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.
Robot performs flawless surgery, alone
Read Online | Sign Up | Advertise
Good morning, robotics enthusiasts. A robot just nailed a critical phase of gallbladder surgery — no hands, no joystick, just pure silicon smarts.
Johns Hopkins’ SRT-H didn’t just follow orders; it listened, adapted, and self-corrected like a junior surgeon learning on the fly. If a robot can now think its way through surgery, how long before it earns the right to operate solo on you?
In today’s robotics rundown:
Robot performs surgery without human help
Chinese robot dog breaks speed record
Futurist Adam Dorr: The robots are coming
Coyote robots guard U.S. military airfields
Quick hits on other robotics news
LATEST DEVELOPMENTS
JOHNS HOPKINS
😷 Robot performs surgery without human help
Image source: Juo-Tung Chen/Johns Hopkins University
The Rundown: A Johns Hopkins robot just nailed a key part of gallbladder surgery with no human hands on deck. Named SRT-H, the robot performed the surgery eight times like a seasoned pro, self-correcting and adapting in real time.
The details:
In trials, the team said SRT-H took spoken commands and adapted throughout the procedure, mirroring a junior surgeon’s learning curve.
The robot operated on a synthetic patient model, using organs from dead pigs, simulating the complexity and unpredictability of live human tissue.
Unlike traditional surgical robots, which follow pre-set instructions, this system demonstrated adaptive learning, adjusting its actions on the fly.
Backed by federal funding, the team calls this a leap toward robots that combine machine precision with a surgeon’s improvisational skill.
Why it matters: SRT-H marks a new era for surgical robots, adapting and learning in real time instead of just following scripts, and even acing simulated medical emergencies. But don’t expect fully automated robot surgeries on humans just yet; human trials aren’t expected until later this decade.
MIRROR ME
🔥 Chinese robot dog breaks speed record
Image source: Mirror Me
The Rundown: When it comes to robot dogs, Boston Dynamics’ Spot often steals the show. But this time, China’s Mirror Me startup is turning heads with Black Panther II — a four-legged robot that just set a new record by sprinting 100m in 13 seconds.
The details:
Weighing in at 38 kg and standing just over half a meter tall, Black Panther II ran a record-breaking 100m dash on live TV, topping 10.4m/second.
The pace beats Boston Dynamics’ WildCat, which has been shown to reach 8.8m/second in testing, and is on par with top human sprinters.
Black Panther II’s legs are crafted from carbon fiber and mimic the anatomy of jumping mice and cheetahs to take giant leaps.
AI-driven algorithms enable the robot dog to adapt its gait in real time, adjusting to different terrains and optimizing stride patterns for speed.
Why it matters: While Usain Bolt still holds the 100m record, China’s Black Panther outpaces almost all humans except the world’s top sprinters. Its AI-driven gait adaptation and biomimetic design, blending animal-inspired mechanics with real-time machine learning, hint that much smarter and faster robots are on the way.
ROBOT REVOLUTION
🤔 Futurist Adam Dorr: The robots are coming
Image source: Ideogram/The Rundown
The Rundown: Adam Dorr, research chief at the tech futurist think tank RethinkX, predicts that within two decades, robots and AI will outpace human labor so completely that “work” as we know it could vanish from the economic map.
The details:
In a Guardian interview, Dorr sketches a future where robots — cheaper, faster, and smarter — replace not just factory hands but nearly every other profession.
His team argues that the current wave of automation isn’t an industrial revolution but an extinction-level event for the labor market.
Dorr’s thesis is as stark as it is sweeping: whatever your job, whatever your sector, machines will soon do it better and for less.
The pattern, he says, is relentless and predictable: Costs keep dropping, capabilities keep climbing, and society pivots overnight.
Why it matters: Many economists argue that the real story is more complex: robots will likely automate repetitive tasks, but humans will remain essential for jobs requiring judgment, empathy, and adaptability. In either case, Dorr says that if the society can adapt, this could be the gateway to an era free from the grind of wage labor.
MILITARY ROBOTS
🐦⬛ Coyote robots guard U.S. military airfields
Image source: ERDC
The Rundown: To avoid disastrous bird strikes on multimillion-dollar aircraft, the U.S. Army Engineer Research and Development Center (ERDC) is swapping out traditional scare tactics for something a bit more cyberpunk: robotic coyotes.
The details:
Dubbed “Coyote Rovers,” these four-wheeled unmanned ground vehicles are topped with life-size plastic coyote figures and programmed to patrol airfields.
Developed with wildlife biologists, the Coyote Rovers have already been tested at bases like Naval Air Station Pensacola, the home of the Blue Angels.
With speeds of 20 mph, these rovers can quickly cover large areas and react to wildlife presence in real time, outperforming earlier robot dog prototypes.
The Coyote Rovers are programmed to autonomously patrol airfields, following set routes, and can operate continuously for days.
Why it matters: Future upgrades aim to add AI and machine learning, letting the bots identify specific species and customize their deterrent tactics. This smart, adaptable approach promises to minimize dangerous wildlife strikes and offers a scalable, cost-saving solution for protecting not only airfields but also other vital infrastructure.
QUICK HITS
📰 Everything else in robotics today
Hugging Face launched pre-orders for its Reachy Mini open-source desktop robots, offering a $449 wireless model with onboard computing and a $299 lite version.
Filics, a robotics company backed by Amazon, secured $16M in funding to scale up its automated pallet handling platform.
A San Francisco “underground” fight club for robots has reportedly emerged, featuring high-end humanoid cage matches.
MIT researchers developed a deep-learning control system that teaches soft, bio-inspired robots to move and follow commands from just a single image.
Chinese startup Shenzhen Dobot posted a clip showing its Atom humanoid cooking a steak while being remotely controlled from nearly 1800 km away via a VR headset.
Chinese startup MagicLab just launched MagicBot Z1, a 4'7" tall Unitree G1 lookalike that can run, kick, and perform advanced yoga poses.
NEURA Robotics teamed up with HD Hyundai Robotics to co-develop and test advanced humanoids for the shipbuilding industry.
Canadian robotics firm Cobionix raised $3M to commercialize CODI, an AI-driven healthcare robot designed to automate and enhance patient care.
Philadelphia startup Daxo Robotics emerged from stealth with its soft robotic “ultra-dexterous” hands featuring “infinite degrees of freedom.”
Chinese humanoid startup Robot Era, a Tsinghua University spinout, raised nearly 500 million yuan ($69M) in a Series A round.
Realbotix just gave its humanoid “girlfriend” a serious upgrade: it now speaks 15 languages fluently and taps the cloud for another 147.
COMMUNITY
🎥 Join our next live workshop
Join our next workshop this Friday, July 11th, at 4 PM EST with Dr. Alvaro Cintas, The Rundown’s AI professor. By the end of the workshop, you’ll confidently be able to install and use Gemini CLI to boost your productivity right from the command line.
RSVP here. Not a member? Join The Rundown University on a 14-day free trial.
See you soon,
Rowan, Jennifer, and Joey—The Rundown’s editorial team
xAI's Grok 4 arrives
Read Online | Sign Up | Advertise
Good morning, AI enthusiasts. xAI’s highly anticipated Grok 4 has arrived — and it’s crushing benchmarks across the board.
