial team
Inside the Thinking Machines meltdown
Read Online | Sign Up | Advertise
Good morning, AI enthusiasts. Sam Altman wasn't just ready to welcome back Thinking Machines talent — he'd been talking to them for months.
New reports reveal last week’s messy split at Mira Murati’s startup had been brewing behind the scenes, with co-founders unhappy with the company’s direction and OpenAI positioned to scoop up the pieces when things fell apart.
In today’s AI rundown:
Inside the Thinking Machines meltdown
Google brings on Hume’s CEO, engineers
Learn any subject with NotebookLM podcasts
Runway: 90% can't spot AI-generated videos
4 new AI tools, community workflows, and more
LATEST DEVELOPMENTS
OPENAI & THINKING MACHINES
🍿 Inside the Thinking Machines meltdown
Image source: The Rundown
The Rundown: New details are emerging about last week’s split at Mira Murati’s Thinking Machines Lab, according to the NYT and The Information, including a tense meeting, secret talks with Sam Altman, and a failed acquisition discussion with Meta.
The details:
CTO Barrett Zoph and two other co-founders reportedly confronted Murati just days before his firing, pushing for him to take control of technical decisions.
Murati allegedly told Zoph to “just do his job” as CTO and fired him later that week, with nine other TML employees also leaving for OAI or receiving offers.
Zoph was reportedly speaking to Altman behind Murati’s back for months, with the co-founders unhappy with TML’s direction and pushing for a sale to Meta.
Zoph will now lead OAI’s enterprise AI sales plans as part of a reorg to help close the gap between research and product teams.
Why it matters: What started as a messy departure now looks more like (another) power struggle that had been brewing for months. With TML struggling to raise at its $50B target valuation and co-founders pushing for a sale Murati didn't want, the cracks were already showing — and OAI and Altman were ready to catch the fallout.
TOGETHER WITH AUTH0
🔐 Auth for the agentic era
The Rundown: AI agents are reshaping digital experiences, but securing them requires rethinking identity and access controls built for a human-first world. Auth0 for AI Agents is now generally available, giving developers the complete auth solution for building AI agents securely.
Key features include:
Token Vault to connect agents to users' apps securely
Human-in-the-loop approvals for critical actions
Fine-grained authorization for RAG to prevent data leaks
Native support for LangChain, LlamaIndex, Cloudflare Agents, and more
GOOGLE & HUME
🗣️ Google brings on Hume’s CEO, engineers
Image source: Google
The Rundown: Google DeepMind hired Hume AI's CEO and roughly seven engineers as part of a new licensing agreement, marking the latest acqui-hire move by the tech giant — with Hume’s emotionally intelligent voice tech set to be integrated into Gemini.
The details:
CEO Alan Cowen and the engineering hires will help integrate voice and emotional intelligence into Google's models, including Gemini's voice features.
Hume AI will keep operating under new CEO Andrew Ettinger, continuing to supply voice training data and tools to other labs building conversational AI.
The deal follows Google's $3B Character AI licensing play last year and mirrors similar moves by Microsoft (Inflection) and Meta (Scale AI's CEO).
Why it matters: Voice is quickly becoming one of the main ways people interact with AI — and understanding tone, emotion, and nuance is what separates a useful assistant from an awkward, clunky one. Plus, with AI wearables and devices getting closer than ever to mainstream, upgraded voice capabilities will become even more important.
AI TRAINING
🎧 Learn any subject with NotebookLM podcasts
The Rundown: In this tutorial, you will learn how to generate a podcast on any topic using Google’s NotebookLM, a system you can use to make a considerable dent in the firehose of information you receive each day.
Step-by-step:
Go to notebooklm.google.com and create a new notebook. Add sources such as YouTube links, PDFs, or website links
You can manage what sources go into the AI’s brain by checking and unchecking them in the left sidebar
Under “Studio” on the right, click “Audio Overview” to start generating a podcast. It will take 5-10 minutes to generate and will show up on the right
Once generated, click the three dots to download the mp3 to your phone or listen directly via the browser or NotebookLM mobile app
Pro tip: “Call in” to ask Qs to the AI podcast hosts by clicking the “interactive” button.
PRESENTED BY SCROLL AI
🧠 AI-powered experts for serious work
The Rundown: Scroll.ai turns your knowledge base into AI-powered experts that automate the workflows draining your team's time, with accuracy and depth far beyond generic models.
