Hi! Welcome to the 7th 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.

_{~6 mins read}

/Crypto /Finance

If you're a crypto enthusiast who wants to understand exactly what's happening in the market - and you don't just want to buy the latest meme coin that popped up in your MetaMask wallet - Perplexity has the answer for you.

Perplexity Finance just launched its Coinbase integration, and honestly, it's like having a Bloomberg terminal that actually speaks human. You can now ask natural language questions like "Why is Bitcoin mooning today?" or "What's driving Ethereum's on-chain activity?" and get real-time, AI-powered explanations backed by institutional-grade data from Coinbase.

The magic happens through Perplexity's Comet browser, where you can "double-click" on any sudden price movement and instantly get context without juggling between CoinGecko, Twitter, and seventeen other crypto tabs. No more frantically Googling "why crypto dump" at 3am while watching your portfolio turn red.

Coinbase CEO Brian Armstrong dropped an interesting stat: crypto searches on Perplexity now match traditional equity searches. That's a pretty clear signal that digital assets are officially mainstream, whether traditional finance likes it or not.

🔗 If this was interesting go play with perplexity finance - the crypto bro’s Bloomberg terminal.

/Productivity /Coding

While everyone's been obsessing over the latest OpenAI drama, Alibaba quietly dropped Qwen3 - and it's genuinely impressive. This isn't just another "competitive" model; it's 7x cheaper than Claude Sonnet 4 and performs almost as well for coding tasks.

Here's the kicker: Qwen3 uses something called "mixture of experts" architecture, which basically means it has multiple specialised mini-models that tag in and out depending on what you're asking. Need to write code? The coding expert jumps in. Want to solve math problems? The math expert takes over. It's like having a whole team of specialists instead of one overworked generalist.

The 235 billion parameter version can switch between thinking mode for complex reasoning and fast mode for quick responses. Plus, it handles 256K token contexts, which means you can feed it an entire codebase and it'll actually remember what you're talking about.

The real story here isn't just another AI model launch. Chinese companies are proving that open-source can compete with - and sometimes beat - the big closed models from Silicon Valley. And they're doing it at a fraction of the cost.

🔗 Not everything that is “Made in China” is poor quality, go try out Qwen yourself by following this link.

/Design /Integration /Automation

Remember when you needed actual design skills to make something that didn't look like it was cobbled together in Microsoft Paint? Those days are officially over. Claude can now create, edit, and manage Canva designs directly through chat - no switching between apps, no downloading files, no tears.

You can literally tell Claude "resize my sales presentation for Instagram Stories" or "create a LinkedIn post design for our Q2 results", and it'll handle everything in Canva behind the scenes. It's like having a designer on speed dial who never sleeps and doesn't judge your terrible colour choices.

This is part of Claude's broader conquest of productivity apps. They've already integrated with Notion, Stripe, and Figma, turning Claude into the ultimate AI assistant that can actually do stuff instead of just talking about doing stuff. The goal seems clear: make Claude the central hub for all your creative and business workflows.

The catch? You need paid subscriptions to both Claude and Canva. But if you're running any kind of business that needs regular content creation, this integration just paid for itself.

🔗 Don’t believe us, see the demo Anthropic posted on x.

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: Learn how to prompt AI properly

It's cool to play with AI tools and generate funny images of cats wearing business suits, but for AI to be genuinely useful in your work, you need to know how to prompt it properly. Think of prompting like learning to Google effectively - the better you get at asking, the better answers you receive.

Most people's prompts look like this: "Write me a marketing email." And then they wonder why the output sounds like it was written by a robot having an existential crisis.

Here's how to level up your prompting game 🚀

For beginners: Start with Google's comprehensive prompting guide. It covers the basics of how to structure requests, provide context, and get consistent results. Download it, read it over coffee, and watch your AI interactions improve dramatically.

Download Google's Prompting Guide 101

For intermediates: Anthropic created an interactive tutorial that you can run on your computer. It's hands-on, practical, and shows you exactly how small changes in prompting can lead to dramatically different outputs. You'll need an Anthropic API key, but the learning is worth it.

Try Anthropic's Interactive Prompting Course

And if you are advanced, please reach out to us via email/LinkedIn to teach us your ways! 💡

The difference between someone who knows how to prompt and someone who doesn't is the difference between having a super-powered assistant and having a very expensive paperweight.

Take the time to learn this skill - your future self will thank you.

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).

In this section, we aim to spotlight something weird & wonderful in the world of AI.

This week: One exhausted human beats OpenAI in a 10-hour coding marathon

In what might be the most David vs. Goliath moment in recent AI history, Polish programmer Przemysław Dębiak (known as "Psyho") defeated OpenAI's latest coding model in the AtCoder World Tour Finals 2025.

Picture this: after three days of competitions and barely any sleep, this absolute legend sat down for a 10-hour coding death match against a specially trained AI model comparable to OpenAI's o3. The challenge? Solve complex optimisation problems that had no perfect solutions - just find the best possible approximations within 10 hours.

The result? Humanity wins! Dębiak scored around 1.81 trillion points while the AI managed 1.65 trillion - a narrow but symbolic 9.5% victory margin. The AI still placed second out of twelve participants, but second place is first loser when you're fighting for the future of human relevance.

Dębiak's exhausted victory tweet said it all: "Humanity has won (for now)." That "for now" is doing a lot of heavy lifting, but we'll take the win. Sometimes the most important battles are won not by being vastly superior, but by being just slightly better when it counts.

It's a reminder that while AI is incredibly powerful, human creativity, intuition, and sheer bloody-minded determination will always have their place.

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 👋