The newest “truth-seeking” AI takes us a step closer to AGI, with Musk suggesting it may start discovering new physics “as soon as this year.” But given the backlash over Grok 3’s recent racist and antisemitic comments, the release is also likely to face more scrutiny than ever before.
In today’s AI rundown:
xAI releases Grok 4 following 3’s crashout
Perplexity’s Comet browser for AI-first web
Turn messy image filenames into descriptive ones
OpenAI snags top engineers from rivals for scaling team
4 new AI tools & 4 job opportunities
LATEST DEVELOPMENTS
XAI
🚀 xAI releases SOTA Grok 4 following 3’s crashout
Image source: xAI
The Rundown: xAI just announced Grok 4 and Grok 4 Heavy, its next-gen reasoning-only models that are “better than PHD levels in every subject” and deliver SOTA capabilities across benchmarks, including Arc-AGI and Humanity’s Last Exam.
The details:
Grok 4 is a single-agent AI with voice, vision, and a 128K context window, while 4 Heavy is its advanced sibling, with multiple agents to tackle complex tasks.
Both mark a major jump in benchmarks, achieving SOTA on Humanity's Last Exam, Arc-AGI-2, and AIME, and surpassing Gemini 2.5 Pro and OpenAI’s o3.
Grok 4 is available with the SuperGrok subscription at $30/month, while Grok 4 Heavy is part of the new SuperGrok Heavy plan priced at $300/month.
The new model is also available via API with a 256K-token context window and built-in search, priced at $3/million input tokens and $15/million output tokens.
The power-packed release comes after a major backlash against Grok 3, which was caught making racist and antisemitic comments after an update.
Why it matters: Despite being a relatively new player, Musk’s xAI is already challenging the AI heavyweights. The latest release showcases the power of its Colossus supercomputer and pushes the scaling frontier further, though in the wake of the Grok 3 controversy, it’s likely to face heightened scrutiny from experts around the world.
TOGETHER WITH ASAPP
♻️ Automate complex customer interactions with AI agents
The Rundown: ASAPP’s GenerativeAgent is an enterprise-grade AI agent that resolves real customer issues on its own across voice and chat. Built for high-stakes environments, it goes beyond conversation to deliver real outcomes backed by security, collaboration, and continuous learning.
With GenerativeAgent, you’ll experience:
Fewer transfers, faster resolutions with human-AI collaboration workflow
Built-in guardrails and enterprise-grade data protection
Real-time quality monitoring and self-improving
Take a self-guided tour of GenerativeAgent today.
PERPLEXITY
🖥️ Perplexity’s Comet browser for AI-first web
Image Source: Perplexity
The Rundown: Perplexity launched Comet, a new AI browser that embeds the company’s search engine alongside an assistant capable of performing agentic tasks—like booking meetings and navigating websites—while integrating with user workflows.
The details:
The Comet Assistant lives in a sidebar that watches users browse, answering questions while automating tasks like email and calendar management.
Users can utilize the agentic assistant to “vibe browse” without interacting directly with sites, using natural language or via voice commands.
The browser promises seamless integration with existing extensions and bookmarks, supporting both Mac and Windows at launch.
Perplexity Max users ($200/mo subscription) get first access along with a rolling waitlist, with Pro, free, and Enterprise users coming at a later date.
Why it matters: Chrome has had a chokehold on the browser for years — but appears to be a step behind on the agentic, AI-driven transition. While there will be hiccups as agents continue to evolve, Dia, Comet, and soon OpenAI (more below) are taking the first steps into a new, inevitable shift in how we navigate and take actions on the web.
AI TRAINING
📸 Turn messy image filenames into descriptive ones
The Rundown: In this tutorial, you will learn how to use Google's Gemini CLI to analyze your images and generate SEO-friendly filenames automatically, improving your content organization and search engine visibility.
Step-by-step:
Install Gemini CLI: npm install -g @google/gemini-cli and authenticate with your Google account
Test a single image analysis typing: gemini "Describe what's in [image1.png]"
Batch process: gemini “Process all images in this folder. For each image, analyze the content and rename it with a descriptive filename with relevant keywords for SEO purposes.”
Pro tip: Start with a small batch to understand how Gemini interprets your content, then scale up to your entire image library.
PRESENTED BY INVISIBLE
📊 AI training trends from those behind the scenes
The Rundown: As the partner for many enterprises and leading foundation models including AWS, Microsoft, and Cohere, Invisible has a front row seat into how the AI training landscape is evolving in real-time.
In this free guide, Invisible reveals:
The most important shifts happening in model development, optimization, and deployment
The rising bar for quality data
A focus on specialized training sets
The emergence of agentic products for real-world outcomes
OPENAI
🥊 OpenAI snags top engineers from rivals for scaling team
Image source: Greg Brockman (@gdb on x)
The Rundown: OpenAI recruited four new senior engineers from Tesla, xAI, and Meta for its scaling team, according to WIRED — joining to work on the Stargate data center and infrastructure initiatives, and coming during a tense AI talent war with tech giants.
The details:
Former Tesla VP of software engineering David Lau will oversee OAI’s backend systems, revealed in an internal message from co-founder Greg Brockman.
Engineers Uday Ruddarraju and Mike Dalton join OAI’s scaling team to work on Stargate after helping build the 200,000-GPU Colossus supercomputer at xAI.
Former Meta AI researcher Angela Fan also joins the scaling team, coming amid Meta’s aggressive recruitment of OAI staff that has poached seven staffers.
Why it matters: These hires mark the first public moves for OpenAI since Meta’s hiring spree that has poached talent from across the AI leaders. It’s also a direct strike at Elon Musk’s engineering crew from xAI and Tesla — and given the past relationship between the two, it may stir the pot even further in their ongoing feud.
QUICK HITS
🛠️ Trending AI Tools
🎬 Marey - Filmmaker-focused AI video model trained on licensed content
📽️ Veo 3 - Google’s video AI, now with first-frame image + audio output
🪪 Higgsfield Soul ID - Personalized, consistent character image generation
🤝 Coachvox - Clone an AI version of yourself to coach clients in your style
💼 AI Job Opportunities
🧠 Databricks - Director, Emerging Enterprise
📈 Harvey - Senior Analytics Engineer
🗂️ Notable - Technical Program Manager
🔬 Deepmind - Data Scientist, GeminiApp, Verticals
📰 Everything else in AI today
Get up to speed on Agentic AI – learn how to build, test, and deploy AI Agents with Postman’s Rodric Rabbah in this free, on-demand webinar.*
OpenAI is set to launch its own web browser in the “coming weeks” that will challenge Google Chrome, featuring a ChatGPT-like chat interface and agentic integrations.
OpenAI will also reportedly release its highly anticipated open-source model next week, rumored to be “similar to o3 mini” with reasoning capabilities.
Microsoft CCO Judson Althoff said the company has saved over $500M in the past year from AI’s infusion in call centers, following last week’s cut of 9,000 jobs.