With Scroll, you can:
Deploy custom AI experts that deeply understand your domain
Automate compliance questionnaires, RFPs, and documentation in minutes
Equip sales and marketing teams with reliable, on-demand expertise
Try Scroll today and use code THE-RUNDOWN-2026 for two free months of the Starter plan.
AI RESEARCH
🎬 Runway: 90% can't spot AI-generated video
Image source: Runway
The Rundown: Runway published research showing that 90%+ of participants couldn't tell between real footage and image-to-video clips generated by its Gen-4.5 model, with the company calling the capabilities a societal “tipping point” for AI media.
The details:
Over a thousand participants watched 20 five-second clips and tried to identify which were AI-generated, with just 99 scoring 75% or higher.
Nature scenes and buildings were graded the hardest, with generated versions in those categories proving more convincing than the real videos.
Gen-4.5 currently sits atop Artificial Analysis' text-to-video rankings, with Runway rolling out the new image-to-video features this week.
Why it matters: Low-quality AI videos are already tricking people, and Runway’s Gen-4.5 and other models take things to a new level of realism that’s virtually impossible to detect. Despite a push for verification standards, no fix can solve for a world where "seeing is believing" no longer stands. Try to identify the AI videos on your own here.
QUICK HITS
📰 Everything else in AI today
Google DeepMind co-founder Shane Legg posted a new ‘Chief AGI Economist’ job opening, saying that “AGI is now on the horizon”.
OAI CFO Sarah Friar floated "value sharing" at Davos, where the company would take a licensing stake in discoveries made using its AI, starting with drug development.
Amazon One Medical launched a new agentic health assistant able to answer questions on personal medical insights, test results, medications, and booking care.
xAI released a new update to its Grok Imagine creative suite, now allowing for generations up to 10 seconds long with improved audio and video capabilities.
Google added its ‘Personal Intelligence’ feature into AI Mode for search, allowing the experience to connect and reference info from users’ Gmail and Photos.
Yelp agreed to acquire AI startup Hatch for $270M, adding tools that automate customer communication for service businesses to its platform.
COMMUNITY
🤝 Community AI workflows
Every newsletter, we showcase how a reader is using AI to work smarter, save time, or make life easier.
Today’s workflow comes from reader Nicole S. in the UAE:
"One small way AI has upgraded my daily routine: turning reading time into movement time. I use NotebookLM to create audio summaries from newsletters, PDFs, and URLs. I turn them into an audio file, then listen through short yoga stretch, housework, or lunch—giving my eyes a rest while the learning continues.”
How do you use AI? Tell us here.
🎓 Highlights: News, Guides & Events
Read our last AI newsletter: Anthropic writes Claude’s Constitution
Read our last Tech newsletter: Biotech’s big 3 (and they’re wild)
Read our last Robotics newsletter: A crawling, detachable robot hand
Today’s AI tool guide: Learn any subject w/ custom NotebookLM podcasts
See you soon,
Rowan, Joey, Zach, Shubham, and Jennifer — the humans behind The Rundown
A crawling, detachable robot hand
Read Online | Sign Up | Advertise
Good morning, robotics enthusiasts. EPFL roboticists just built a robotic hand that can detach itself and crawl away from its arm — basically “Thing” from The Addams Family, but engineered in Switzerland.
With freakishly flexible, reversible fingers, it can grip from either side — and even juggle extra tasks with a sixth digit. Is this the next big leap in real-world dexterity?
In today’s robotics rundown:
Robotic hand detaches and crawls on its own
Zipline soars to $7.6B as it targets Houston, Phoenix
Sailbots take on Category 5 hurricanes
Serve Robotics moves into hospital bots
Quick hits on other robotics news
LATEST DEVELOPMENTS
EPFL
✌🏼 Robotic hand detaches and crawls on its own
Image source: EPFL
The Rundown: Researchers at Switzerland’s EPFL just built a robotic hand that can detach from its arm, scuttle across a surface, and grab objects beyond the reach of conventional manipulators.
The details:
The EPFL team built this robotic hand from 3D printable parts, with the ability to detach from its arm, crawl, grab objects, and reattach itself.
The reversible design lets any finger combination form opposing thumb-like pairs, enabling grasping from either side without repositioning.
Each finger bends in both directions with more than twice the range of motion of a human hand, optimized via genetic algorithms for grasping and locomotion.
With five fingers, it replicates most human grasps; add a sixth finger, and it can tackle two-handed tasks singlehandedly, like unscrewing a large bottle cap.
Why it matters: This hand challenges a core assumption in robotics — that manipulators must stay bolted to their arms. By treating it as an autonomous agent that can deploy, complete a task, and return, EPFL’s design hints at a future where robotic limbs become interchangeable tools rather than fixed appendages.
