Today’s Highlights

• AI: Thinking Machines Lab: Stealthily Erected by Mira Murati, OpenAI’s Ex-CTO

• Design: How to “Show, Not Tell”

• Science & Tech: AI Designs Enzyme to Break Down Plastics

• Founding: Your Not-To-Do List: The Ultimate Productivity Hack

• Product: 2025 Tech Trends: Adapt or Fade

• Today’s AI image: An Enzyme to Digest All That Plastic

• Quote for the day: Crazy Thinkers Grant us A Unique Perspective

AI

Thinking Machines Lab: Stealthily Erected by Mira Murati, OpenAI’s Ex-CTO

Former OpenAI CTO Mira Murati has stepped into the spotlight with Thinking Machines Lab, an AI research startup aiming to make artificial intelligence more transparent, adaptable, and versatile.

With a strong focus on open science, the company is set to challenge the current AI landscape.

All You Need To Know

• Thinking Machines is developing cutting-edge models with a heavy focus on science, programming, and human-AI collaboration.

• Murati has brought in OpenAI veterans John Schulman and Barret Zoph, plus top talent from DeepMind, Character AI, and Mistral.

• Unlike many AI labs, this one promises to share research, datasets, code, and model details regularly.

• Expect AI that isn’t just good at text or images, but a seamless blend of multiple forms of input and output.

• The lab’s reveal comes just six months after she suddenly left OpenAI, citing the need for time and space to explore new ideas.

Why is this a big deal?

Murati’s move follows a growing trend of ex-OpenAI leaders launching their own AI ventures, including Ilya Sutskever’s SSI, which is reportedly raising over $1 billion.

While Thinking Machines’ all-star lineup makes it a serious contender, its commitment to open science could be the real industry game-changer, potentially nudging AI development towards a more transparent and collaborative future.

Read More

Also in AI

• OpenAI’s SWE-Lancer Tests AI vs. Devs: $400K “Earned,” Still Needs a Job

• Fiverr’s AI Hub Automates Freelancers: Innovation or Self-Destruction?

• OpenAI Poll Picks o3-mini as the Next Open-Source Model

• Musk Unveils xAI’s Gaming Studio: AI Now Makes Games Too

• xAI’s Grok Gets a Voice: New Mode Launching in a Week

• HP Acquires Humane’s AI, Scraps AI Pin, Focuses on AI Devices

• Meta Announces Llamacon: First GenAI Dev Event on April 29

Design

How to “Show, Not Tell”

Some products don’t just succeed. They captivate.

DeepSeek, a rising AI challenger, Nike Air’s legendary sneakers, and ChatGPT’s conversational magic all have one thing in common:

They made their innovations visible, intuitive, and irresistible.

Because logic alone doesn’t sell.

◆ DeepSeek’s Thought Process Display

Instead of just presenting an answer, DeepSeek shows its reasoning step by step, making the AI’s decision-making process feel more transparent and intelligent.

◆ Nike Air’s Visible Cushioning

Nike’s Air technology was an engineering leap, but no one noticed until the Air Max 1 made it visible.

Inspired by the Centre Pompidou’s exposed structure, Tinker Hatfield cut a window into the sole, letting people see the cushioning.

◆ ChatGPT’s Animated Typing Effect

Many AI models could chat, but ChatGPT felt different because it mimicked real conversation.

The animated text effect made interactions feel dynamic and human-like, transforming AI from a robotic responder to something more engaging and natural.

Innovations don’t sell themselves.

They need a "magic moment" that makes their impact instantly clear.

Read More

Also in Design

• Adobe’s AI upgrades bring smarter edits, distraction removal, and duplicate detection

• YouTube tweaks its red logo, adds a gradient, and sparks mild indifference

• HP buys Humane for $116M, shuts down AI Pin, and offers limited refunds

• AI is reshaping design, forcing creatives to adapt or fall behind

• UX is more than interfaces. Branding, pricing, and support shape the experience

• OpenAI simplifies its UI while Google’s design remains chaotic

Science & Tech

AI Designs Enzyme to Break Down Plastics

AI has successfully designed an enzyme capable of breaking ester bonds—found in both biological molecules and plastics.

This breakthrough paves the way for more efficient plastic recycling and industrial applications.