AI2 introduced FlexOlmo, a new language model training paradigm that enables data owners to contribute to AI development without sharing their raw data.
Google integrated Gemini into WearOS smartwatches from Pixel, Samsung, Xiaomi and more, enabling natural voice interactions and task management on the devices.
OpenAI announced that its acquisition of Jony Ive’s firm, io, has closed, with Ive and his LoveFrom team staying independent but embedded in OpenAI’s design direction.
*Sponsored Listing
COMMUNITY
🎥 Join our next live workshop
Join our next workshop this Friday, July 11th, at 4 PM EST with Dr. Alvaro Cintas, The Rundown’s AI professor. By the end of the workshop, you’ll confidently be able to install and use Gemini CLI to boost your productivity right from the command line.
RSVP here. Not a member? Join The Rundown University on a 14-day free trial.
See you soon,
Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team
Meta poaching spree hits Apple
Read Online | Sign Up | Advertise
Good morning, AI enthusiasts. Apple’s AI struggles somehow just got worse — and Mark Zuckerberg just added even more talent to Meta’s new Superintelligence unit.
With the chief of foundation models, Ruoming Pang, leaving for Meta and others rumored to follow, an AI brain drain may leave the tech giant scrambling even harder to catch up to its already distant competitors.
P.S.: Our next live workshop is today at 4 PM EST — join and learn how to confidently build and deploy scalable apps using Bolt. RSVP here.
In today’s AI rundown:
Meta poaches Apple’s AI leader
Teachers' union launches $23M AI academy
Prepare for meetings using Google Gemini Apps
Moonvalley debuts filmmaker-friendly video AI
4 new AI tools & 4 job opportunities
LATEST DEVELOPMENTS
META
🍏 Meta poaches Apple’s AI leader
Image source: Flux Kontext / The Rundown — original image via Mark Zuckerberg on Facebook
The Rundown: Meta poached Ruoming Pang, Apple’s head of foundation AI models, with a reported offer worth tens of millions — alongside other new additions to its new Superintelligence unit, including OpenAI’s Yuanzhi Li and Anthropic’s Anton Bakhtin.
The details:
Pang led Apple’s 100-person foundation models group, shaping the AI behind Apple Intelligence and next-gen Siri features.
Meta’s offer reportedly included a compensation package in the “tens of millions,” joining the Superintelligence division led by Alexandr Wang.
The Apple departure follows internal tensions after leadership explored replacing in-house AI models with options from OpenAI or Anthropic.
Pang is not alone, with Bloomberg reporting that several engineers from Apple's foundation models group plan to exit for Meta or other competitors.
Why it matters: Apple Intelligence and the AI-infused Siri have been nothing short of a disaster, but Meta is still picking off Apple talent — and leaving the tech giant even further away from any sort of internal turnaround. Meta’s poaching may be the headline, but the real story is an Apple AI brain drain coming to its lowest point.
TOGETHER WITH HUBSPOT
⛏️ The AI gold rush: Are you in or out?
The Rundown: HubSpot’s “AI For Business Builders” free guide shows you how to join the savvy entrepreneurs quietly leveraging artificial intelligence for real business growth — with no P.H.D. needed.
This no-BS guide cuts through the hype, giving you:
A jargon-free breakdown of large language models (the engine behind AI's magic)
Concrete steps to leverage AI for immediate business growth
Real-world examples that turn "prompt engineering" from gibberish into a secret weapon
AI & EDUCATION
📚 Teachers' union launches $23M AI academy
Image source: Reve / The Rundown
The Rundown: The American Federation of Teachers partnered with Microsoft, OpenAI, and Anthropic to create a national AI training hub that will prepare 400,000 educators to integrate the technology into classrooms across the U.S.
The details:
The academy will offer workshops, online courses, and professional development, with its flagship campus in NYC, and plans to scale nationally.
OpenAI is committing $10M in funding and technical support, with Microsoft and Anthropic also contributing to cover training, resources, and AI tool access.
Teachers will gain access to priority support, API credits, and early education-focused AI features, with an emphasis on accessibility for high-needs districts.
Why it matters: With AI rapidly reshaping both the workforce and the classroom, this partnership helps better equip teachers for being on the frontlines of a tech revolution for the younger generation. The academy’s success could influence how AI is taught, governed, and trusted in education for years to come.
AI TRAINING
📅 Prepare for meetings using Google Gemini Apps
The Rundown: In this tutorial, you’ll learn how to use Gemini AI to prepare for meetings — by analyzing your calendar, reviewing past emails, and researching participants to generate comprehensive briefings before every call.
Step-by-step:
Enable Gemini’s Google Workspace in settings to connect Gmail and Calendar
Ask: “Check my calendar for this week and show me all my meetings with participant details”
Review history: “Search my Gmail for previous conversations with [name] and summarize our past interactions in [timeframe]”
Get strategy: “Based on my upcoming meeting with [name], our previous chats, and current [company] context, suggest 5 questions I should prepare”
Pro Tip: After meetings, ask the AI to draft follow-up email templates based on the talking points you prepared and discussed.
PRESENTED BY WARP
🏆 Warp surpasses Claude Code on T-Bench
The Rundown: Warp just launched the first Agentic Development Environment, and it’s already surpassed Claude Code by over +20% on Terminal-Bench — also scoring 71% on SWE-bench Verified to make it the best-performing AI coding agent on the market.
Warp pulled ahead using:
Long-running commands — something no other tool can support
Agent multi-threading, running multiple agents in parallel all under your control
Deployment across the entire development lifecycle
Try Warp's SOTA agent for free — or for a limited time get 2.5x AI credits on paid plans.
MOONVALLEY
🎬 Moonvalley debuts filmmaker-friendly video AI
Image source: Moonvalley
The Rundown: Moonvalley, a startup founded by ex-DeepMind researchers, just released Marey — a filmmaker-focused AI video model trained exclusively on licensed content that gives directors granular control over scenes.
The details:
Marey is trained exclusively on licensed footage to avoid copyright issues that plague other AI startups, heavily sourced from indie filmmakers and agencies.
The model gives directors precise control over camera moves, character motion, backgrounds, and lighting, integrating directly into VFX workflows.
Pricing starts at $14.99 monthly for 100 credits, scaling up to $149.99 for 1,000 credits — with each five-second clip costing roughly $1-2 to render.
The company has raised over $100M to date and launched Marey alongside Asteria Film Co., an AI animation studio acquired by Moonvalley.
Why it matters: As Hollywood’s AI anxiety peaks, Marey is a test of whether ethical AI can win over creators. With a “clean” training set and a platform that acts as a precise director tool instead of a “luck” based video slot machine, Marey could help shift AI’s narrative around tinseltown from existential threat to creative partnership.
QUICK HITS
🛠️ Trending AI Tools
🧊 Hunyuan 3D-PolyGen - AI model for professional, art-grade 3D outputs
🤖 Proactor - A context-aware, memory-augmented “self-active” AI teammate
⚙️ Emergent 2.0 - Agentic vibe coding platform for production-ready apps
🤏 SmolLM3 - Hugging Face’s small, multilingual, long-context reasoner
📰 Everything else in AI today
Meta invested $3.5B into Ray-Ban maker EssilorLuxottica SA, giving the company a 3% stake in the world’s largest eyewear maker and expanding its AI glasses partnership.
