In partnership with |
 |
|
|
2 minutes. Your URL. A customer profile worth using. |
|
Most founders can describe their product. They can't describe their customer. Not in a way that actually changes how they sell. |
HubSpot for Startups built a free tool to fix that. Paste in your URL, answer a few quick questions, and it generates a structured profile of your best-fit customer. Firmographics, buying triggers, the works. |
Takes 2 minutes. No spreadsheet required. |
Build Your Free ICP. |
Runway just gave software a pulse. Its new Solaris model skips code entirely, generating every button, menu, and screen frame-by-frame as you click and drag - the same way it generates video. |
In blind tests, people preferred Solaris's on-the-fly interfaces over ones built the traditional way with code, in the majority of matchups. If a model can improvise a working app instead of assembling one from prebuilt parts, how much of the software you use every day still needs a developer behind it? |
Today in AI: |
Runway's Solaris renders apps like video, no code MiniMax's video model powers an AI TV channel Hugging Face's $399 robot duck sells out fast |
|
|
|
What's new? Runway introduced Solaris, its first "Interface World Model" built on Gen-4.5 - a system that generates the screen you're looking at frame-by-frame in real time, instead of writing code that a browser then renders. |
What matters? |
In a study of 250 evaluators, Solaris's interfaces beat Claude Opus 5-coded UIs in 71% of natural-behavior comparisons and 61% of instruction-following tests. Because the model draws each frame on demand, developers wouldn't need to prebuild every button, menu, or screen state a user might reach. Runway says accessibility, saved state, and session recovery still need to work reliably even though there's no fixed code underneath - early access is request-only for now. |
Why it matters? |
Solaris turns the model itself into the thing responsible for what you see next, instead of a set of prebuilt screens a developer assembled ahead of time. That's a real shift in how "an app" gets made, even if today's version still needs the boring, consistent parts - a button that acts the same way five clicks later - to catch up with the flashy parts. |
|
GUIDE |
|
|
|
What's new? Fal.ai launched H3 Max, a post-trained version of MiniMax's H3 model tuned to generate video faster than you can watch it - and developer Pieter Levels used it to build Infinite Slop, a livestream that writes and renders its own episodes as viewers vote on what happens next. |
What matters? |
H3 Max costs as little as $0.0125 per second of 480p video at launch pricing, a 75% discount that ends September 7, after which the rate roughly quadruples. Levels says Infinite Slop even scans X and Hacker News to turn breaking stories into a self-generated "Slop News Network" segment, all rendered live. Running the channel around the clock, without sponsorship, would cost roughly $4,000 a day - a real number for what "infinite" AI content actually costs to keep on air. |
Why it matters? |
Video generation has been too slow and too expensive for anything beyond short demo clips, and this is the first time it's been cheap and fast enough to run as a live, audience-steered channel. Anyone curious what that speed feels like can try the same model on fal.ai's playground before the launch pricing disappears. |
|
SPONSORED BY SUPERHUMAN AI |
Go from AI overwhelmed to AI savvy professional |
|
AI will eliminate 300 million jobs in the next 5 years. |
Yours doesn't have to be one of them. |
Here's how to future-proof your career: |
Join the Superhuman AI newsletter - read by 1M+ professionals Learn AI skills in 3 mins a day Become the AI expert on your team |
Start learning AI now |
|
|
|
What's new? Hugging Face and Pollen Robotics opened pre-orders for Microduck, a 25 cm bipedal robot built for reinforcement learning, and the $399 device pulled in $2.6 million in orders on day one. |
What matters? |
Microduck ships with 15 motors, a camera, LiDAR, two IMUs, and seven pre-trained moves you can retrain and redeploy straight to the physical robot. The entire software stack - SDK, simulator, and training pipeline - is published on GitHub under an open Apache 2.0 license. Add-on packs for chargers, developer tools, and accessories are optional, keeping the base $399 price the real entry point for anyone who wants to try RL on real hardware. |
Why it matters? |
A $399 open-source robot with a full retrainable software stack lowers the bar for hobbyists to experiment with reinforcement learning outside a simulator. It's a small, playful example of a bigger trend: physical AI tools that used to require a lab now fit in a backpack. |
|
Everything else in AI |
ChatGPT connects straight to Figma now, letting you pull a design frame into chat and turn it into working UI code without leaving the conversation. |
Google extended Gemini's Omni 1.1 Flash video model to read up to 10 seconds of prior footage, letting creators chain scenes into a single 40-second clip. |
Bria released Fibo Generate 1.5, an open, JSON-native image model that cuts generation down from 50 steps to as few as 4, without losing quality. |
Complete AI Bundle bundles prompt libraries, courses, and templates into one download for readers who want to turn everyday AI use into paid skills. |
|
Essential AI Guides - Reading List: |
|
|
Let us know! |
What did you think of today's email?
Before you go, please give your feedback to help us improve the content for you!
|
|
|
Work with us |
Reach 100k+ engaged Tech Professionals, Engineers, Managers and decision makers. Join brands like MorningBrew, HubSpot, Prezi, Nike, Ahref, Roku, 1440, Superhuman, and others in showcasing your product to our audience. Get in touch now → |
