Hey — It’s Nico.

Today’s issue takes 5 minutes to read. If you only have one, here are the five main things:

• A FT report comes out sharing the story behind Aaqua’s failure and £100mn lawsuit — read more below.

• A founder shares some lessons from building a VC-backed startup.

• Menlo Ventures and Anthropic launch a $100M AI fund.

• Fei-Fei Li builds a $1B startup in 4 months and raises $100M from A16z.

• People are using Claude 3.5 Sonnet to build 6-figure startups — learn more below.

Let’s get into it.

Presented by FundraisingOS

Get Failory’s Fundraising OS for $97 $297

90% of startups fail because they run out of cash.

Fundraising is hard and a challenge all founders face.

This is why I’ve built Fundraising OS, which has everything you need to raise funding for your startup. This includes:

• 5 lists of 3,500+ investors.

• 7 tools for tracking and talking with investors.

• 18 templates for different emails to investors.

• 3 learning resources.

And as a thank you for being a Failory subscriber, I’m giving you a 67% discount.

This Week In Startups

🔗 Resources

Lessons from building a VC-backed startup.

List of accelerators you can apply for.

12 strategic 'edges' for startups.

Network effects for AI startups.

Coatue’s views on the state of the markets. 

📰 News

AI startup Cohere cuts staff after raising $500M.

Meta won’t release Llama AI in Europe over unpredictable regulations.

Wiz walks away from Google’s $23B acquisition offer.

Indian ed-tech giant Byju faces total shutdown.

Menlo Ventures and Anthropic launch a $100M AI fund.

💸 Fundraising

'Godmother of AI' Fei-Fei Li builds a $1B startup in 4 months and raises $100M from A16z.

Legal tech firm Clio raises $900M, plans expansion into AI and fintech offerings.

Posh raises $22M to personalize events and gatherings.

17-year-old Eric Zhu’s private market data startup Aviato raises $2.3M.

Fail(St)ory

2 Years, No Product

A few weeks ago, the Financial Times published an excellent report on what happened to Aaqua, a startup that closed in 2022.

The company, which promised a revolutionary approach to social media, ultimately never delivered on its grand vision.

What Was Aaqua: Aaqua aimed to create a social media platform focused on communities built around shared passions.

Unlike other social media giants that monetize user data through ads, Aaqua pledged to use data only to suggest relevant content tailored to users' interests.

The fundamental concept of Aaqua was to develop a "circular economy" where content and interactions would be organically generated, driven by users' genuine interests instead of profit-driven algorithms.

This vision, and the high salaries they offered quickly drew top talent from the leading tech companies. In a short time, Aaqua had over 175 employees.

However, despite having the greatest talent and significant funding, Aaqua’s app was never launched.

The Numbers:

• 📅 Founded in 2020.

• 💰 Raised substantial funding from investors, including Candy Ventures.

• 🧑‍💼 At its peak, it had 175 employees.

• 🔥 Burned $1M+ every month initially, escalating to around $4M per month by the end of 2021.

• 📉 Never launched its app despite the significant investment and resources.

Reasons For Failure:

• Legal Troubles: Aaqua's founder, Robert Bonnier, made grand promises about the platform's potential, including false claims that Apple and LVMH were backing Aaqua. These claims were critical to securing investment from Candy Ventures. When Candy Ventures discovered that Apple had never intended to back Aaqua, they sued Bonnier, resulting in a freezing order that prevented Aaqua employees from being paid in July 2022.

• Poor Leadership and Transparency: Leadership issues plagued Aaqua, with Bonnier's confrontational management style contributing to internal strife. According to the FT report, Bonnier once publicly and aggressively fired a key employee, stating: “As founder and CEO of Aaqua, I hereby fire you on the spot... I have no idea what you ever did for Aaqua.”

• Financial Mismanagement: Aaqua was burning through cash at an unsustainable rate. Their large workforce meant they were spending millions per month just on salaries. This could have been acceptable if Aaqua had something to show for all this money, but in its two-year lifetime, they never released even a prototype of the app.

• Lack of a Viable Product: Despite the grand vision and substantial funding, Aaqua never developed a market-ready product. The inability to launch an app meant the company could not generate revenue or validate its business model. 

Why It Matters:

• Aaqua’s journey reflects the importance of transparency and realistic goal-setting in startups.

• It is another example of the great difficulties of building a new social media app.

• It highlights the importance of building an MVP before hiring hundreds of employees.

Go Deeper: Check Financial Times’s report.

Trend

Coding with Sonnet

A few weeks ago, I talked about Anthropic's newest model, Claude 3.5 Sonnet. This week, I noticed many people discussing how they're using Sonnet to quickly build software businesses. Surprisingly, many seem to prefer Sonnet over GPT for coding.

Why It Matters:

• Developers say Sonnet makes them 10 times more efficient.

• People without coding experience have created complex projects (see examples below).

• As AI improves, more people can build cool things without being expert programmers.

Some Examples:

• Pietro Schirano made EverArt with the help of Sonnet. The app is now making $100K per month.

• The Otto team wrote half their code using Sonnet.

• Riley Brown made a working note-taking app without coding experience. He just kept asking Claude questions until the app was finished.

• Min Choi shared a list with several other examples.

Why Sonnet: When Sonnet was released, I mentioned that benchmarks showed its coding abilities were better than Gemini, Llama, and even GPT-4.

We know benchmarks don't always reflect how people actually use the model. But in this case, it seems these tests were right. Many developers say Sonnet is much better than GPT for coding.

I'm not completely sure if this is true, but what I do find very useful are Sonnet's artifacts, which let the model run JavaScript code right in the chat. This has many uses, but one I find really interesting is this idea by Pietro Schirano.

Help Me Improve Failory

That's all of this edition.

Cheers,

Nico