Microsoft and Replit announced a new partnership to bring the startup’s agentic coding capabilities to Azure enterprise customers.
OpenAI ramped up its security with fingerprint scans, isolated computer environments, and military expertise hires over espionage concerns from Chinese rivals.
Google rolled out the ability to use first-frame image-to-video generations in Veo 3 with audio output, enhancing character consistency.
A U.S. diplomatic cable revealed that someone used AI to impersonate Secretary of State Marco Rubio on Signal, targeting at least five people, including foreign ministers.
IBM unveiled its next-gen Power11 chips and servers, designed for simplified AI deployment in business operations.
COMMUNITY
🎥 Join our next live workshop
Join our next workshop today at 3 PM EST with Tomek Sułkowski, Founding Engineer and DevRel Lead at Bolt. By the end of the session, you’ll confidently build and deploy scalable apps using Bolt.
RSVP here. Not a member? Join The Rundown University on a 14-day free trial.
See you soon,
Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team
Meta poaches Apple's AI head
Read Online | Sign Up | Advertise
Good morning, tech enthusiasts. Apple’s top AI architect, Ruoming Pang, is jumping ship to join Meta’s ambitious Superintelligence Labs, just as the Cupertino giant’s models struggle to impress.
Is this a critical blow to Apple’s AI plans? And could Meta’s talent grab shift the balance in the AI race?
In today’s tech rundown:
Meta poaches Apple’s head of AI
Jack Dorsey’s ‘Bitchat’ messaging app
The American version of TikTok
OpenAI’s military-grade lab lockdown
Quick hits on other major tech news
LATEST DEVELOPMENTS
META/APPLE
🍏 Meta poaches Apple’s head of AI
Image source: Ideogram/The Rundown
The Rundown: In a major win for Meta’s expanding AI ambitions, Apple’s head of AI models, Ruoming Pang, is leaving Cupertino to join Mark Zuckerberg’s newly launched AI division, Superintelligence Labs, Bloomberg reports.
The details:
At Apple, Pang played a key role in developing foundation models powering Apple Intelligence and the upcoming Siri update.
His exit comes amid mounting troubles at Apple over its AI models, which have lagged behind the offerings of OpenAI, Anthropic, and even Meta itself.
Meta has been aggressively recruiting top AI talent for its Superintelligence Labs, led by Scale AI’s Alexandr Wang and former GitHub CEO Nat Friedman.
Apple has reportedly explored licensing third-party AI models, such as those from OpenAI or Google, to supplement its technology.
Why it matters: Meta’s aggressive hiring spree — offering multimillion-dollar salaries — has become a major draw for top talent, while Apple’s AI division is reportedly grappling with internal friction. Pang’s departure could spark a broader wave of exits, as other key players may question Apple’s AI strategy and its ability to stay competitive.
JACK DORSEY
🔥 Jack Dorsey’s ‘Bitchat’ messaging app
Image source: Steve Jurvetson/Wikimedia Commons
The Rundown: Block CEO and Twitter co-founder Jack Dorsey posted on X that he spent his weekend building a new messaging app called Bitchat, which he described as a peer-to-peer chat service that operates entirely over Bluetooth mesh networks.
The details:
The app, available in beta via TestFlight (which quickly maxed out its user limit), doesn’t require the internet, phone numbers, or email addresses to use it.
It’s designed for encrypted, ephemeral communication that can persist even when traditional networks are blocked or censored.
Each device acts as both a client and a relay, allowing messages to “hop” from device to device, extending the range far beyond standard Bluetooth limits.
The app supports password-protected group chats, and it features mechanisms like timing randomization to further obfuscate user activity.
Why it matters: Dorsey, a proponent of decentralized tech like Bluesky and Nostr, called Bitchat an experiment in mesh networking, store-and-forward messaging, and end-to-end encryption. Wi-Fi Direct support is coming, enabling even richer peer-to-peer sharing offline.
TIKTOK
💃🏻 The American version of TikTok
Image source: Ideogram/The Rundown
The Rundown: TikTok is reportedly building a U.S.-only version of its app as pressure mounts on Chinese parent ByteDance, which is targeting a Sept. 5 launch — well ahead of the March 2026 shutdown deadline for the current version.
The details:
The new app is expected to operate independently from ByteDance’s global infrastructure, with all U.S. user data migrated to U.S.-based servers.
The report follows President Trump’s announcement that talks with China on a TikTok sale are starting this week, with a deal being “pretty much” in place.
U.S. users will need to download a new app, and it’s unclear if ByteDance will transfer TikTok’s recommendation algorithm to the new U.S. version.
The switchover could disrupt ad targeting, follower retention, and algorithm-driven engagement — a potential blow to creators and brands.
Why it matters: Amid the ongoing U.S.-China standoff, Trump has extended the TikTok ban, pushing the sell-by deadline to September 17. A potential buyout could top $50B. Now, the fate of TikTok’s 170M U.S. users rests on whether a stripped-down version can survive, potentially without its core algorithm.
OPENAI
🔒 OpenAI’s military-grade lab lockdown
Image source: Steve Jennings/Getty Images for TechCrunch/Wikimedia Commons
The Rundown: OpenAI has significantly tightened its security measures amid rising concerns over intellectual property theft, following the January release of a rival open-source AI model by DeepSeek, the Financial Times reports.
The details:
OpenAI’s overhaul includes biometric access controls and a “deny-by-default” internet policy that requires explicit approval for any external connections.
Proprietary tech is now isolated on offline systems, and only a tightly vetted group of employees can access or even discuss certain projects.
OpenAI claims DeepSeek used “distillation” methods to extract knowledge from its models through structured API queries, breaching usage terms.
In response, the AI giant now has stricter verification for API access, requiring government-issued IDs to unlock advanced features.
Why it matters: OpenAI is seriously locking down the lab to protect its $300B AI powerhouse as concerns rise about Chinese corporate espionage. In the AI arms race, it’s not just outsiders they’re likely worried about — it’s loose lips and poaching from their own backyard as well.
QUICK HITS
📰 Everything else in tech today
Dubai’s XPANCEO reportedly raised $250M in a Series A round of funding for its night-vision and optical zoom contact lens.
Amazon’s four-day Prime Day sales event is expected to boost online spending to an eye-watering $23.9B, a 28.4% increase from last year.
Waymo started robotaxi testing in Philadelphia and NYC, deploying sensor-equipped vehicles to map and collect data as it explores expanding its service.
Epic Games, the maker of Fortnite, reportedly withdrew its antitrust lawsuit against Samsung after the two companies reached an undisclosed settlement.
Top European executives from Mistral, ASML, Airbus, and others have requested that the European Commission delay enforcing its landmark AI act for two years.
Apple is appealing a €500M ($587M) EU fine for allegedly breaching new Big Tech regulations.
China’s BYD is set to begin assembling EVs at its new Brazilian factory as early as this month, aiming to cut imports as rising tariffs impact its largest overseas market.