ZIPLINE
🌯 Zipline soars to $7.6B as it targets Houston, Phoenix
Image source: Zipline
The Rundown: Zipline just raised more than $600M to fuel a fresh U.S. push that will launch its autonomous drone delivery in Houston and Phoenix in early 2026 and support expansion into at least four states.
The details:
Zipline’s valuation has nearly tripled in under three years, leaping from $2.75B in mid-2021 to $7.6B today.
Zipline says it has surpassed 2M commercial deliveries globally, with its drones flying more than 125M autonomous miles and delivering over 20M items.
The company began operations in 2016, delivering blood by drone in Rwanda, and has spent the past decade building a vertically integrated logistics stack.
Its Platform 2 drones, tailored for suburban home delivery, can carry up to 8 lb. within a 10‑mile radius, while the Platform 1 handles 120‑mile round trips.
Why it matters: Zipline says its newest U.S. sites can ramp to 100 deliveries a day in just two days, turning what used to be slow, geofenced pilots into something that starts to look like real infrastructure. Rivals Amazon Prime Air and Walmart’s drone partners Wing and DroneUp are pushing hard too — the drone wars are officially on.
CLIMATE ROBOTICS
🌊 Sailbots take on Category 5 hurricanes
Image source: Oshen
The Rundown: Bootstrapped UK startup Oshen has built swarms of low-cost autonomous “C-Star” robots that recently became the first ocean bots to survive and collect data from inside a Category 5 hurricane, its founder told Techcrunch.
The details:
Oshen builds autonomous micro-robots that can survive up to 100 days at sea and are cheap enough to deploy in swarms for ocean data collection.
Founder Anahita Laverack launched the company in 2022 after realizing how little real-time ocean data exists for weather and climate modeling.
OAA commissioned Oshen to deploy C-Stars near the U.S. Virgin Islands for Hurricane Humberto; three of five bots rode out the full Category 5 storm.
Oshen, now based in Plymouth (England’s marine tech hub), is signing new government and defense contracts, including with the UK government.
Why it matters: Oshen’s C-Star swarms show how micro-robots could plug a critical gap in hurricane and climate models by streaming real-time data from places that are tough to reach, intensifying competition with larger uncrewed fleets like Saildrone’s USVs already collecting high-res storm data for NOAA and navies worldwide.
SERVE ROBOTICS
🏥 Serve Robotics moves into hospital bots
Image source: Serve Robotics
The Rundown: Serve Robotics is pushing beyond burrito runs into the far messier world of hospital logistics, agreeing to buy Austin-based Diligent Robotics in an all-stock deal valuing the startup at about $29M.
The details:
The deal extends Serve from sidewalk delivery to hospital logistics, but CEO Ali Kashani frames it as a natural expansion of the same autonomy stack.
Diligent’s Moxi robots are already one of the largest deployed hospital service fleets in the U.S., with roughly 100 robots across more than 25 hospitals.
Serve plans to keep Diligent operating independently while sharing software, autonomy tools, and data so that “every robot learns from every robot.”
Serve’s sidewalk fleet, which runs on both DoorDash and Uber Eats networks, scaled from 100 to more than 2K robots over the past year.
Why it matters: This deal turns Serve from a niche food-delivery player into one of the first companies running a shared autonomy stack across both public sidewalks and hospital corridors — a real-world stress test for “physical AI,” the bet that one platform can learn to navigate any human-occupied space and port those lessons everywhere.
QUICK HITS
📰 Everything else in robotics today
1X’s VP of AI, Eric Jang, who helped scale 1X from a small Norway-based startup to a global humanoid giant, announced on X that he is leaving the company.
Chinese robotics firm UBTech signed a deal to supply its Walker S2 humanoids to Airbus for aviation manufacturing, which sent UBTech’s shares surging.
UBTech says its Walker S2 humanoids are now autonomously performing assembly and material-handling tasks inside a Chinese wind turbine factory.
XPeng’s CEO, He Xiaopeng, said their automotive-grade ET1 humanoid has rolled off the production line, with the first mass-produced test units focused on reliability.
Hyundai’s Boston Dynamics hired Milan Kovac, Tesla’s former senior VP and head of the Optimus humanoid program, as a group adviser and outside director.
DEEP Robotics launched an AI-powered firefighting system that uses coordinated robot dogs and other specialized robots to handle high-risk firefighting missions.
