Microsoft introduces Copilot in Excel

Welcome back, Superhuman. The days of aggressive hiring in Tech may be gone for good. With AI coming onto the stage, “getting fit“ is the new mantra in the industry. Shopify has set the bar: the company’s revenue is close to doubling, and headcount is down 30% compared to 2022. Shopify’s CEO says employees must prove AI can't do a better job before hiring anyone new. Others look set to follow this example.

P.S. We’re introducing a new mini section today called 📡 Dev Radar in the Trending on Socials section that covers the latest releases for developers and builders. Let me know if you’d like us to continue developer updates in the poll at the bottom of today’s email.

TODAY IN AI

1. Microsoft launches Copilot AI directly into Excel: The company is bringing AI directly into Excel sheets with a new COPILOT function that lets users perform generative tasks right in their spreadsheets. The function can summarize feedback, categorize data, and even pull information from outside sources — all by typing a simple formula. Currently available to beta users, you'll need the $30/month Microsoft 365 Copilot subscription to access it. You can watch a quick tutorial here.

2. Grammarly’s new AI agents can predict your grade: The new tools, rolling out for free to existing Grammarly users, can generate proper citations, predict reader questions, and most impressively, help students gauge whether their writing will land an A or a C. The app can do this partly by analyzing how your professor might react based on "publicly available" information about them. On the flip side, educators get access to plagiarism checkers and AI detectors to spot machine-generated text in submissions. You can see it in action here.

3. Google study claims 9 out of 10 video game developers are using AI: Nearly 90% of video game developers are now using AI agents to automate the grunt work of video game creation, according to a new Google Cloud study. The technology is helping studios tackle rising development costs and endless creation cycles by handling repetitive tasks like optimizing content and processing code, audio, and video — freeing up human talent for the creative heavy lifting.

PRESENTED BY WISPR

Your keyboard called—it’s taking a long weekend

Typing is slow. Wispr Flow lets you speak naturally and turns your voice into clean, formatted text—4x faster than typing.

It auto-edits filler words, fixes your grammar, and works everywhere you do (Mac, Windows, and iPhone).

Free your hands. Try Flow today.

FROM THE FRONTIER

AI is designing scientific experiments that leave experts stumped

LIGO’s gravitational wave detector is one of the world’s most sensitive scientific instruments. It can measure changes smaller than the width of a proton. But after spotting the first gravitational wave in 2015, physicists had wanted to push sensitivity even further to detect different types of cosmic events.

Caltech physicist Rana Adhikari decided to turn to AI. His team fed AI systems the building blocks for complex physics experiments. Unlike humans, the AI had no preconceptions about what "good" experimental design should look like, allowing it to explore configurations that would seem ridiculous to human scientists.

The results looked like "alien things" (at first). The machine learning system created such bizarre, asymmetrical designs for improving LIGO's gravitational wave detectors that researchers initially dismissed them as incomprehensible. One researcher said his students would have been rejected for proposing anything similar.

But the AI was onto something. After months of analysis, the team realized the system had rediscovered obscure decades-old theoretical principles from Russian physicists — ideas that had never been tested experimentally. The AI's counterintuitive design could make LIGO 10-15% more sensitive — improvement for an instrument that could already sense the distance to Alpha Centauri down to the width of a human hair.

Similar breakthroughs are happening across the sciences. Just last week, scientists created AI-generated antibiotics that can kill gonorrhoea and MRSA superbugs. At Nanjing University, researchers used AI to create quantum entanglement through a method no human had conceived. AI systems are also churning out new equations for dark matter and discovering symmetries in particle data. The pattern is clear: AI is thinking outside human assumptions to solve bottlenecks that have persisted for decades.

THE AI ACADEMY

How to turn your ideas into live apps

• Go to Trickle AI and sign up.

• Select Canvas Mode to start building your site.

• Enter your prompt in detail to create website/landing page.