Intel has begun a major round of layoffs, notifying employees individually this week, with reports indicating that thousands of jobs are on the line.
AI cloud infrastructure provider CoreWeave says it plans to buy Core Scientific, a data center infrastructure operator, in a reported all-stock deal valued at $9B.
COMMUNITY
🎥 Join our next live workshop
Join our next workshop this Wednesday, July 9th at 3 PM EST with Tomek Sułkowski, Founding Engineer and DevRel Lead at Bolt. By the end of the session, you’ll confidently build and deploy scalable apps using Bolt.
Watch it here. Not a member? Join The Rundown University on a 14-day free trial.
See you soon,
Rowan, Jennifer, and Joey—The Rundown’s editorial team
The AI aiming to 'solve all diseases'
Read Online | Sign Up | Advertise
Good morning, AI enthusiasts. "Solving all diseases" sounds like Silicon Valley hype — until you realize the company making the claim has $600M in funding and a Nobel Prize-winning technology.
Google DeepMind spinoff Isomorphic Labs is set to take its AlphaFold-powered treatments from computer simulations into human trials, chasing a vision where AI can generate cures on demand.
In today’s AI rundown:
Isomorphic Labs’ AI-created drugs near human trials
Chinese giant under fire over model copying
Get code documentation in AI coding assistants
AI takes the wheel for managerial decisions
4 new AI tools & 4 job opportunities
LATEST DEVELOPMENTS
ISOMORPHIC LABS
💊 Isomorphic Labs’ AI-created drugs near human trials
Image source: Midjourney / The Rundown
The Rundown: Alphabet's AI-powered drug discovery company, Isomorphic Labs, is preparing to start its first human clinical trials for its AI-designed cancer drugs, with an ultimate goal of “solving all diseases.”
The details:
The DeepMind spinoff has spent four years developing drugs using AlphaFold 3, an AI system for predicting protein structures and molecular interactions.
The team secured $600M in fresh funding in April, fueling both in-house drug candidates and major multi-billion dollar partnerships with Novartis and Eli Lilly.
The company envisions creating a "drug design engine" that could eventually generate treatments on demand to “solve all diseases.”
Human dosing is expected to begin soon, with oncology as the first clinical focus, and plans to license successful candidates after early trials.
Why it matters: If Isomorphic’s approach delivers, pharma’s previous trial-and-error model could give way to a faster, more precise era where AI can design new treatments that get tested via simulations before entering the lab. “Solving all diseases” is a utopian vision — but at least one Nobel Prize winner agrees that it is in sight.
TOGETHER WITH VANTA
🛡️ Live Event: Security, AI, and Trust
The Rundown: Vanta’s Trust Maturity Report proves mature security is a catalyst, not a cost center. Join Vanta and security researcher Matt Johansen on July 23 to dig into the report findings and explore what trust maturity looks like at every stage of growth.
In this live session, you’ll discover:
What’s changed (and what hasn’t) about how teams build trust through security
Deep insights from Vanta’s data research team
Tips for how high-maturity organizations are scaling security through AI
Tactical advice for getting buy-in and making measurable progress
HUAWEI & ALIBABA
🔥 Chinese giant under fire over model copying
Image source: Reve / The Rundown
The Rundown: Chinese giant Huawei’s research arm is pushing back on accusations that its new Pangu Pro model was copied from Alibaba's Qwen 2.5, coming after whistleblowers posted technical analysis showing similarities between the two systems.
The details:
A GitHub group called HonestAGI initially published (now deleted) findings accusing Pangu of having an “extraordinary correlation” with Qwen 2.5-14B.
Huawei's Noah Ark Lab denied the claims, saying Pangu was independently developed and the first system built on the company’s Ascend chips.
A whistleblower claiming to work at Huawei then posted on GitHub, alleging Pangu cloned third-party models while under pressure to catch up to rival labs.
Why it matters: The Chinese AI wave has felt more united than the deeper rivalries of closed Western leaders, but high-stakes domestic competition looks to be pushing teams towards ethical shortcuts. Will Chinese giants remain committed to the open-source push if their work is getting re-skinned by one of their biggest competitors?
AI TRAINING
🔧 Get code documentation in AI coding assistants
The Rundown: In this tutorial, you will learn how to use Context7 MCP to connect your AI coding assistant to real-time, version-specific documentation, eliminating outdated code examples and hallucinated APIs.
Step-by-step:
Open Cursor Settings and select “Tools and Integrations”
Click “Add new global MCP server” and paste the Context7 configuration URL found here
Test with a project prompt: “Create a React to-do list application. use context7”
Compare the updated, accurate code vs. generic AI responses
Pro tip: End every coding prompt with “use context7” to get real-time, version-specific documentation instead of outdated training data.
PRESENTED BY RETELLIO
🤝 Meet your AI Product Manager
The Rundown Retellio’s AI Product Manager finds the most common bugs and feature requests in your customer conversations, writes PRDs, and sends them to AI coding agents like Cursor — so you get pull requests that automatically fix real issues and get work done.
Retellio’s AI Product Manager can:
Listen to your customers across your calls, tickets, etc. to capture feedback
Surface the most important problems and draft specs on demand
Collaborate with Cursor to auto-deploy pull requests for lightning-fast bug fixes
AI RESEARCH
💼 AI takes the wheel for managerial decisions
Image source: Resume Builder
The Rundown: A new survey from Resume Builder found that 60% of managers are using AI tools to make critical business and personnel decisions, allowing the tech to determine raises, promotions, and firings with minimal oversight or training.
The details:
Resume Builder surveyed 1,342 managers and found that 78% use AI to determine raises, 77% for promotions, and 64% for terminations.
ChatGPT dominated as the primary tool for 53% of AI-using managers, followed by Microsoft Copilot at 29% and Google Gemini at 16%.
One in five managers also frequently allow AI to make final decisions without human review, despite most never receiving formal AI training or guidelines.
Nearly half of the managers were asked to evaluate whether AI could replace their team members, with 43% following through on replacements.
Why it matters: AI is already entrenched in the managerial department — but just as entry-level jobs have been the first to be automated, lower-level employees are again those being impacted by supervisors offloading decisions to ChatGPT. As models scale in intelligence, will owners automating managers out of the equation be next?
QUICK HITS
🛠️ Trending AI Tools
📄 Genspark AI Docs - Create a variety of document types with text prompts
📈 Context - AI office suite for working smarter, faster, and more efficiently
🗣️ Voicelab - High-performance API to run top open-source voice models
🔎 StepFun Diligence Check - AI-powered search with agent-verified citations
💼 AI Job Opportunities
🎨 Beautiful AI - Content Designer
📏 Figure AI - Metrology Technician
📊 Kumo - GTM Operations Manager
📣 Lakera AI - Director of Marketing
📰 Everything else in AI today
Elon Musk revealed that xAI’s highly-anticipated Grok 4 model will be released on Wednesday, July 9.
Anthropic published a Transparency Framework, pushing to require AI labs to release plans for assessing model risks, system cards, whistleblower protections, and more.
Tencent’s Hunyuan released Hunyuan 3D-PolyGen, a new 3D AI model designed for professional art-grade outputs for game development and artist modeling.
