Today’s Highlights

• AI: Anthropic Awakens: A Smarter Claude on the Horizon

• Design: Should You Trust ChatGPT’s Design Advice?

• Science & Tech: Brain Implant to Boost Mood Enters NHS Trials

• Founding: “No Version 2” So They Say

• Product: How to Build Viral Products

• Today’s AI image: Your New Pizza Guy: Metal & Wires, Not Flesh & Bone

• Quote for the day: A Strong Belief Becomes Reality

AI

Anthropic Awakens: A Smarter Claude on the Horizon

Anthropic is gearing up to launch a new AI model in the coming weeks, blending traditional language skills with cutting-edge reasoning.

The upgrade promises stronger coding abilities and greater flexibility for developers, allowing them to fine-tune performance to suit their needs.

What’s New?

• This model isn’t just a chatbot; it can switch between a standard LLM and a deep reasoning engine, adapting to different tasks seamlessly.

• A new sliding scale lets developers decide how much reasoning muscle to flex for each query, balancing speed and complexity.

• At full reasoning capacity, the model reportedly handles large-scale programming projects with ease, making it a powerhouse for developers.

• Whispers suggest Anthropic already has an AI stronger than OpenAI’s GPT-4o, but it’s been kept under wraps due to safety concerns.

• With rivals like OpenAI and Google pushing new releases, this could be Anthropic’s moment to reclaim the spotlight.

Why It’s a Big Deal

Anthropic has been unusually quiet since Sonnet 3.5, while the rest of the AI world raced ahead.

This launch could change the game, especially as AI moves toward hybrid models that mix reasoning and speed.

If Anthropic plays its cards right, it might just set the new industry standard.

Read More

Also in AI

• YouTube Shorts Gets AI Magic: Type, Generate, Done

• Google’s Gemini Flash 2.0 Outshines GPT-4o in AI Showdown

• Musk Teases Grok 3 Release, Claims It Outthinks the Competition

• Major Publishers Sue Cohere for AI-Powered Copycat Content

• Baidu to Make Ernie AI Free from April 1 to Spur Adoption

• Apptronik Secures $350M to Scale Apollo Robots into New Markets

• Musk Offers $97.4B for OpenAI, Sets May 10 Deadline

Design

Should You Trust ChatGPT’s Design Advice?

When designer Anthony Hobday asked ChatGPT to recommend experts in interface design, it confidently named Adam Silver—a spot-on choice.

But when it comes to actual design advice, ChatGPT’s responses often lack depth, context, and practical wisdom.

Where ChatGPT’s Advice Falls Short

• Ignores Users' Needs – Suggests tooltips for hints, hiding helpful info behind clunky, inaccessible interactions.

• Style Over Substance – Prioritises "clean design" over clarity, missing the point of usability.

• Vague & Unhelpful – Uses wishy-washy terms like "you can" and "sometimes" instead of giving clear, tested advice.

• Inconsistent Logic – Treats radio button hints differently based on form structure when they should work the same way.

• No Rationale – Gives advice without explaining why, making it hard to learn or apply to future designs.

How to Get Better at Spotting Bad Advice

• Learn core design principles, not just patterns.

• Follow research-backed design practices rather than surface-level trends.

• Test solutions in real-world scenarios instead of blindly following AI-generated suggestions.

Read More

Also in Design

• Framer Awards 2024: Digital Design’s Best Take the Spotlight

• Adobe’s AI Video Tool Hits Beta—Sora Has Competition

• Phase Snags $13M to Challenge Figma with No-Code Magic

• UX Pros Now Need to Do It All—Thanks, AI!

• 8 AI Design Breakthroughs Changing How We Create

• UX Innovation Tip: Think Beyond Screens & Out of the Digital Box

Science & Tech

Brain Implant to Boost Mood Enters NHS Trials 

A cutting-edge NHS trial is set to explore whether an implantable brain device can enhance mood and treat neurological conditions using targeted ultrasound.

This £6.5 million study, backed by the UK’s Advanced Research and Invention Agency (Aria), will assess the safety and potential of this non-invasive technology.

How It Works

• Implanted under the skull, the device maps brain activity and stimulates neurons with ultrasound.

• Unlike electrode implants, it monitors and adjusts brain function using sound waves.

• The trial involves 30 patients, focusing on those with skull defects to avoid new surgeries.

