Claude is shipping features faster than you can read about them, Perplexity just launched something that wants to replace your entire team, and a Twitter founder just cut 4,000 jobs while posting his best quarter ever. Buckle up.

Hi! Welcome to the 38th 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 /Claude /CoWork /ClaudeCode

If you blinked this week, you missed four new Anthropic features. Let's run through them fast - because honestly, that's the energy Anthropic is bringing right now.

Cowork gets scheduled tasks. Claude can now run recurring jobs automatically - a morning briefing when you wake up, a weekly spreadsheet that updates itself, a Friday summary sent to your team without you lifting a finger. It works even better with plugins, which give Cowork specialist knowledge across design, engineering, operations, and more. There's also a new “Customise” tab in the sidebar where you can manage all your plugins, skills, and connectors in one place.

Auto-memory is live. Claude now remembers things across sessions - your project context, your preferred ways of working, patterns it's noticed from past conversations. You don't write anything down. It just learns. This is a big deal for anyone who's been frustrated by having to re-explain themselves every single session.

Claude Code gets Remote Control. Start a task in your terminal, then leave your desk. Walk to a meeting. Take a lunch break. Claude keeps working on your machine, and you can check in, redirect, or approve decisions from your phone via the Claude app or claude.ai/code. The era of leaving your laptop running while you go to a meeting just got a whole lot more deliberate.

You can now bring your memory from other AI models. Spent months teaching ChatGPT or Gemini how you work - your tone, your projects, your preferences? You no longer have to start from scratch when switching to Claude. Head to claude.com/import-memory, copy a provided prompt into your current AI, paste the results into Claude's memory settings, and you're done. Your first Claude conversation will feel like your hundredth. Available on all paid plans.

We are not sure what they are feeding the developers at Anthropic, but one thing is sure: the speed at which they are shipping is very impressive. The AI race is heating up, and consumers are the ones who are benefiting.

/Perplexity /AIAgents /Automation

Most AI tools answer questions. Perplexity Computer wants to do the entire job.

Launched this week, Perplexity Computer is best described as an autonomous digital worker - the kind you'd hire for research, analysis, design, coding, and project management, except it never sleeps, never calls in sick, and runs on credits instead of a salary. The name is intentional: Perplexity's CEO, Aravind Srinivas, points out that the word "computer" originally meant a person who performed complex calculations. He's reclaiming that original meaning - an AI that takes a massive workload and gets it done.

Here's what makes it different from every other AI tool out there: instead of relying on a single model, it orchestrates 19 or more AI models simultaneously. Claude Opus 4.6 handles deep reasoning. Gemini covers research. GPT-5.2 holds long-context memory. Specialist models handle images and video. Think of it less like a chatbot and more like a well-staffed agency working asynchronously in the background while you focus elsewhere.

You give it a high-level goal - "build me a market analysis for the South African logistics sector" - and it breaks that down into dozens of sub-tasks, executes them inside a sandboxed cloud environment with a real browser and file system, and only comes back when it needs a decision or credential. It can work for hours. Months, even.

There's no flat subscription - it runs on a credit system. But before you commit, you can browse free live examples of what it's capable of at perplexity.ai/computer/live. Worth five minutes of your time.

/Google /NanoBanana /ImageGeneration /Design

Google dropped Nano Banana 2 this week - an image generation model that combines the quality of their top-tier Pro model with the speed of their Flash tier. For most people, that means one thing: you describe an image, and it's done before you've had time to second-guess your prompt.

What's genuinely impressive:

Nano Banana 2 is now the default across the Gemini app, Google Search (AI Mode and Lens), and developer tools. Pro and Ultra subscribers keep access to Nano Banana Pro for precision tasks.

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: Generate a stunning image with Nano Banana 2 (free, takes 2 minutes)

South African AI creative lead Jacques Pienaar tested Nano Banana Pro head-to-head against Nano Banana 2 this week - check his comparison on LinkedIn. His verdict? Faster, snappier, and holds its own on most creative tasks.

Here's how to try it yourself right now, completely free:

Need a prompt to start with? Try this Cape Town-inspired one:

Pro tip: The more specific you are about lighting, angle, style, and setting, the sharper the result. Nano Banana 2 responds particularly well to photography-style language - "golden hour lighting", "low angle shot", "editorial photography" all push the output noticeably further.

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

Block cuts 4,000 jobs after its best quarter ever - and the AI agent is why… or is it?

Block - the fintech company founded by Jack Dorsey (yes, the Twitter guy) - just announced it's reducing from 10,000 employees to just under 6,000. The timing is jarring: this came immediately after Block reported its strongest quarter in company history. Q4 gross profit hit $2.87 billion, up 24% year-on-year. Cash App grew 33%. Operating income went from $13 million to $485 million in twelve months.

So what happened?

Dorsey wrote a memo to the company that was unusually direct. He didn't claim the business was struggling. He said the opposite: "Something has changed." Block had quietly built an open-source AI agent called Goose - powered by Anthropic's Model Context Protocol - and rolled it out across the entire organisation. Engineers reported that 90% of their code was now written by Goose. Non-technical teams were using it to write SQL queries, close support tickets, and manage inventory without waiting for engineers. Block's CTO said it was saving employees eight to ten hours per week.

When you run those savings across thousands of people and look at your Q4 results, the maths becomes uncomfortable. Block with 6,000 people produces roughly the same output as Block with 13,000. Revenue per employee just jumped from ~$2.2 million to ~$4.2 million overnight. Dorsey chose one clean cut over a slow, demoralising trickle of redundancies. The severance package - 20 weeks’ salary, equity through May, six months of healthcare, and a $5,000 transition stipend - is above average for the industry.

Here's the part that makes this different from every other "AI layoffs" headline you'll see this year: Block built the tooling first. They didn't wave at "AI transformation" as a post-hoc justification. They measured the productivity gains, confirmed the numbers in Q4, and then acted.

A fair counterpoint is worth noting: some observers argue that Block massively overhired during COVID and that AI is a convenient, stock-price-friendly narrative for a correction they needed to make anyway. The stock did jump 23% on the announcement, which tells you something about how markets are reading the "AI productivity" story right now, whether or not the full picture is as clean as the memo suggests.

The takeaway: This may be the first time a large public company has restructured specifically around AI productivity gains it could actually measure. If the bet pays off - and the Q4 numbers suggest it already has - every CEO with a bloated org chart and an AI roadmap is going to be watching very closely.

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 👋