The Mayo Clinic introduced Vision Transformer, an AI system for detecting surgical-site infections quickly and accurately via photos during outpatient monitoring.
AI semiconductor startup Groq announced its first European data center in Helsinki, Finland, aiming to position its LPU chips as a cheaper alternative to Nvidia.
Several publishers filed an EU antitrust complaint against Google for its AI Overviews, saying the AI summaries are causing “significant harm” to traffic and revenue.
COMMUNITY
🎥 Join our next live workshop
Join our next workshop this Wednesday, July 9th at 3 PM EST with Tomek Sułkowski, Founding Engineer and DevRel Lead at Bolt. By the end of the session, you’ll confidently build and deploy scalable apps using Bolt.
RSVP here. Not a member? Join The Rundown University on a 14-day free trial.
See you soon,
Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team
Universal robot brain gets $105M boost
Read Online | Sign Up | Advertise
Good morning, robotics enthusiasts. Genesis AI just emerged from stealth with $105M and a moonshot plan: build one brain to power almost any type of robot.
Instead of training bots for one job at a time, Genesis is using massive simulated worlds to teach them how to learn, adapt, and act across tasks. The goal? Break the mold of brittle, single-purpose machines to create robots that can handle the real world.
In today’s robotics rundown:
Genesis nabs $105M for universal robot brain
AI-powered robot tattooist debuts in NYC
Low-energy robot eyes ‘see’ like humans
MIT’s AI robot boosts semiconductor testing
Quick hits on other robotics news
LATEST DEVELOPMENTS
GENESIS AI
🧠 Genesis nabs $105M for universal robot brain
Image source: Genesis AI
The Rundown: Vancouver-based robotics startup Genesis AI has burst out of stealth mode with $105M in fresh funding and a mission to build what it calls a “universal robotics foundation model,” all while backed by a who’s who of tech billionaires.
The details:
The startup’s funding comes with backing from Khosla Ventures, Eclipse, former Google CEO Eric Schmidt, HSG, and French billionaire Xavier Niel.
Genesis was founded last December by Zhou Xian, a robotics PhD, and Théophile Gervet, who worked on the French AI lab Mistral.
The company’s core mission is to develop a large-scale, adaptable AI that can power a wide range of robots across different industries and use cases.
Genesis says its synthetic data engine allows it to develop models faster, marking an advantage over competitors that rely on Nvidia’s software.
Why it matters: Genesis AI is tackling robotics’ core problem: fragmentation. Unlike rivals like Skild and Physical Intelligence, which lean on real-world data, Genesis is taking a simulation-first approach to build a single intelligence layer designed to work across robot types and break down the industry's task-specific silos.
BLACKDOT
🔥 AI-powered robot tattooist debuts in NYC
Image source: Blackdot
The Rundown: A high-end NYC tattoo studio is now booking sessions with an AI-fueled robotic tattooist, developed by Austin-based startup Blackdot. In the works since 2019, this cutting-edge ink machine has just received a $7M funding boost.
The details:
The arm features computer vision, a laser to measure skin thickness, and a microscope to inspect the reference dots the device sets at different depths.
The system also aims to make the tattooing experience less painful for clients by optimizing needle movement and reducing the need to wipe irritated skin.
Sophisticated algorithms guide the device as it adapts in real time to the unique topography of each client’s body for high-precision tattoos.
The latest $7M in funding comes from a blend of strategic and angel investors, crowdfunding, and venture capital.
Why it matters: Blackdot says its tattoo robot has capacities beyond the human hand, with a wide grayscale range and dots the diameter of a human hair for incredible detail and precision. Plus, it says the tech allows human artists to license their artwork for use on human skin, meaning the future is all about scale.
QUEENSLAND UNIVERSITY
👀 Low-energy robot eyes ‘see’ like humans
Image source: Queensland University of Technology
The Rundown: Scientists in Australia have created an ultra-efficient robotic vision system dubbed LENS that uses less memory than a single smartphone photo and just a tenth of the energy of standard systems — and it could work on almost any robot.
The details:
LENS features a highly compact AI model that enables location learning without relying on bulky neural networks or external processing.
At its core is Speck, a SynSense chip-and-sensor that mimics the human eye by activating solely in response to changes, ignoring static visuals.
This neuromorphic chip processes data in an event-driven, brain-inspired way, optimizing tasks like localization and mapping with minimal energy use.
Potential uses include space and underwater exploration, medical microrobots for tasks like gastrointestinal inspection, and lightweight drones.
Why it matters: This kind of low-power “eye” could be a game-changer for machines operating where every joule counts. The LENS system’s architecture, which combines neural networks, event-based vision, and on-chip learning, also marks a radical departure from the energy-hungry, cloud-dependent AI of today.
MIT
☀️ MIT’s AI robot boosts semiconductor testing
Image source: Ideogram/The Rundown
The Rundown: MIT researchers have developed a fully autonomous robotic system that can rapidly measure key properties of new semiconductor materials, which could potentially streamline the development of more powerful solar panels.
The details:
The system uses a robotic probe, guided by a machine-learning model, to identify the best spots to test a material’s photoconductance.
A specialized planning algorithm determines the most efficient path for the probe, allowing it to move quickly between points.
The system combines machine learning with knowledge from materials scientists, allowing it to make informed decisions about where to probe.
In controlled tests, the robot was able to perform over 125 unique measurements per hour, surpassing manual or previous AI-driven methods.
Why it matters: MIT says this tech enables scientists to gather data much faster and more accurately than manual methods, potentially accelerating advances in solar cells and electronics. While specialized for solar cells, the method could be adapted for other types of materials testing, expanding its impact on research and tech innovation.
QUICK HITS
📰 Everything else in robotics today
Figure CEO Brett Adcock predicts that humanoids taking over and excelling at most human jobs could fuel anxiety regarding human purpose.
K-Scale just launched their U.S.-made open-source humanoid, starting at $9K for the first 100 units, with the highest trim priced at $16K.
China’s robotics companies are reportedly now cutting out U.S. chips and producing up to 90% of the components required for humanoids domestically.
A Unitree Robotic’s G1 humanoid dubbed ‘Jake the Rizzbot’ has been roaming around the streets of Austin in a cowboy hat, giving compliments to passersby.
Yamaha Robotics has unified its subsidiaries into a single robotics-focused company to accelerate innovation and global growth in semiconductor back-end automation.
The University of East Anglia, England, just published a study that finds humans perceive robots as more humanlike when they play games together.
China’s Agibot just released new video of its X2-N humanoid switching from wheeled locomotion to bipedal walking and stair climbing.
Chinese company Robotera has introduced Q5, a humanoid with 44 degrees of freedom and capable of responding up to 10 times per second.
Tacta Systems, a California startup, has secured $75M to enhance robots’ spatial awareness with its innovative “smart nervous system.”
Market research firm Persistence reports that the U.S. mining robotics market will hit $2.69B by 2031.
COMMUNITY
🎥 Join our next live workshop
Join our next workshop this Wednesday, July 9th at 3 PM EST with Tomek Sułkowski, Founding Engineer and DevRel Lead at Bolt. By the end of the session, you’ll confidently build and deploy scalable apps using Bolt.
