Hi! Welcome to the 35th 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.

_{~5 mins read}

/Anthropic /Opus /Enterprise /Coding

Anthropic just dropped Claude Opus 4.6, and it's not just an incremental upgrade - this is the best AI model in the world right now. On professional work tasks, Opus 4.6 is beating every other model on the market, including OpenAI's latest GPT-5.2. The gap? A massive 144 Elo points ahead of OpenAI's best offering.

What makes this release special isn't just the raw performance numbers. Opus 4.6 can now handle teams of AI agents working together, coordinate complex projects across multiple agents, and sustain focused work over much longer sessions without losing the plot. Think of it like upgrading from a solo freelancer to a coordinated team that never sleeps.

The model achieved something remarkable that demonstrates its capabilities: it autonomously built a working C compiler from scratch. For non-technical folks, that's like asking someone to build a car engine without instructions - it requires deep understanding, careful planning, and the ability to catch and fix your own mistakes. Previous AI models would get lost partway through such complex tasks.

Here's what matters for businesses: Opus 4.6 doesn't just write better code. It can run financial analyses, prepare documents and presentations, conduct research, and handle the kind of knowledge work that traditionally required skilled professionals. And with its new 1-million token context window, it can work with massive amounts of information - equivalent to about 750,000 words - without forgetting what it read at the beginning.

/OpenAI /Enterprise /Agents /Platform

While companies have been scrambling to deploy AI, most are hitting the same wall: isolated AI tools that don't talk to each other, don't understand company context, and can't actually do real work at scale. OpenAI just announced Frontier, and it might be the solution that unlocks the efficiency gains everyone's been chasing.

Here's the problem Frontier solves: 75% of enterprise workers say AI helps them do tasks they couldn't do before, but companies struggle to scale this beyond individual use cases. Every AI agent lives in its own silo, doesn't understand how the business actually works, and needs constant hand-holding. Frontier changes this by giving AI agents what human employees get: shared context, proper onboarding, learning through feedback, and clear permissions.

Think of it this way - when you hire a new employee, you don't just give them a laptop and hope for the best. You explain how the company works, where information lives, what good decisions look like, and what they're allowed to do. Frontier does exactly this for AI agents, creating a "semantic layer" that all agents can reference to understand the business.

The early results are staggering. A manufacturer reduced production optimisation from six weeks to one day. A global investment firm freed up over 90% more time for salespeople to spend with customers by deploying agents across the entire sales process. An energy producer increased output by up to 5% - that's over a billion rand in additional revenue.

What makes Frontier particularly powerful is that it works with the systems companies already have - no need to rip out your existing tech stack. It connects siloed data warehouses, CRM systems, and internal applications, giving AI agents the full picture they need to make good decisions.

/Anthropic /PowerPoint /Enterprise /Productivity

Microsoft's Copilot promised to revolutionise how we work with Office applications, but most users found it underwhelming. Anthropic just released Claude in PowerPoint, and it's doing what Copilot couldn't: actually making presentations easier.

Claude in PowerPoint reads your templates, understands your brand guidelines, respects your slide masters, and makes changes that actually stay on-brand. You can start with a blank deck and say, "Create 10 slides walking through our market assessment", and Claude builds a professional draft with logical structure. Then you can refine: "make slide 3 more visual," "combine slides 5 and 6," or "turn these bullet points into a process flow diagram."

What sets this apart is precision. You can point Claude to exactly what you want changed - a specific slide, a particular element - and it makes surgical edits without regenerating the entire deck. It creates native PowerPoint charts and diagrams that you can edit directly, not static images you're stuck with.

The tool is currently available to Max, Team, and Enterprise plan customers in research preview. For consultants, analysts, and anyone who spends hours building slide decks, this could be a genuine productivity multiplier.

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

You know that feeling when you do the same task every week - formatting reports, processing data, drafting similar emails - and think "there has to be a better way"? Claude Skills is that better way.

Claude Skills lets you teach Claude to automate repetitive workflows by creating custom "skills" that capture your exact process. Instead of explaining the same task over and over, you define it once, and Claude handles it from then on. Think of it like training a really smart assistant who never forgets your preferences.

Important: You'll need Claude Desktop for this, which currently only works on Mac 🍏.

Here's a perfect example: Let's say you run a weekly sales report where you pull data from a spreadsheet, calculate key metrics, format it in a specific way, and email it to your team. You follow the exact same process every Monday morning.

Here's how to set it up:

Step 1: Enable Claude Skills

Step 2: Define Your Repetitive Problem

Step 3: Create the Skill

Step 4: Use Your New Skill

Pro Tips:

Skills turn Claude from a helpful chatbot into a personalised automation engine. The 10x output claim? It's real when you're no longer manually doing the same task dozens of times.

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

Something unprecedented happened this week: a single AI product announcement triggered one of the largest sell-offs in software company history. When Anthropic introduced Claude Cowork - particularly its legal and financial work plugins - investors panicked, and billions of rand evaporated from stock markets within hours.

The carnage was brutal. Analysis companies like Gartner and S&P Global plummeted 21% and 11%, respectively. Financial data providers FactSet dropped 10%, while Moody's and Nasdaq saw sharp declines. Major software companies - the aristocracy of enterprise tech - went into freefall: SAP lost over a third of its value from its peak, Microsoft dropped 9% year-to-date and is trading 21% below its annual high. Even advertising giants like Publicis (-9%), WPP (-12%), and Omnicom (-11%) weren't spared.

A JPMorgan index tracking US software stocks dropped 7% in a single day and has now accumulated 18% losses for the year. Private equity firms that bet big on software - Ares Management, KKR, Apollo - all saw significant losses.\

Why is this happening? The traditional Software-as-a-Service (SaaS) business model is built on a simple premise: more employees = more licenses = more revenue. But if AI agents can do the work of entire departments, why would companies need dozens of Salesforce licenses? If Claude reviews and drafts contracts autonomously, DocuSign subscriptions become less essential. If AI generates designs in seconds, Adobe license requirements drop.

Analysts from Mizuho Securities noted that many institutional investors currently see no reason to hold software stocks - regardless of how cheap they've become. The fear is existential: these companies might face an explosive increase in user productivity coupled with a collapse in their licensing revenues.

The bigger picture: This market reaction reveals two critical truths about where we are in the AI revolution. First, AI is no longer a distant threat to traditional business models - it's actively disrupting them right now. Second, for anyone running a business or building a career, staying informed about AI developments isn't optional anymore. Market movements worth billions of rand are now being triggered by AI product launches.

When a mid-week product update from an AI company can wipe out more value than most market crashes, it's a signal: the AI transformation isn't coming, it's here. And whether you're an investor, a business owner, or an employee, understanding these shifts could be the difference between riding the wave and being swept away by it.

The analysts at Gartner might be right that predictions of the "SaaSapocalypse" are premature - large organisations don't change overnight. But the direction of travel is clear, and this week's market reaction is a warning shot: the future is arriving faster than most people expected.

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 👋