Key Developments

• Targeting Ester Bonds – Essential in plastics like polyester, breaking these bonds is crucial for recycling.

• AI-Driven Design – RFDiffusion generated enzyme structures based on natural ester-breaking proteins.

• Fluorescence Test – Only 2 out of 129 designs showed activity, proving enzyme design is tricky.

• PLACER Refinement – A second AI helped improve enzyme structure, enabling multi-step reactions.

• Final Success – After several refinement cycles, researchers created an enzyme that worked repeatedly.

• Plastic Breakdown – The same method was applied to PET plastic, showing real-world potential.

Why is this a big deal?

This marks a shift from modifying existing enzymes to creating entirely new ones.

AI-designed enzymes could revolutionise waste management, carbon capture, and industrial processes, with evolution potentially improving AI’s work even further.

Read More

Also in Science & Tech

• Scientists create a nanomaterial as strong as steel and light as foam, ready for real-world use

• Bioengineers develop a waterproof, bacteria-proof glue inspired by mussels and human proteins

• Researchers uncover how Staph bacteria steal iron, opening new doors for infection treatments

• The world’s oldest wild bird is still hatching chicks and defying the odds

Founding

Your Not-To-Do List: The Ultimate Productivity Hack

Businesses often chase expansion. New features, markets, or products; assuming growth comes from adding.

Yet, most features go unused, creating unnecessary complexity and waste. The Pareto principle (80/20 rule) suggests that only a fraction of what’s built delivers real value.

The Danger of Excess

Unused features don’t just sit idly—they increase technical debt, slow development, and complicate maintenance.

Worse, some additions cause negative value, unintentionally harming users or creating unforeseen risks.

Robinhood’s tragic case highlights the ethical responsibility of product decisions.

The Power of Doing Less

Reducing complexity lowers risk and enhances focus. Instead of chasing unproven ideas, refining existing strengths yields predictable results.

High work-in-progress spreads teams too thin, leading to poor customer experiences, staff burnout, and declining business performance.

The ‘Not Doing’ Mindset

Success is mostly about cutting what doesn’t serve the core mission.

Prioritising what matters over relentless expansion creates clarity, efficiency, and sustainable growth.

Read more

Also in Founding

• Hallmark’s Tagline Power – A slogan so strong it made buying cards a moral duty

• Winning Talent with “Why Us” – A three-step pitch to make top hires say yes

• Speeding Up Software Releases – If Amazon and Netflix can ship fast, so can you

• News Without Overload – Filter information based on what you need, not what it pushes

Product

2025 Tech Trends: Adapt or Fade

1. Tech Economy on the Rise
   After a tough period, the tech sector is rebounding. Stabilising interest rates, active mergers & acquisitions, and potential IPOs point to a strong recovery in 2025.

2. AI Moves to Mainstream
   AI is no longer experimental. It’s now integral to business operations, driving innovation across sectors like customer service and data management.

3. Big Goals, Small Budgets
   With high targets and tight budgets, CMOs must optimise teams for efficiency, not just hire more staff, making resourcefulness a key focus for 2025.

4. SEO Declines with AI's Rise
   AI-driven search is disrupting traditional SEO. Organic search is losing ground to instant AI responses, prompting B2B marketers to rethink engagement strategies.

5. Leaders as Influencers
   CEOs with strong social media presences are gaining an edge, using personal influence on platforms like LinkedIn and X to connect with audiences in ways branded accounts can't.

6. Video Takes Over
   Short-form video is the future. Platforms like TikTok and Instagram Reels demand companies create engaging content, with AI tools streamlining the process.

7. Newsletters Make a Return
   Despite social media hype, newsletters are making a comeback. By building engaged email lists, companies can bypass fluctuating social media algorithms.

In 2025, AI, video, and personal influence will dominate.

Companies that balance innovation with strong strategies and team building will lead.

Read more

Today’s AI Image

An Enzyme to Digest All That Plastic

Quote of the Day

Crazy Thinkers Grant us A Unique Perspective

"Blessed are the weird people: poets, misfits, writers, mystics, painters, and troubadours, for they teach us to see the world through different eyes."

Jacob Nordby

What we’re working on

🧞Your wish is my command