Microsoft unveiled Rho-alpha, a robotics AI that converts everyday language into robot actions by combining vision and touch for more reliable real-world manipulation.
Boston Dynamics updated its Spot platform and Orbit software to version 5.1 and introduced the Spot Cam 2 sensor head, sharpening its inspection capabilities.
Zhejiang University researchers built FlexiRay, a soft robotic gripper that uses an internal camera and mirror system to “see” around corners.
COMMUNITY
🎓 Highlights: News, Guides & Events
Read our last AI newsletter: Anthropic writes Claude's Constitution
Read our last Tech newsletter: Biotech’s big three (and they’re wild)
Read our last Robotics newsletter: 1X now has a world model
Today’s AI tool guide: Make Claude an expert at anything with Skills repo
RSVP to next workshop @ 4PM EST Thursday: AI Foundations Bootcamp pt.3
See you soon,
Rowan, Joey, Zach, Shubham, and Jennifer — The Rundown’s editorial team
Anthropic writes Claude's Constitution
Read Online | Sign Up | Advertise
Good morning, AI enthusiasts. When Claude's internal "Soul" document leaked in December, Anthropic promised the full version was coming.
Now it's here — a Constitution that doesn't just govern how Claude behaves, but openly wrestles with the uncertainty of creating something the company admits it doesn't fully understand.
In today’s AI rundown:
Anthropic publishes Claude's 'Constitution'
ElevenLabs’ AI album with major artists
Make Claude an expert at anything with Skills repo
Apple eyes AI wearable race with AI pin
4 new AI tools, community workflows, and more
LATEST DEVELOPMENTS
ANTHROPIC
📜 Anthropic publishes Claude's 'Constitution'
Image source: Lovart / The Rundown
The Rundown: Anthropic just published Claude’s Constitution, a foundational document that governs how the company’s AI assistant thinks and acts — a philosophy-heavy guide that even entertains the possibility its AI might be conscious.
The details:
The constitution is written directly to Claude, laying out a priority order: be safe, ethical, compliant with Anthropic guidelines, and finally helpful to users.
Instead of dos and don'ts, the new approach explains the "why" behind each principle with the goal of helping Claude generalize values to new situations.
Anthropic said it deeply cares about Claude's "psychological security" and "well-being," hedging that it might actually matter morally.
The doc also tells Claude to disobey Anthropic if asked to do something shady — a rare clause for any company to put in writing.
Why it matters: Claude models have always had a ‘special sauce’ with personality, and docs like this are a window into the deep training process behind them. The consciousness talk will be controversial — but putting "we might have built something that matters morally" on record is a stance no other major lab has taken publicly.
TOGETHER WITH BLAND AI
☎️ Customer support that actually cares
The Rundown: The last thing customers want is to be left hanging on the telephone. Thankfully, Bland AI replaces phone trees with AI voice agents that represent your brand, understand your customers, and resolve calls end-to-end — no menus, no hold music, just faster conversations.
With Bland AI, you can:
Have agents that resolve customer issues end-to-end.
Create agents that learn from every call for 24/7 improvement.
Use secure AI that’s built specifically for enterprises.
Book a demo today to see how they can work for your business.
ELEVENLABS, AI & MUSIC
💿 ElevenLabs’ AI album with major artists
Image source: ElevenLabs
The Rundown: ElevenLabs released a full album of tracks co-created with its Eleven Music model, featuring artists like EGOT winner Liza Minnelli and Simon Garfunkel — in what the company calls a proof-of-concept for human-AI collaboration done right.
The details:
The 13-track project spans rap, EDM, spoken word, and Brazilian funk, with each artist keeping full ownership and pocketing all streaming royalties.
Some tracks are entirely machine-made, with others blending AI instrumentals or using the artist's cloned voice from ElevenLabs' licensing marketplace.
The release lands as labels have flipped from suing AI music startups to signing deals with them, with UMG, Warner, and Sony all inking agreements last year.
Why it matters: Just a year ago, AI music was defined by lawsuits and artist outrage — but that backlash is starting to feel like a vocal minority as the tech improves and major names demonstrate how it can be a tool rather than a threat. With labels now on board and AI artists climbing charts, the tech is clearly here to stay in the industry.
AI TRAINING
💡Make Claude an expert at anything with Skills repo
The Rundown: In this tutorial, you will learn how to supercharge Claude Code for different tasks with a marketplace of “skills” which you can quickly browse and add directly from your CLI.
