Shipping design systems? One method that actually works

And more in today's news ...

Author

Charbel Zeaiter
June 12, 2025

TODAY’S HIGHLIGHTS

Well hello (again),

Hope you’ve had a great week. We’re almost half way through 2025 and things have no sign of slowing down.

Let’s dive in.

With love,
Charbel
Velvet Onion & Friends

  • Product: 100xing the quality of your market research

  • Founding: Your ambition and your funding

  • Design: Shipping design systems? One method that actually works

  • Science & Tech: A big step toward HIV treatment

  • AI: More and more people are confiding in with AI

  • For Young Founders: [Insert a hint to the cards]

  • Quote for the day: - Some truth from Jung

100xing the quality of your market research

For decades, companies spent billions on clunky, slow surveys and pricey panels, only to end up with insights that were outdated by the time decisions were made.

Now, AI-native startups are burying the old school methods.

Instead of running surveys, they’re simulating entire societies of digital agents who act, think, and evolve like real consumers.

Scrolling fake TikToks, reviewing imaginary skincare, even gossiping in AI-generated group chats.

These generative agents, powered by persistent memory and multimodal learning, allow brands to observe, test, and adapt in real time.

AI tools may not always be perfect, but when they deliver 70% of the accuracy in a fraction of the time and cost, that’s reasonable. Mostly.

MORE PRODUCT NEWS

Your ambition and your funding

A growing software startup with 1,000+ customers and solid, if not rocket speed, growth is staring down the barrel of its next funding round.

The catch is it’s thriving, but not quite at the $500M ARR. Not at 30%+ growth league needed to woo public markets or big-name VCs.

Growth equity firms might bite, but only for modest returns.

The founder's got grit and grand plans, but the company’s current trajectory may not match that scale.

Why a big deal?

Many founders face this quiet cliff.

Do you stay lean, stretch the roadmap, or cash in and reboot?

In any case, abide by this: In funding, “from where” matters just as much as “how much”

MORE STARTUP & FOUNDING NEWS

Shipping design systems? One method that actually works

Three attempts, one real winner.

First, they tried a full redesign. Looked fab in Figma, crumbled in code. Then came the bit-by-bit approach, which stalled mid-upgrade.

Finally, the one that works: redesign one feature, build the system alongside it, and scale from there.

Turns out, design systems flourish when rolled out with surgical precision and not creative chaos.

Why it matters:

Fancy mockups mean zilch if they never ship.

We often confuse implementation with aesthetics. When most of it is architecture.

A bland-but-working system might gather more love than a beautiful one stuck in limbo every single time.

MORE DESIGN NEWS

A big step toward HIV treatment  

Scientists in Melbourne have pulled off a stunner. They’ve coaxed HIV out of hiding inside immune cells; something long deemed near impossible.

The tech:

A next-gen fat bubble (LNP X) that sneaks mRNA into white blood cells, forcing them to expose the lurking virus.

This could eventually allow the body (or future therapies) to flush out HIV entirely.

The method builds on tech behind COVID mRNA jabs, but this time it’s gunning for a cure.

Results in lab dishes were jaw dropping. Now come animal and human trials.

Why a big deal?

HIV’s biggest trick has always been vanishing into cells untouched by drugs or immunity. Cracking that disguise? It’s a seismic shift.

Finally, we may be within striking distance of a cure.

MORE SCIENCE & TECH NEWS

More and more people are confiding in with AI

People are starting to chat with ChatGPT like it’s their best mate.

Thanking it (ah, most of us did do this at some point, didn’t we? XD), confiding in it, even calling it “alive.”

OpenAI’s got its eye on this emotional tango, focusing on perceived consciousness. Means how much the AI seems alive, rather than actually claiming it is.

They’re walking a fine line: making ChatGPT warm and friendly without turning it into some digital melodrama of feelings and a secret agenda.

Too much humanising, and folks might start expecting it to do the emotional heavy lifting, leaving real human bonds out in the cold.

Something feels wrong?

Valid because we’re talking about friendships with… technically a lifeless pile of code.

This strongly signals towards the need for more gathering, talking and empathising within ourselves.

I still can’t help but wonder, “Aren’t there enough humans for humans?”

MORE ARTIFICIAL INTELLIGENCE NEWS

CARD #2: Understanding Who You’re Building For

What's this?

No business can serve everyone. The fastest way to fail is to build for “everyone” because different groups have different needs, frustrations, and expectations. To succeed, you need to focus on a specific audience and deeply understand:

  • Who they are (age, profession, habits, interests)

  • What they struggle with (pain points, daily frustrations)

  • What they’re willing to pay for (what is urgent and valuable to them?)

Instead of trying to appeal to millions of people, start by serving a small, well-defined audience really well. 

As the saying goes, “The Riches are in the Niches.”

Why do this?

  • If you try to build for everyone, you’ll serve no one. The more targeted your audience, the stronger your marketing, messaging, and product.

  • Clear audience = clear marketing. Knowing your customers means you’ll understand where to find them, what messaging works, and what features they actually need.

  • Understanding your customers early saves time and money. Instead of wasting months guessing, defining your audience early helps you reach product-market fit faster.

  • It’s easier to scale when you dominate a niche first. Most startups grow by winning a small market first, then expanding later.

Example: Facebook started just for Harvard students before expanding to other colleges, then the world.

Real-world examples:

✅ Superhuman – Instead of marketing to everyone, this $30/month premium email app focused on busy executives, investors, and startup founders who valued speed and were willing to pay for a faster email experience.

✅ Twitch (originally Justin.tv) – It started as a general live-streaming platform but realised that gamers were the most engaged users. By pivoting to focus exclusively on gaming, Twitch became a billion-dollar company.

✅ Nike – Initially targeted elite athletes before expanding to the mainstream. Even today, their marketing still speaks directly to high performers.

✅ Tesla – Instead of launching an affordable car for everyone, they started with luxury EVs for early adopters before scaling down to mass-market models.

✅ Lululemon – Built for a specific audience (yoga practitioners) before expanding into broader fitness wear.

Action: Write a one-sentence problem statement that describes:

  • Who has the problem? (Target audience)

  • What the problem is? (Pain point)

  • Why existing solutions don’t work? (Gap in the market)

Example:
"Freelancers struggle to manage their invoices and get paid on time, and existing accounting tools are too complex."

Bonus Step:

  • Find 5 people in your target audience and ask:

    • "What’s your biggest frustration with [your problem area]?"

    • "What tools or solutions have you tried? What’s missing?"

Document their responses - this is real-world insight you can use to refine your business.

FROM THE YOUNG FOUNDERS KIT (COMING SOON) →

I HOPE YOU ENJOYED TODAY’S NEWSLETTER

“You are what you do, not what you say you’ll do.” - Carl Jung

Until next time,

With love,
Charbel

From Velvet Onion & Friends,
The House of Better