AI just made the billion-dollar solo founder real
PLUS: Turn any flat image into a fully editable design
Read Online | Sign Up | Advertise
Good morning, AI enthusiasts. It took $20K, two months, and a stack of AI tools for Matthew Gallagher to launch a telehealth startup from his house in LA. A year and a half later, Medvi is on pace to do $1.8B in sales.
Sam Altman predicted in 2024 that AI would make the solo billion-dollar company possible. Gallagher may have just delivered the proof — not by building AI himself, but by using it to move fast and replace an entire corporate workforce.
In today’s AI rundown:
AI turns solo founder into $1.8B operator
OpenAI acquires TBPN in first media deal
Turn any flat image into a fully editable design
Google’s powerful new open-source family
4 new AI tools, community workflows, and more
LATEST DEVELOPMENTS
MEDVI
🚀 AI turns solo founder into $1.8B operator

Image source: Medvi / NYT
The Rundown: Matthew Gallagher just scaled his startup, Medvi, from a $20K AI experiment to $1.8B in projected annual sales, the NYT reported — becoming one of the first to fulfill Sam Altman’s prediction of AI-driven, solo billion-dollar companies.
The details:
Medvi sells GLP-1 drugs online, outsourcing doctors, prescriptions, and shipping to telehealth platforms CareValidate and OpenLoop.
Gallagher used ChatGPT, Claude, and Grok for code, Midjourney and Runway for ad creatives, and ElevenLabs and custom AI agents for customer service.
The whole operation took two months and $20K to stand up, with the company bringing in $401M in revenue in its first year.
He then brought on his brother as the only full-time hire, and uses contract engineers and account managers, with the team on pace for $1.8B this year.
Why it matters: Altman predicted that a one-person billion-dollar company "would have been unimaginable without A.I., and now it will happen." The first real example isn't some revolutionary AI product; it's selling weight-loss drugs from a living room. AI tools, combined with strong builder instincts and action, can yield pretty wild results.
TOGETHER WITH VANTA
🛡️ Turn compliance into a competitive advantage
The Rundown: With customer expectations rising and compliance needs shifting, keeping up can be daunting. Vanta’s Agentic Trust Platform helps fast-moving startups and security teams get audit-ready fast and stay continuously compliant, turning compliance into a deal accelerator, not a blocker.
Join to learn how Vanta can help you:
Automate evidence collection, policies, and remediation across major frameworks
Build real security foundations, not check-the-box fixes
Show credibility faster with a public Trust Center and AI-powered questionnaires
Keep engineers focused with guided workflows and developer-native automation
OPENAI & TBPN
🎙️ OpenAI acquires TBPN in first media deal

Image source: Jordi Hays on X
The Rundown: OpenAI just announced the acquisition of TBPN, the daily live tech talk show that's become a go-to for Silicon Valley CEOs, in a deal reportedly worth low hundreds of millions — marking the AI giant's first media acquisition.
The details:
TBPN goes live every weekday on YouTube and X, pulling around 70K viewers per episode and hosting major tech CEOs and figures.
Fidji Simo said “the standard comms playbook just doesn’t apply to us” with OAI driving a tech shift, aiming to foster real, constructive convos on AI.
TBPN's 11-person team will report to OAI chief of global affairs Chris Lehane, and will drop its ad business but retain editorial independence on the show.
Co-founders Jordi Hays and John Coogan debuted the live show 17 months ago, with the company reportedly on pace for $30M in revenue this year.
Why it matters: This is a fascinating one, with the AI leader buying a direct channel to both the cultural tech vibes TBPN has fostered and the founders and CEOs who tune in every day. OAI’s public perception has taken hits this year, and bringing in the team behind one of the tech bubble’s most beloved shows could help shake up the approach.
AI TRAINING
🌄 Turn any flat image into a fully editable design
The Rundown: In this guide, you'll learn how to use Canva's new Magic Layers feature to turn any flat AI image into a fully editable design — enabling you to fix small details in generations without recreating entire outputs.
Step-by-step:
On the Canva homepage, select Magic Layers, click Select Media, and choose an image
Wait 30-60 seconds while Canva processes the image. It reads the layout, identifies text, objects, and background, and splits them into individual layers
Click any layer and edit it directly. Swap a date, change a tagline, fix a typo. Or click an object layer to move it, resize it, or delete it entirely
Drop in a replacement image (new product photo, logo), and it slots into the existing layout. Adjust the background color using the color picker if needed
Pro tip: To use Magic Layers on an existing design, open the design, click Uploads to add your image, select it, click Edit in the toolbar, then select Magic Layers.
PRESENTED BY GOOGLE CLOUD
🧠 Build scalable AI agents faster
The Rundown: Google Cloud's updated Startup technical guide: AI agents is your essential blueprint, providing technical founders with the practical frameworks needed to design, build, and deploy intelligent autonomous systems that scale seamlessly.
Inside the updated guide, you’ll discover:
Frameworks to design autonomous agent architectures
Best practices for prompt engineering workflows
Scalable deployment strategies using Google Cloud
💎 Google’s powerful new open-source family