Sample Prompt: Design a modern, cozy, aesthetic landing page for a donut café called “DOUGH with the Flow. Café”. The vibe is warm, slightly vintage, pastel tones (cream, blush, muted olive, soft browns) with donut & dessert-themed illustrations and cozy coffee photography.

• In just a few seconds, Trickle AI will create a fully designed landing page for you.

• You can click on the ‘Select to Edit‘ option and choose an element to edit it. Make the page feel exactly like your brand.

• Once you’re happy with your edits, you can publish your site directly or export the code for further development.

PRESENTED BY GALILEO

How are you evaluating your AI outputs?

Learn how the experts quickly and accurately evaluate AI using LLM judges! Enjoy 70 pages of content on:

• How to automate evaluations to score, explain, and flag quality issues

• Advanced techniques like token-level scoring, Chain-of-Thought, and pairwise comparison

• Practical frameworks and code examples for building your own LLM judges

Get the free eBook

IN THE KNOW

What’s trending on socials and headlines today

🥊 Bot Battle: Researchers created a social media network with AI bots. The bots ended up at war with each other.

🔈️ Speaks Volumes: Builder Josh Whiton built a completely local voice AI with no internet on a MacBook M1 with only 16GB RAM.

🤖 Bad Robot: Facebook is allegedly rolling out AI characters like “Russian girl,“ and some users are upset. (We checked our accounts and the feature is not showing up for us, so we cannot confirm its existence).

📡 Dev Radar: Gemini API now comes with URL Context to visit websites, ElevenLabs releases Music API, Alibaba’s Qwen-Image-Edit comes with image text editing, and Nvidia releases mini model with reasoning toggle.

OpenAI is rolling out a new affordable plan in India that costs less than $5, as the company looks to accelerate its growth in the country.

The Attorney General of Texas announced plans to investigate Meta and Character AI over 'misleading' mental health claims allegedly made by their chatbots.

PRODUCTIVITY

5 New AI Tools to Boost Your Productivity

• 💻️ Splash: Transform your data into insights with AI.

• 💁 Gitcord: Track and analyze your GitHub repos with real-time insights

• ⚡️ Scrybe: AI that turns your conversations into viral LinkedIn content.

• 📽️ Miniflow: Access all the AI models and automate your workflows.

• 📺️ Verbite: Generate publish-ready, high-ranking SEO content with AI.

^{* indicates a promoted tool, if any}

PROMPT STATION

Meeting Notes Summarizer

Prompt: You are a senior executive assistant with expertise in meeting documentation and project management.

Context: You are processing meeting notes to extract actionable insights and ensure nothing gets overlooked.

Instructions: Review the provided meeting notes and create a structured summary highlighting key decisions, action items, and next steps.

Constraints:
Focus only on concrete decisions and actions
Include responsible parties and deadlines when mentioned
Ignore off-topic discussions or small talk
Keep bullet points concise but informative

Output Format:
## Key Decisions:
- [Decision 1]
- [Decision 2]
## Action Items:
- [Task] - Owner: [Name] - Due: [Date]
- [Task] - Owner: [Name] - Due: [Date]
## Next Steps:
- [Next meeting date/follow-up required]
Reasoning: Apply hierarchical processing - scan entire content first, then categorize information by importance, finally extract actionable elements using chain-of-thought methodology.

User Input: [Paste your raw meeting notes here]

3D Model Scenery

Midjourney Prompt: 3D model scenery, an endless <type of> landscape with a <natural element>, terrain, cut-out diagram --ar 4:3 --c 12 --s 100

<type of> could be natural landscape (like mountains & forests) or cultural landscape (like urban areas & architecture).

<natural element> could be the sunset or a waterfall.

Whenever you’re ready to take the next step

• Check out our Top 125 AI Tools

• Get better at prompting with our Top 1,000 Prompts

• Learn how to use AI at work and get certified with our free course

Grow customers & revenue: Join companies like Amazon, Hubspot, and Salesforce. Showcase your product to our 1M+ readers and 2M+ followers on socials. Get in touch.

Until next time — Zain & the Superhuman AI team