• It provides brain imaging 100 times sharper than standard fMRI scans.

• If effective, it could help treat depression, addiction, OCD, and epilepsy.

• Patients will wear it for two hours while researchers assess mood and motivation changes.

Why It’s a Big Deal

Neurotechnology has been advancing rapidly, with companies like Elon Musk’s Neuralink already testing brain implants for paralysis and stroke recovery.

However, this new approach offers a less invasive alternative, expanding the potential for treating a wider range of conditions.

Despite the promise, ethical concerns loom.

Privacy, data ownership, and misuse risks need scrutiny.

Researchers also worry about unintended personality or impulse changes.

With the first phase of regulatory approvals set to last eight months, this study marks the beginning of what could be a game-changing shift in how we treat neurological and mental health conditions.

Read More

Also in Science & Tech

• Uber Eats Rolls Out Delivery Bots: Robot Couriers Hit the Streets

• AI-Designed Enzymes Could Revolutionise Biotech: If We Solve the Data Puzzle

• Musk’s Boring Company Digs Deep in Dubai with a 10-Mile Tunnel Project

• AI-Powered Prosthetics Now Move and Feel More Like the Real Thing

• Evolution’s Unexpected Twist: Complex Brains Evolved More Than Once

Founding

“No Version 2” - apparently.

“There is no version 2.”

The reality? That’s nonsense. The real issue isn’t that version 2 never happens. It’s that version 1 often doesn’t deserve one.

Why This Myth Exists

Most projects get launched, then abandoned. Priorities shift, roadmaps overflow, and version 2 gets lost in the chaos.

But look at any successful product—Facebook is on version 495, Tesla’s self-driving software is at 12, and Substack updates constantly. Clearly, version 2 is real.

Your First Attempt Was Weak

If a feature solves a real problem, customers will demand improvements.

If version 2 never arrives, it’s not because no one had time. It’s because version 1 wasn’t compelling enough.

Build a First Version That Matters

• Solve a genuine customer pain point.

• Strip out the ‘nice-to-haves’—those can wait.

• Gather feedback fast.

• Ensure version 1 is strong enough to spark demand for version 2.

Make Version 1 Worth Upgrading

Version 2 exists for products that deserve it. If no one’s asking for the next iteration, your first attempt wasn’t good enough. Solve real problems, and the next version will write itself.

Read more

Also in Founding

• Stablecoins: The Future of Open Finance?
  Beyond crypto trading, they could power global liquidity and banking at scale

• Venture Capital’s Harsh Wake-Up Call
  Overhyped valuations meet reality—reset or regret

• Why Startups Fail (And How to Survive)
  No market, bad leadership, or cash woes? Avoid the usual pitfalls

• The Superhuman Paradox: Slow Growth Wins
  Scaling too fast breaks companies. Pace yourself

Product

How to Build Viral Products

1. Product Philosophy: Hook Fast or Lose

Users decide in three seconds. Fail that, and they’re gone.

Teens (13–18) drive organic growth, but engagement drops 20% per year with age. Don’t just claim value—bake social proof into the experience.

2. Testing: More Shots, Fewer Guesses

If a product works, you’ll know.

Test in stages: do users engage, share within their circles, and spread it further? Remove distractions, focus on strong signals, and go deep in small groups before scaling.

3. User Acquisition: Frequent Problems Win

Solve daily or weekly issues—rare-use products struggle.

Find where users battle inefficiencies. Adults need ads; organic adoption is tougher. People download apps to make money, find love, or escape reality.

4. User Experience: Friction Kills, Value Wins

Every tap must feel worthwhile. Live chat boosts support, and social features should add—not distract—from core value.

Prioritise activation before scaling. Reduce every obstacle to action.

5. Growth & Retention: Build, Test, Repeat

Launch small, iterate fast—two months max for an MVP. Start with tight-knit adoption, then expand.

Retention matters as much as growth. Prioritise impact, cut distractions, and start now—don’t wait for funding.

Viral products don’t HAPPEN. They’re engineered. Sharp execution, constant testing, and user obsession drive real growth.

Read more

Today’s AI Image

Your New Pizza Guy: Metal & Wires, Not Flesh & Bone

Quote of the Day

A Strong Belief Becomes Reality

"Whether you think you can or you think you can’t, you’re right."

Henry Ford

What we’re working on

🧞Your wish is my command