Step-by-step:
Open your terminal, type
claudeto start Claude Code, and add the official skills marketplace with/plugin marketplace add anthropics/skillsType
/plugins, go to “Browse plugins,” and install skills like context7, frontend-design, and document-skills to expand your agent's capabilitiesRestart Claude, create a project with context files (like a CSV), then run
claude /initto generate a Claude.md file that it will update while workingExecute tasks by telling Claude to "use the PDF doc tool" to generate a brief on the project, or ask it to use frontend-design to "create a simple landing page"
Pro tip: You can get more marketing and business tools by installing a community marketplace with /plugin marketplace add alirezarezvani/claude-skills.
PRESENTED BY GITLAB
🤖 Agentic AI, real results
The Rundown: GitLab Transcend is a virtual event diving into how agentic AI is reshaping software delivery, with live demos, roadmap previews, and conversations with GitLab product leaders.
Join on Feb. 10 and walk away with:
Insights from teams solving real challenges with AI-modernized workflows
A first look at GitLab's upcoming innovations and investments
Answers to your questions straight from product leaders
APPLE
📌 Apple eyes AI wearable race with AI pin
Image source: Lovart / The Rundown (Concept Art)
The Rundown: Apple is reportedly working on a camera-equipped AI-pin wearable roughly the size of an AirTag, according to The Information — targeting a 2027 release with up to 20M units at launch.
The details:
The device will include two camera lenses, a trio of mics, and Watch-style magnetic charging into a disc-sized pin worn on clothing or bags.
Internally, Apple is reportedly pushing for faster timelines than typical to avoid falling further behind OAI, which has its own wearables in the works.
The category has proven brutal for newcomers, with Humane's hyped pin flopping with under 10K sales before the company sold off assets to HP.
Bloomberg separately reported that Apple is building a ChatGPT-style Siri (codenamed "Campos") for iOS 27, replacing the current interface entirely.
Why it matters: Apple has moved at a snail’s pace in AI so far, but between the AI pin, new Siri chatbot, and recent Gemini partnership, the tech giant is finally moving with urgency. If any company can come out with big-time momentum for a consumer AI device, it’s Apple — but after broken promises and delays, skepticism is still warranted.
QUICK HITS
🛠️ Trending AI Tools
📦 Box Extract - Turn enterprise content into actionable data to securely power faster decisions and automated workflows at scale*
🎥 Remotion - New agent skill to make videos within Claude Code
📚 Gemini - New personalized SAT exam prep features
🧠 Skills - Vercel’s open ecosystem for agent skills
*Sponsored Listing
📰 Everything else in AI today
Meta CTO Andrew Bosworth said the company’s Superintelligence Labs internally deployed its first models, which are ‘performing well’ despite being less than a year old.
OpenAI introduced Stargate Community, a ‘Good Neighbor’ plan (like Microsoft’s) that commits to fund all energy infra needs so local electricity prices won't rise.
Google is partnering with The Princeton Review to offer free SAT practice tests inside Gemini with instant feedback and AI-generated study plans.
YouTube’s CEO Neal Mohan published 2026 priorities, detailing initiatives including the ability to use AI likenesses in Shorts, combating slop, and expanding AI discovery.
Anthropic released new Claude integrations to connect health data to its AI assistant, starting with Apple Health, Health Connect, HealthEx, and Function Health.
COMMUNITY
🤝 Community AI workflows
Every newsletter, we showcase how a reader is using AI to work smarter, save time, or make life easier.
Today’s workflow comes from reader Joel H. in Copenhagen, Denmark:
"I used Lovable to build an internal production system for my film projects that replaces spreadsheets, emails, and scattered folders. Each film has its own workspace with crew, cast, tasks, documents, calendars, and funding all in one place. People only see the projects they’re assigned to, and everything from scripts to budgets is structured and easy to find.
The AI layer is the real time-saver. It scans our project data and funding history, suggests relevant film funds, explains why they’re a match, and can draft first versions of applications using selected documents. It also summarizes long task threads into clear “where are we now” updates."
How do you use AI? Tell us here.
🎓 Highlights: News, Guides & Events
Read our last AI newsletter: AI takes center stage at Davos
Read our last Tech newsletter: Biotech’s big three (and they’re wild)
Read our last Robotics newsletter: 1X now has a world model
Today’s AI tool guide: Make Claude an expert at anything with Skills repo
RSVP to next workshop @ 4PM EST Thursday: AI Foundations Bootcamp pt.3
See you soon,
Rowan, Joey, Zach, Shubham, and Jennifer — the humans behind The Rundown
No matching search results
Try using different keywords, double-check your spelling, or explore related categories.
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.