Image source: Google
The Rundown: Google DeepMind rolled out its Gemma 4 family, four open models with sizes for devices from phones to computers — released under Apache 2.0 for the first time to remove legal barriers that pushed enterprises toward Qwen/Mistral instead.
The details:
All four models handle code, vision, and multi-step agent tasks, with the smallest variants adding voice and running entirely offline on a phone.
Gemma 4’s 31B and 26B models place near rivals like Kimi K2.5, GLM-5, and Qwen 3.5 in terms of intelligence while coming in at a fraction of the size.
The switch from a custom license to Apache 2.0 means developers can modify, deploy, and sell commercially with zero legal friction, a first for the Gemma line.
Why it matters: Chinese models have dominated the open-source frontier, but this week has seen two U.S. releases to challenge them: Arcee AI’s Trinity-Large and now Gemma 4. Google is trending in the opposite direction of its Chinese rivals, who are moving towards closed systems, with Gemma getting an even more permissive license.
QUICK HITS
🛠️ Trending AI Tools
🔀 Merge Gateway - Ship production AI faster. Routing, cost controls, and observability already built in*
💎 Gemma 4 - Google's new open-weight AI with SOTA intelligence for its size
🧠 Qwen3.6-Plus - Alibaba's reasoner with 1M context window, strong coding
🎧 MAI-Transcribe-1 - Microsoft's speech-to-text model for 25 languages
*Sponsored Listing
📰 Everything else in AI today
ByteDance’s Seedance 2.0 AI video generator is now broadly available across major platforms, taking the top spot on Artificial Analysis’ video leaderboards.
Cursor unveiled Cursor 3, a new rebuilt interface that lets developers run fleets of local and cloud coding agents in parallel across multiple repos from one workspace.
Alibaba released Qwen3.6-Plus, a reasoning model that rivals Opus 4.5 on coding agent benchmarks while natively supporting 1M-token context and multimodal inputs.
Microsoft launched MAI-Transcribe-1 in public preview, a new speech-to-text model that tops benchmarks on accuracy across 25 languages.
Japanese AI startup Sakana AI opened beta testing for Marlin, an autonomous AI research assistant that can work up to 8 hours straight on business-related tasks.
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 Phillip H. in Barcelona, Spain:
"I have a Claude Cowork automation that connects with my HubSpot, Slack, Gmail, Google Calendar, and Fellow AI. It runs automatically every workday at 8 AM to give me an update on anything I might have missed and what's on my plate for the day!"
How do you use AI? Tell us here.
🎓 Highlights: News, Guides & Events
Read our last AI newsletter: Dorsey makes the AI case against managers
Read our last Tech newsletter: This startup wants to grow your next body
Read our last Robotics newsletter: Waymo hits 500k weekly rides
Today’s AI tool guide: Turn any flat image into a fully editable design
See you soon,
Rowan, Joey, Zach, Shubham, and Jennifer — the humans behind The Rundown

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.








