🔈 Listen on Spotify

Hi! Welcome to the 55th 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}

/OpenAI /FrontierAI /Access

OpenAI pulled the wraps off a new flagship family this week: meet GPT‑5.6 Sol, Terra, and Luna. Sol is the new top dog - OpenAI calls it a “step function” better than GPT‑5.5, setting a new state of the art on Terminal‑Bench 2.1 (think complex command‑line workflows that need planning, iteration and tool juggling). Terra is the balanced everyday model at roughly half the cost of GPT‑5.5, and Luna is the cheap‑and‑fast one for high‑volume work.

So far, so exciting. Here’s the catch: you can’t use any of it. OpenAI is launching with a limited preview to “a small group of trusted partners” - and crucially, they say this is at the request of the U.S. government. General availability is promised “in the coming weeks,” but for now, frontier intelligence is gated to a connected few.

Why the velvet rope? Sol is also OpenAI’s most capable model yet for cybersecurity - it pushes the frontier on “long‑horizon security tasks including vulnerability research and exploitation.” Read that twice. When a government quietly asks you to not release your best model to the public, it starts to feel like the model crossed a threshold, too capable, and maybe too dangerous, to hand to everyone at once. Whatever the reason, the message is hard to miss: the bleeding edge of AI is no longer guaranteed to be available to those who aren’t inside the room.

/Claude /Slack /Agents

Until now, working with Claude meant opening a separate app. Anthropic just changed that with Claude Tag - a new way for whole teams to work with Claude inside Slack.

Here’s the magic: Claude joins your workspace as a team member with access to the channels and tools you choose. Then you just @‑tag it the same way you’d tag a colleague - “@Claude pull last month’s sales numbers” or “@Claude draft a first version of this feature” - and it breaks the task into steps, does the work, and tags you back in the thread when it’s done. Everyone in the channel sees the same Claude, so teammates can pick up where someone else left off.

It runs on Opus 4.8 and is in beta for Claude Enterprise and Team customers for now. The eyebrow‑raising stat: Anthropic says 65% of its own product team’s code already flows through its internal version of this tool. If that holds up beyond Anthropic’s walls, “tagging” an AI teammate could become as normal as tagging a human one.

/Microsoft /Excel /Finance

There’s a pattern to how new tech reaches the world: developers first, finance next. Microsoft just leaned hard into that with Skills for Copilot in Excel - a way for teams to bottle their financial expertise and reuse it across workbooks.

A skill gives Copilot a defined process, structure and formatting rules for a recurring task, so it does it the same way every time instead of you re‑prompting from scratch. Think building a discounted cash flow model, closing the books, refreshing a monthly forecast, or preparing a variance analysis. (Sound familiar? Skills are written in an open‑standard SKILL.md file saved to OneDrive - the same idea we use to run this newsletter.)

Microsoft also wired in trusted financial‑data connectors - CB Insights, Daloopa, FactSet, Morningstar, PitchBook and S&P Global - so analysis starts from real market data, not a manual copy‑paste. And because finance lives and dies on showing your work, Copilot can now propose a plan before it touches anything, then log every edit it makes in Excel’s Show Changes panel. That’s the bar AI has to clear in finance: trusted data, traceable numbers, no black boxes.

One honest caveat: the fancy data connectors may need separate licensing from each provider, and custom skills are still rolling out (Insiders now, broader availability next month).

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

If you’re a runner, cyclist or weekend warrior, your Strava history is a goldmine - but squinting at charts only gets you so far. Strava launched an official MCP connector that plugs your training data straight into Claude, so you can ask questions in plain English and get answers from your actual activity history.

One heads‑up before you start: this needs a paid Strava subscription, and the connection is read‑only (Claude can analyse your data but can’t change anything), and you can revoke it anytime from your Strava settings.

Don’t believe us? Try it yourself…

Here’s a prompt to get you going:

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

Investec just became the first organisation in South Africa to publicly announce a full‑workforce Microsoft Copilot rollout - every one of its 8,000 colleagues across South Africa, the UK and other international markets now has access.

This isn’t a standing start. Investec already runs 800+ AI agents across the Group, automating repetitive tasks and accelerating knowledge work - and by its own numbers, those agents are already freeing up more than 350,000 colleague hours a year, capacity it says is being reinvested into client service, advisory work and growth. The strategy has a clear thesis, summed up by Global Head of Digital and Technology Lyndon Subroyen: “higher tech leads to higher touch.” The idea is that AI is an amplifier of people and their judgement, not a substitute, by taking on routine drafting, reconciliation and admin, it frees teams to spend more time with clients and on the nuanced work that actually creates value. Investec is pairing the tools with company‑wide training and a formal AI governance framework, and is already pushing into agentic systems that execute tasks end‑to‑end under human oversight. Microsoft, its strategic partner here, called it a milestone for AI adoption in South Africa - proof of what happens when AI is treated as a business‑transformation opportunity rather than a tech install.

Our take? Well done, Investec, giving 8,000 people access to paid tools is genuinely the right first move. But access is step one, not the whole game. A rollout this size runs into real money (think tens of millions of rand a year), and tools alone tend to deliver a small productivity bump rather than real business value. The harder, more valuable unlocks come next: training people to apply AI to their actual workflows, moving beyond chat into coding harnesses like Claude Code and Cowork, and building an org‑wide context layer that captures your real processes and SOPs, because a powerful agent pointed at messy, ambiguous work just scales the mess. And eventually the truly difficult part surfaces: none of it sits neatly on top of an organisation built for a pre‑AI world. Like factory floors during the industrial revolution, the whole structure may need rethinking. Implementing AI across an enterprise might be the hardest thing in business right now, the tech learning curve is real, but the change management is the mountain. We’ll be watching this one closely. 👀

Bottom line: Access is the easy 10%. The real work - training, context, harnesses, and reshaping how the org actually operates - is the other 90%.

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 👋