Google Stitch gets a serious upgrade and becomes the easiest design tool on the internet, Claude can now run scheduled tasks from the cloud so it works even while you sleep, NVIDIA builds the safety net that AI agents desperately need - and Peter Thiel just backed a startup putting AI collars on cows at a $2 billion valuation. What a week.

Hi! Welcome to the 41st edition of the TomorrowToday newsletter.

We’re here to decode the AI chaos so you don't have to. Think of us as your friendly neighbourhood tech translators - we cut through the chaos, translate the jargon, and spotlight new AI tools that matter for founders, builders, and curious minds.

Buckle up, because the future's moving fast and we're here to make sure you don't get left behind! ⚡

If you enjoyed today’s newsletter, please forward it to a friend & subscribe by following this link.

_{~7 mins read}

/Design /Google /NoCode

If you've ever stared at a blank Figma canvas and thought, "Where do I even start?", Stitch might just be your new best friend. Google's AI-native design tool just dropped a major update, and it's essentially Pinterest, a web designer, and a brand consultant rolled into one - and it requires zero design experience to use.

The five headline upgrades are worth knowing. There's an infinite canvas that holds images, code, and text all at once, so you can throw in your brand colours, a competitor screenshot, and a product brief and let the AI make sense of it all. The smarter design agent now understands your entire canvas context, meaning you can ask it to "swap out the logo for the one I just uploaded" and it actually knows what you mean. There's also voice design (yes, you can literally talk your way through a redesign), instant prototypes at the click of a button, and a new consistency feature called DESIGN.md.

DESIGN.md is genuinely clever and worth understanding. Think of it as a plain-text document - readable by both humans and AI - that captures all the visual rules of your project: your colour palette, your typography, your button styles, your spacing. Every new screen the AI generates automatically follows these rules, so your designs look like they were made by the same person, even if you're building session by session with no memory of what came before. You can generate it from a prompt, import it from an existing website URL, or write it by hand. It's the design equivalent of handing someone a brand bible before they start work.

Bottom line: if you've been avoiding design because it felt too technical, Stitch just removed most of those barriers. Watch the walkthrough video here.

/Anthropic /Automation /Productivity

One of the things that's always made autonomous AI exciting is the idea of an assistant that's truly always on - one that doesn't just wait for you to ask it something, but runs tasks on a schedule and reports back. The problem? Setting that up has always been a pain. You'd need to leave your laptop running, keep Claude Code open, and hope nothing fell over overnight.

Anthropic just fixed that. Claude now supports cloud-hosted scheduled tasks, meaning you can tell Claude to do something on a recurring schedule and it will go do it - from the cloud, without you needing to keep anything open on your machine. Set a repo, a schedule, and a prompt, and Claude handles the rest.

The practical implications are significant. You could have Claude pull your Slack messages and GitHub activity every morning and compile a daily standup summary. You could set it to monitor a news source and send you a weekly briefing on a specific topic. You could build a recurring data pipeline, an automated project check-in, or a weekly report - all without writing a single line of infrastructure code. It's the kind of always-on automation that used to require a developer and a server. Now it's a few prompts away. This is available on Pro and Max plans via the Cowork feature. Read more here.

/NVIDIA /Agents /Security

At GTC last week, Jensen Huang made a bold claim: OpenClaw is becoming the operating system for AI agents. OpenClaw is a universal runtime for autonomous AI agents - the kind that can take actions, write code, use tools, and complete long-horizon tasks without you holding their hand the whole time. The hype is real.

But here's the thing: anyone who's actually tried deploying autonomous agents in a real business environment knows the problem. Agents that can take actions can also take the wrong actions, expose private data, or behave in ways you didn't anticipate. It's genuinely hard to do safely, and that's kept a lot of businesses on the sidelines.

That's exactly what NVIDIA NemoClaw is designed to solve. It's an open-source security and privacy layer that wraps around OpenClaw and adds policy-based guardrails - essentially a set of rules that controls how your agent behaves and what it's allowed to touch. It also evaluates your available hardware and, where possible, runs AI models locally using NVIDIA's Nemotron models, which is better for both privacy and running costs. One command to install, and your agent deployment gets a serious safety upgrade. Check it out on GitHub.

In this section, we aim to spotlight an incredible AI tool or use case and guide you on how you can try it.

This week's challenge: Design a landing page for your own matcha brand using Google Stitch 🍵

For this week's challenge, we're using Google Stitch to design a landing page for a fictional matcha brand called Matcha Do About Nothing (yes, we're very proud of that name). Think clean sage green aesthetic, minimalist typography, modern product cards - the kind of thing that would make a Joburg coffee shop owner weep with envy.

The best part? You don't need any design skills whatsoever. Here's how:

Pro tip: Once you have a DESIGN.md, you can import it into any future Stitch project to keep your brand consistent - even if you're designing something completely new weeks later.

In this section, we're spotlighting real businesses using AI to solve actual problems.

When the world's most connected investor backs AI cow collars, you pay attention. 🐮

If you don't know Peter Thiel, here's the short version: he co-founded PayPal with Elon Musk, wrote the first cheque into Facebook, built Palantir (the AI data analytics company now deeply embedded in governments and intelligence agencies worldwide), and runs Founders Fund - arguably the venture capital firm with the sharpest eye for genuinely transformative companies. When Thiel backs something, it usually means the idea is either brilliant, insane, or both. This week, it's both.

Halter, a New Zealand-based agritech startup, just closed a new funding round led by Founders Fund at a $2 billion valuation - and what they make is exactly as wild as it sounds. AI-powered solar collars for cows.

The collars track each animal's location, health, and behaviour in real time, feeding data back to a farmer's smartphone. The really clever part is the virtual fencing system: instead of physical fences, farmers draw boundaries on a map in the app, and the collars guide cows away from restricted areas using vibrations and audio cues. To bring the entire herd in for milking, a farmer taps a single button. The cows come. That's it. The technology is already running on over 600,000 collars across farms in New Zealand, Australia, and the United States.

The business case is solid: less labour, better land management, healthier herds, and real-time visibility over an asset class - livestock - that has historically been one of the hardest to monitor at scale. The deal is reportedly heavily oversubscribed, meaning investors are queuing up.

The lesson for founders: some of the most defensible AI businesses are being built in industries that look boring from the outside. Wherever there's a manual, analogue process happening at scale, there's probably a billion-rand opportunity hiding in plain sight.

AI is full of jargon, and we’re here to decode it. Each week, we’ll give you a plain-English definition of a buzzy term you’ve probably seen (but never fully understood).

Thanks for reading TomorrowToday! We’d love to hear from you:

➡️ What would you like us to cover next?
➡️ Have a tool or topic we should feature?

We’re building this with (and for) you. 🚀
See you next Tuesday 👋