This week, Anthropic launched Claude Design and sent Figma's stock into a tailspin, Perplexity turned your Mac mini into an always-on AI agent, OpenAI gave Codex a serious glow-up, and we show you how to build something beautiful in Claude Design - no design degree required. Plus: why 74% of AI's economic value is going to just 20% of companies.

Hi! Welcome to the 45th 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.

_{~8 mins read}

/Design /Anthropic /Claude

Designing has always been a specialist sport. You needed the skills, the tools, and - if you've ever fumbled around in Figma - a whole lot of patience and a very forgiving undo button. Not anymore.

Anthropic just launched Claude Design, a new tool that lets anyone create polished designs, prototypes, pitch decks, one-pagers, and more - just by describing what you want in plain English. It's powered by Claude Opus 4.7 (their sharpest vision model yet), and is rolling out as a research preview to Pro, Max, Team, and Enterprise subscribers.

Previously, creating in Claude was hit and miss - it could write the copy, but the visual output often looked like something generated by a confused robot. That era is over. Claude Design can read your codebase or import your existing design files, automatically apply your brand colours and typography, and build a design system that stays consistent across every project. You can then refine through inline comments, direct edits, or custom sliders that Claude builds specifically for your design. When you're done, export to Canva, PowerPoint, PDF, or hand it straight to Claude Code to bring it to life as a working product.

The market noticed immediately. Figma - which holds an estimated 80–90% of the UI/UX design market - saw its stock drop nearly 7% on the day of launch. Adobe fell 2.7%, Wix dropped 4.7%. Anthropic's CPO quietly stepped off Figma's board just days before the launch, which, in hindsight, was a fairly loud signal. This isn't an isolated incident either - we've seen this pattern before: when Claude Code launched, coding tool stocks trembled. When Notion started shipping AI features, productivity incumbents felt the squeeze. AI labs are no longer just building models; they're building entire product categories, and they're doing it fast.

For most people, though? This is genuinely exciting. AI designs that don't look like a spad (South African slang fully applies here) are now within everyone's reach. Welcome to the design-for-everyone era.

/Agent /Mac /Productivity

You probably know Perplexity as the AI search engine that actually cites its sources - the one that gives you real answers with real links, rather than confidently hallucinated nonsense. It was one of the first AI tools to be permanently connected to the internet, which made it genuinely useful for research when most chatbots were stuck in the past.

But Perplexity has been quietly pivoting into something much bigger: always-on AI agents that work for you, autonomously, on your actual machine.

This week they launched Personal Computer, and it's exactly what it sounds like. It integrates directly with the Perplexity Mac app and turns your Mac - ideally a Mac mini you leave running overnight - into a persistent AI agent that can work across your local files, native apps like Mail, Messages, Calendar, Finder, and Slack, and the live web, all at once. You can even kick off tasks from your iPhone while you're on the go, and Personal Computer will pick it up on your desktop.

The use cases are delightfully practical: ask it to read your to-do list and actually do it, sort your chaotic Downloads folder into sensible project folders, or compare local files against live information on the web. It can coordinate across more than 20 AI models to handle complex multi-step tasks, and every action it takes is auditable and reversible - there's even a kill switch, which is reassuring when you're handing an AI the keys to your file system.

The catch is pricing: Personal Computer is currently only available to Perplexity Max subscribers at $200/month (~R3,700). That's not cheap. But as a preview of where personal computing is heading, it's a glimpse worth paying attention to.

/AI /Coding /Agent

While Anthropic is reinventing design and Perplexity is reinventing your desktop, OpenAI isn't sitting still. Their developer tool Codex just received a significant upgrade - and it's moving squarely into "autonomous agent" territory.

Codex can now operate your computer the way a human would: seeing, clicking, and typing with its own cursor. Multiple Codex agents can run in parallel on your Mac without interrupting your own work. It now has an in-app browser for iterating on frontend designs, can generate and refine images, and connects to 90+ new plugins - including Jira, GitLab, CircleCI, Microsoft Suite, and more. It can also schedule work for itself, meaning you can kick off a task today and Codex will wake up tomorrow (or next week) to continue it autonomously.

The memory feature is particularly clever: Codex now remembers your preferences and previous corrections, so it gets progressively better at your specific workflow. Think of it less like a coding assistant and more like a self-improving junior developer who never sleeps, never complains, and never charges overtime.

OpenAI is clearly watching what Anthropic did with Claude Cowork and what Perplexity did with Personal Computer - and decided they wanted a slice of that always-on agent pie. The race to own your full desktop attention is officially on.

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

Claude Design is live for Pro and Max subscribers, and the outputs are genuinely impressive. You don't need a style guide, a mood board, or a single hour of design experience. You just need a description and a bit of curiosity.

Here's a prompt making the rounds that consistently produces beautiful results - adapt it to whatever you're working on:

Don't believe us? Try it yourself…

⏱️ Time needed: 10–15 minutes. Output: a polished design you'd actually be proud to share.

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

If this week's launches taught us anything, it's that AI agents are no longer a future promise - they're a present reality. And the companies moving fastest are pulling very, very far ahead.

A PwC study published this week - based on interviews with 1,217 senior executives across 25 industries - found that 74% of AI's economic value is being captured by just 20% of companies. The divide between AI leaders and the rest is widening sharply, and it has almost nothing to do with budget.

The leaders aren't simply spending more on AI tools. They're using AI as a catalyst for growth and business reinvention - deploying autonomous agents to open up entirely new revenue streams, not just to shave costs off existing processes. These are the companies building agentic workflows into sales, product development, and customer experience, and they're measuring the results in actual rand (and dollars).

The implication for everyone else is stark: if your AI strategy is still limited to "use ChatGPT to write emails faster," you're in the 80%. This week's launches - Claude Design, Perplexity Personal Computer, and Codex - aren't coincidences. They're the infrastructure the top 20% will be using to pull even further ahead. The question isn't whether agents are coming to work. They're already here. The question is whether your business is ready.

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 👋