Watch it here. Not a member? Join The Rundown University on a 14-day free trial.
See you soon,
Rowan, Jennifer, and Joey—The Rundown’s editorial team
LLMs show signs of strategic intelligence
Read Online | Sign Up | Advertise
Good morning, AI enthusiasts. Researchers just found a way to expose the hidden personalities of AI models… and all it took was a classic game theory experiment.
Mapping how different LLMs respond to 140,000 rounds of the Prisoner's Dilemma revealed that LLMs developed distinct strategic approaches — providing a ‘fingerprint’ that suggests models are doing a lot more than just matching patterns.
In today’s AI rundown:
LLMs show signs of strategic intelligence
AI coding tool Cursor faces mass cancellations
How to build interactive 3D websites from images
Researchers game peer reviews with hidden prompts
4 new AI tools & 4 job opportunities
LATEST DEVELOPMENTS
AI RESEARCH
🧠 LLMs show signs of strategic intelligence
Image source: Reve / The Rundown
The Rundown: Researchers just tested whether AI models can be strategic reasoners by running 140,000 Prisoner's Dilemma decisions — discovering that models from OpenAI, Google, and Anthropic each developed unique strategic approaches.
The details:
Researchers ran Prisoner’s Dilemma tournaments where agents chose to cooperate or defect, earning points based on mutual choices.
Each AI generated written rationales before decisions, calculating opponent patterns and match termination probabilities that influenced their choices.
The results found distinct strategies across models, with Gemini being ruthlessly adaptive and OpenAI models acting cooperative even when exploited.
Researchers also mapped ‘fingerprints’ showing how models respond to being betrayed or succeeding, with Anthropic’s Claude being the most forgiving.
Why it matters: Seeing LLMs develop distinctive strategies while being trained on the same literature is more evidence of reasoning capabilities over just pattern matching. As models handle more high-level tasks like negotiations, resource allocation, etc., different model ‘personalities’ may lead to drastically different outcomes.
TOGETHER WITH H COMPANY
🚀 Browser automation, unlocked
The Rundown: Backed by a historic $220M seed round, H Company just open sourced Holo1 — the open-source action model behind Surfer H that’s now the top-ranked web-browsing agent on WebVoyager.
With Holo1, you can:
Automate multi-step browser workflows to reclaim hours of repetitive work
Experience SOTA accuracy that outperforms OpenAI’s Operator, Gemini Flash, & more
Integrate instantly with RAG workflows, RPA suites, and multi-agent hubs
Trim costs with full browsing flows at just $0.11 – $0.13 per run
Holo 1 is now freely available for deployment, fine-tuning, and scaling — learn more here.
CURSOR
🤯 AI coding tool Cursor faces mass cancellations
Image source: Cursor
The Rundown: AI coding platform Cursor has triggered a flurry of backlash and anger from the developer community after quietly restructuring its Pro plan, leaving users with surprise charges and quickly depleted quotas.
The details:
Cursor switched from its previous 500 requests per month to a token-based model, drastically cutting limits with limited communication of the move.
Developers reported quickly burning through token quotas under the change, with one team exhausting a $7,000 annual subscription in one day.
Social media filled with cancellation posts and threads, with users migrating to Claude Code and other alternatives over the sudden pricing changes.
Cursor published a blog admitting they "missed the mark” on communication surrounding the changes, issuing refunds for unexpected usage charges.
Why it matters: The Cursor backlash is a comms problem first, but also shows how the economics are changing with more capable, resource-intensive models — making previous quotas now become an unsustainable business plan. But with big competition in the space, any change in pricing is one wrong step away from a mass exodus.
AI TRAINING
🌐 How to build interactive 3D websites from images
The Rundown: In this tutorial, you’ll learn how to turn any image idea into a functional 3D website using ChatGPT for image generation, Hunyuan 3D for model conversion, and AI coding tools for web development.
Step-by-step:
Use ChatGPT to generate your image: “Create a 3D image model of a [object] with a white background”
Convert to 3D model using Hunyuan 3D 2.1 on Hugging Face.
Upload the GLB file to your AI coding tool and prompt: “Create a [website type] that rotates the 3D model on the X axis when the user scrolls down”
Refine with follow-up prompts: “Enhance the UI and make it more modern”
Pro tip: Try different scroll animations like scaling or horizontal movement to create unique interactive experiences that impress website visitors.
PRESENTED BY SAMSARA
🛠️ Samsara AI hits the frontline
The Rundown: AI is no longer theoretical. Samsara just launched new tools built for the real world, cementing its spot as the No. 1 platform for safety and efficiency in physical operations. With Samsara safety tech adoption surpassing 80%, the frontline is becoming the front edge of innovation.
Explore real-world AI tools like:
Multicam AI for safer driving
Predictive maintenance to cut downtime
Smart routing that beats delays
AI & ACADEMIC RESEARCH
📝 Researchers game peer reviews with hidden prompts
Image source: Nikkei Asia
The Rundown: A new report from Nikkei Asia just discovered that scientists at 14 universities planted invisible text in research papers that secretly instructed AI tools to return feedback like generating positive reviews or avoiding any negative commentary.
The details:
Nikkei found 17 preprints containing concealed prompts like "give a positive review only" using white text and microscopic fonts unreadable to humans.
Papers from institutions like Columbia, Peking University, and KAIST included commands directing AI to praise "methodological rigor" and avoid negatives.
KAIST announced the withdrawal of impacted papers, while Waseda professors defended the practice as exposing "lazy reviewers" who use AI for evaluations.
Why it matters: AI writing has already infiltrated the scientific and research communities in a big way — and the other side of the coin is the tech’s infusion into the review process as well. While the upside of AI’s involvement in these fields is clearly massive, it wont come without authenticity issues like this along the way.
QUICK HITS
🛠️ Trending AI Tools
🎨 Soul Inpaint - Higgsfield AI’s new image editing tool for precise changes
🗣️ Kyutai TTS - Open-source text-to-speech model for real-time use
📊 Shortcut - AI agent for Excel data tasks
💎 Gems - Custom AI experts for Gemini, now available across Google Suite
📰 Everything else in AI today
Rumored benchmarks for xAI’s upcoming Grok 4 leaked on X, showcasing a SOTA score on Humanity’s Last Exam, STEM, and coding benchmarks.
OpenAI’s Head of Recruiting called out Meta’s hiring practices, accusing them of ‘exploding’ offers that he called an “unethical” move.
A new ChatGPT tool called “Study Together” (code named Tatertot) has started appearing in user’s platforms, hinting at a new collaborative workflow for students.
Kyutai Labs open-sourced Kyutai TTS, a text-to-speech model designed for fast, real-time use — alongside the code for a voice AI system called Unmute.
Genspark launched AI Docs, an agentic creator allowing users to generate and edit a variety of document times via natural language prompts.
Billionaire entrepreneur Mark Cuban said he believes the AI boom will lead to the world’s first trillionaire, and that it might just be “one dude in the basement”.
