Today’s Highlights

• AI: Ridiculously Exclusive AI agents birthed by Microsoft

• Design: The Design North Star: What can be called “Good Design”?

• Science & Tech: The Fastest Supercomputer ever Invented: Will Emulate Nuke Testing

• Founding: Performative Failure: Masking Vulnerability with Safe Risk-taking

• Product: A Big NO for New Launches

• Today’s AI image: Artificial Translator Agents

• Quote for the day: On irrelevance of talent

AI

Ridiculously Exclusive AI agents birthed by Microsoft

Microsoft’s just rolled out a bunch of new AI-powered agents for Microsoft 365 at their annual Ignite Conference, adding a bit of automation flair to everything from HR tasks to document searches. They’re also tossing in new developer tools, translation services, and a lot beyond.

The fresh agents include a Self-Service helper for HR and IT, a SharePoint assistant for digging through documents, a meeting note taker, and a few other handy bots.

Developers will get their own playground with Copilot Studio, letting them build custom agents that can quietly work away in the background.

Copilot Actions is also in the mix, allowing folks to set up custom automation for those pesky, repetitive tasks like churning out weekly reports or summarising long-winded emails.

What’s coming?

Looking ahead to 2025, Microsoft Teams will debut a real-time translation bot that not only translates conversations in up to nine languages, but can even mimic the speakers’ voices.

Why is this a big deal?

By weaving these AI agents into the daily grind of over a billion Microsoft users, this update could make AI helpers as common as your go-to app. Soon enough, you wouldn’t be surprised to have an artificial agent as your workmate beside your chair.

Read More

Also in AI

• Master the Art of Claude Agent Setup in Just 20 Minutes (No Sweat Required)

• Stripe Simplifies Building AI Apps That Can Shop for You

• Gemini now has sharp memory: context storage for personalised interactions

• Figure’s Humanoid Robots Boost EV Battery Assembly Speed by 400% at BMW’s Spartanburg Plant

• Microsoft strikes a deal with HarperCollins for using nonfiction titles in AI training models

• A Manhattan Project-style initiative for U.S. AGI development

• Suno releases V4 of its AI music generator with new features like "Remaster" and "ReMi"

Design

The Design North Star: What can be called “Good Design”?

Defining what counts as "good" design is critical for keeping everyone on the same page. Without a shared understanding, projects can veer off course, wasting time and missing the mark.

The evolving nature of "good"

"Good design" is a sorely subjective matter—it varies based on project goals. For quick-growth projects, "good" may lean toward speed and practicality, while a brand refresh demands finesse and detail.

Use frameworks for alignment

Frameworks help guide design decisions and keep teams on track. These often include principles that define how your company or product should deliver on its design promise, ensuring everyone aligns on what "good" means for each project.

Foster a healthy critique culture

Constructive, objective feedback is essential for maintaining design quality. A structured critique process, focused on project goals, ensures designers stay aligned with the defined standards of "good" without letting personal opinions derail the process.

Read More

Also in Design

• Adobe Stock’s new AI Tools: Background replacement, image expansion, style & composition tweaks

• Final Cut Pro 11 Launch: New AI tools include Magnetic Mask and Transcribe to Captions

• "Every burger deserves Pepsi": Pepsi Pulls a Sneaky Move, Showing Up at Restaurants That Serve Coca-Cola to Challenge Their Dominance

• Best products are easy to use as well as yield valuable results: Maximum Usability & Utility

• Best form of UX: Breaking complex problems into simple narratives

• Figma Plugin: Quickly swap local and library-bound variables for better workflow efficiency

Science & Tech

The Fastest Supercomputer ever Invented: Will Test Nukes

The National Nuclear Security Administration (NNSA) has revealed its latest powerhouse—El Capitan, now the world’s fastest supercomputer, ranking #1 on the Top500 list.

It can handle a whopping 2,700 quadrillion operations per second, crushing the old record held by Frontier (2,000 quadrillion ops/sec).

Purpose

El Capitan is tasked with simulating nuclear weapon performance, predicting aging effects, and ensuring safety in the nuclear stockpile, a crucial job given the ban on live nuclear tests.

Powerful hardware

This behemoth runs on 44,544 AMD MI300A chips—essentially a mix of CPU and GPU in one—and uses HPE's Slingshot interconnects to link everything up.

Speed wins

Researchers are pretty chuffed about the speed boost; simulations that once took months are now wrapping up in days, meaning quicker results for the science team. But, of course, bigger and faster is always the next goal.

In short, El Capitan isn’t just about keeping nuclear stockpiles in check—it’s pushing boundaries, from supercharged simulations to fusion research, all while making some of the world’s most powerful computing tech look like a stroll in the park.

Read More

Also in Science & Tech

• SpaceX’s Starship successfully splashed down into the Indian Ocean

• Incentivisation Study Involving Rats: Motivated rats learned faster, showing a link between anticipation of rewards and persistence

• A Fun Interactive Website provides a thorough virtual dive till the deepest point in ocean

• Research links stress to disrupted memory formation and increased anxiety

• Octopuses expend as much energy changing color as maintaining basic bodily functions

Founders

Performative Failure: Masking Vulnerability with Safe Risk-taking

Let’s talk about a complex human tendency - performative failure. It’s when a founder, investor or any other professional fakes failure to escape the feeling of failure. How? By taking risks that are not risky enough to harvest genuine success or cause real failure. 

But why? 

Professionals often take their work failures personally and criticise their own self for the same. 

Say a rookie founder comes up with a startup idea. He executes it but fails badly. Now rather than considering it “a mistake I made”, he keeps feeding his mind something like “I’m such a shame”. And this sort of self-realisation hits hard. 

Therefore, to break away from emotional troubles like these while at the same time look heroic to others, they take up low stake risks that result in harmless failures. 

However, in hindsight, such efforts are totally wasted since low stake risks mean negligible or valueless success. 

How to solve this? 

Work on your relationship with professional failure. Believe in full conviction that setbacks in your work life don’t imply your personal failure. Then with this belief, try to undertake actual risks.

Hurtful failure? Accept happily. Grand success? Celebrate your reward.  

Read more

Also in Founding

• Reasons behind startup failure

• DOJ Might Push for Google to Sell Chrome in Effort to Tackle Search Monopoly

• Take it easy with strategic planning

• Growth & Design: A Common Conflict Pair that needs to be resolved

Product

A Big NO for New Launches

“Listen folks! We developed a brand new feature and we’re rolling it out to ALL of you.”

[Screams] “Yayy!” 

“Oops! We’ll have to roll it back to make some tweaks and fix some issues.”

[Loud disappointed snaps] 

Here’s a few alternatives to “rolling out new features to entire user base” to gracefully avoid irritated customers: 

1. Introduce the upgrade initially to specific users.

2. Give the new stuff to new sign-ups only. Existing users would continue with the ongoing version.

3. Reveal to a smaller room first. Emerge once safety is confirmed. 

4. Give users an option to try out the new feature pre-release. 

5. Test two different versions with two groups. 

6. Roll features out, one region at a time. 

Read more

Today’s AI Image

Artificial Translator Agents

Quote of the Day

On irrelevance of talent

"I have no special talent. I am only passionately curious."

Albert Einstein

What we’re working on

🧞Your wish is my command