COMMUNITY
🎥 Join our next live workshop
Join our next workshop this Wednesday, July 9th at 3 PM EST with Tomek Sułkowski, Founding Engineer and DevRel Lead at Bolt. By the end of the session, you’ll confidently build and deploy scalable apps using Bolt.
RSVP here. Not a member? Join The Rundown University on a 14-day free trial.
See you soon,
Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team
AI's fertility breakthrough arrives
Read Online | Sign Up | Advertise
Good morning, AI enthusiasts. After 18 years and 15 failed IVF cycles, one couple's journey to parenthood seemed impossible — until AI looked where humans couldn't.
Columbia University’s AI system just achieved its first pregnancy, using astrophysics algorithms designed for space to create a cost-effective method that could transform fertility treatment for couples worldwide.
Reminder: Our next workshop is today at 4 PM EST — join and learn how to wire MCP into both conversational AI and your favorite coding IDEs. RSVP here.
In today’s AI rundown:
AI helps a couple conceive after 18 years
Meta chatbots to message users first
Build your own AI app with Claude Artifacts
What a real 'AI Manhattan Project' could look like
4 new AI tools & 4 job opportunities
LATEST DEVELOPMENTS
AI & MEDICAL BREAKTHROUGHS
🍼 AI helps a couple conceive after 18 years
Image source: o3 / The Rundown
The Rundown: Columbia University doctors achieved the first pregnancy using an AI system called STAR, which helped a couple conceive after an 18-year struggle by discovering viable sperm in a man with severe infertility.
The details:
STAR uses AI to scan semen samples from men with azoospermia, a condition with nearly zero measurable sperm, instead of the typical 200-300M cells.
The system scanned 8M microscopic images in under an hour, locating 44 cells, whereas human technicians found zero after two days of searching.
Columbia's team developed the approach over five years, adapting astrophysics algorithms for new stars to detect microscopic reproductive cells.
STAR is only used at the Columbia University Fertility Center for now, with an estimated $3K cost compared to as high as $15-30K for a single IVF cycle.
Why it matters: Fertility rates are plunging across the globe — and for many, the costs for expensive cycles of IVF treatments (which don’t guarantee success) are an insurmountable barrier. With STAR and new AI-driven systems, doctors can hopefully provide solutions to infertility at a more accessible price to hopeful parents.
TOGETHER WITH INNOVATING WITH AI
💼 Does your business need an AI expert on call?
The Rundown: Innovating with AI has trained 1,000+ AI consultants, and now they’re launching their new directory of IWAI Certified AI experts, with exclusive early access for readers of The Rundown.
With this free directory, you can:
Connect instantly with dozens of professional AI experts
Get help with AI training, strategy, and implementation of workflows and agents
Implement modern AI solutions with no fees, commission, or lock-in
Click here to request access to the IWAI Certified Professional directory.
META
💬 Meta chatbots to message users first
Image source: …
The Rundown: Meta is training customizable AI chatbots that send unprompted messages within its messaging apps, according to a new report from Business Insider, hoping to increase engagement and user retention with the proactive companions.
The details:
Data labeling firm Aligner is helping develop the bots, which can remember past chats and maintain consistent personas like movie critics and chefs.
Chatbots created through Meta's AI Studio can initiate conversations within 14 days of user contact, requiring five prior messages to activate the feature.
Meta confirmed testing shows bots won't continue messaging without user responses, limiting outreach to one follow-up per conversation thread.
Court documents revealed Meta projects generative AI products will generate $2-3B in revenue by 2025, potentially reaching $1.4T by 2035.
Why it matters: It was only a matter of time before AI started being more proactive with messaging, but it’s an area that needs to be tread very lightly. While on the surface, it may seem more “human” to have a bot message first, it could quickly become cringey and spammy if not implemented correctly.
AI TRAINING
✍️ Build your own AI app with Claude Artifacts
The Rundown: In this tutorial, you will learn how to use Claude Artifacts' new API integration to create custom AI-powered tools directly inside Claude.
Step-by-step:
Click the artifacts button on Claude's left sidebar and then “New artifact”
Use a prompt that requests AI tool capabilities, e.g., "Create a grammar AI checker with two text areas (Original vs Fixed text) and a 'Fix Grammar' button."
Customize with additional features: "Add word count and change highlighting"
Test with sample text and save to your artifacts library for future use.
Pro tip: Your custom AI tool uses Claude via API, without you needing to set it all up.
PRESENTED BY WARP
🏆 Warp surpasses Claude Code on T-Bench
The Rundown: Warp just launched the first Agentic Development Environment, and it’s already surpassed Claude Code by over +20% on Terminal-Bench — also scoring 71% on SWE-bench Verified to make it the best performing AI coding agent on the market.
Warp pulled ahead using:
Long-running commands — something no other tool can support
Agent multi-threading, running multiple agents in parallel, all under your control
Deployment across the entire development lifecycle
Try Warp's state-of-the-art coding agent for yourself — download for free.
AI RESEARCH
🏗️ What a real 'AI Manhattan Project' could look like
Image source: Midjourney
The Rundown: Research Lab Epoch AI just published an analysis of what a U.S.-led AI Manhattan Project could look like, believing the initiative could significantly accelerate progress and achieve a 10,000x increase in AI training scale over GPT-4 by 2027.
The details:
Researchers modeled a national AI project after historical efforts like the Apollo program, involving government leadership and private-sector resources.
An investment level similar to the Apollo program's peak would fund an estimated 27M GPUs and train a model 10,000x larger than GPT-4 by late 2027.
The US-China Economic and Security Review Commission recommended a Manhattan Project AI program, calling it a top priority for achieving AGI.
Epoch estimated massive power needed, suggesting leveraging the Defense Production Act and other national efforts to speed power plant construction.
Why it matters: The AGI/ASI race is no longer just a corporate competition, but a global sprint to mobilize resources for energy and AI infrastructure. Just like major national efforts of the past (Apollo, Manhattan Project), being the first country to achieve this technological milestone will hold humanity-altering implications.
QUICK HITS
📰 Everything else in AI today
OpenAI co-founder Ilya Sutskever formally announced that he will be taking on the role of CEO for SSI, following the departure of Daniel Gross to Meta.
Together AI open-sourced DeepSWE, a coding agent that achieves SOTA results for open-weight agents on SWE-Bench-Verified for software tasks.
Higgsfield introduced Soul Inpaint, a new image editing tool allowing users to make granular changes to then combine them with video and motion control.
Replit released Dynamic Intelligence, new features for its agentic coding tool that enhance context awareness, reasoning, and autonomous behavior.
xAI’s Grok updates will reportedly include a “Games” option to build and create games, with Grok-4 expected to be released next week.
ByteDance researchers released X-UniMotion, a new framework that animates still images with extremely realistic whole-body, hand, and facial motion.
COMMUNITY
🎥 Join our next live workshop
Join our next workshop today at 4 PM EST with Dr. Alvaro Cintas, The Rundown’s AI professor. By the end of the workshop, you’ll confidently wire MCP into both conversational AI and your favorite coding IDEs.
RSVP here. Not a member? Join The Rundown University on a 14-day free trial.
See you soon,
Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team
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.