The biggest opportunity for companies right now lies in data.

But before signing a data deal, ask who holds the raw files, and for how long.

Companies are starting to license their Slack, tickets and CRM to train AI models. Every support ticket has a customer in it. Every eam in it.

It's possible to use 15 demographic details to correctly re-identify 99.98% of Americans in a dataset. Blacking out names also strips what makes the data useful: who handed what to whom, and how the work moved through the team.

Spirit Airlines' data sale is stuck on exactly this. The privacy ombudsman cleared it on customer data. The flight attendants' union objected that employee records weren't covered.

The questions you should ask before talking price:

1/ Where are the raw files held, and when are they deleted?

2/ Who does the de-identification, and how has the method been tested?

3/ How are indirect details in free text handled (health, location, family)?

4/ Does the method keep the workflow intact, or just black out names?

5/ Who keeps ownership of the data?

6/ What can it be used for, and by whom?

And check your customer contracts first. Slack's API terms have barred third-party apps from using Slack data to train LLMs since May 2025.

This is what the likes of @micro1_ai built for. Privacy is built in from the start through its own PII de-identification model, which replaces each real person with a consistent synthetic identity across every file, so the record of how work moved stays intact. Personal information is de-identified, data is handled in isolated pipelines, and the company keeps ownership of its data.

Check out what your data is worth below!

@Willob @etnshow You are correct, and so is our intern

Please see the chart he created with the correct KPI

You are king, sorry @destraynor

Our intern showed great initiative and cooked up another version of the only benchmark that matters

The Time per Pint benchmark

Our intern has rerun the numbers...

@willlob is the new king of the time per pint benchmark shaving 42% off the last one, sorry @destraynor
M

@soleio @etnshow @elonmusk @collision @patrickc @destraynor Your wish is our intern's command

He has delivered

@destraynor is still king of time per pint per kilo

@etnshow @elonmusk @collision @patrickc I think you're missing the point young sirs.

We had our intern go and make the real chart of what really matters

Tis not how many pints one drinks but how fast one drinks them

@destraynor nabs it on this one with 1 pint per 43 minutes, sorry @elonmusk

2/ Stripe famously pitched to potential customers that they could do payments in "seven lines of code", rather than dozens and API calls.

1/ Most payment companies hired sales teams to pitch executives.

The Collisons were developers building for developers, and shared what they genuinely cared about beyond just payments.

In 2010, Patrick and John Collison started Stripe when online payments required weeks of integration hell.

By 2025, Stripe hit $91.5B valuation, processing $1.4T annually.

Here's how two Irish brothers won payments through founder-led content that never felt like marketing:

11/ After fundraising, most startups spend on R&D.

Instead, Cluely spent on content and marketing:
•$140K launch video
•$500K on a Ferrari
•$600K on Times Square billboard
• $1M on merch

The goal? Stay in the public spotlight with crazy viral content.

7/ Roy doubled down. He rebranded the tool as Cluely.

His pitch? “Cheat on Everything”.

From job interviews, sales calls, exams and meetings, AI sits invisibly on your screen giving you real-time answers.

5/ Roy’s first insight: AI will be seamlessly integrated into every user interface.

His AI tool could be used for more than just “cheating” in interviews – it could be used for sales calls, meetings, anything.

Imagine an invisible assistant feeding you answers. In real-time.

4/ After an Amazon executive emailed Columbia, Roy was expelled and blacklisted from Big Tech.

Roy posted the whole saga online. It immediately took Silicon Valley by storm.

Everyone was talking about him.

Tech journalists called him unethical.

But Roy had 2 key insights:

3/ Instead of joining Big Tech, Roy wanted to go viral to launch his company.

He posted the interviews on YouTube and got millions of views overnight.

He started selling Interview Coder for $20/month.

Within weeks, it reached $1m ARR.

But it was extremely risky…

2/ They built “Interview Coder” to cheat in coding interviews.

An AI tool that can listen, read and generate answers in real-time during interviews – undetected.

Roy used it to land internship offers from Amazon, Meta and TikTok.

Then he turned them all down. Why?

1/ Two sophomores at Columbia – Roy Lee and Neel Shanmugam – were sick of Big Tech’s use of LeetCode in job interviews.

They were coding tests completely irrelevant to the real-world.

Super stressful. A waste of time.

In early 2025, they decided to break the system…

At 21, Roy Lee deliberately got himself expelled from Colombia University.

His goal? Go viral and start a media firestorm.

3 months later, his startup @cluely was worth $120M.

Here’s how a college dropout turned media controversy into a growth hack:

9/ Here’s the playbook:

1. Create deep, valuable content (podcast/long-form)
2. Reshare it across platforms (X, LinkedIn, TikTok)
3. Engage consistently to build relationships
4. Share strong perspectives to establish thought leadership
5. Convert visibility → access → deals → more content

6/ By the time he launched 20VC in 2020, Harry had done hundreds of episodes and posted thousands of times on LinkedIn and X.

Founders already trusted him. LPs knew his name. Co-investors wanted to work with him.

This gave him the credibility to raise a $400M fund in Oct 2024.

4/ On X specifically:

• Extract best quotes from episodes
• Tag guests (keep relationships warm)
• Reply to founders/VCs consistently
• Share contrarian takes on market trends

And Harry would relentlessly cold DM new followers to build a relationship with them.

The result: lowers across 3 X accounts.

3/ Harry was one of the first people to focus on building a scalable content machine:

• Podcast = depth & credibility
• X (Twitter) = daily insights, hot takes, guest amplification
• LinkedIn = longer reflections, deal announcements, portfolio wins
• YouTube/TikTok = 15M viewclips

Each platform fed the others. Every episode became 20+ pieces of content.

1/ Most VCs build a track record, then start creating content.

Harry did it backwards.

He built an audience first - across podcast, Twitter, and LinkedIn.

In 2015, @HarryStebbings started a podcast in his dorm room at 17.

No connections. No credentials. Just a microphone and cold emails.

In 2024, he raised a $400M VC fund.

Here’s how he turned content into one of the most powerful fundraising engines in European venture:

Kleiner Perkins sold their stake back to Sahil Lavingia for $1.

After raising a $7M Series A, Gumroad couldn’t close a Series B.

Despite growing steadily (~60–80% annually), VCs didn’t see hypergrowth. Sahil calls this the real VC trap: being “good but not great.”

Instead of n, he downsized, ran lean, and made Gumroad profitable. A few years later, Kleiner reached out:

“We’d be interested in selling our stake back to you for one dollar.”

That was the entire email.

Why would a top-tier VC do that?

Today, Gumroad runs like a mini public company: profitable, lean, distributing dividends, and running share buybacks via open auctions.

Early investors who got in at $0.60 have sold some of their shares back at $4+.

There’s real math behind this “third path” between unicorn success and total failure. It demands strong unit economics and ruthless operational discipline but it works. Most founders don’t even consider it.

We’re taught to think binary: unicorn or bust. But the middle path (profitable, sustainable, customer-first) can often win in the long run.

Yann LeCun is the godfather of AI. Yet, not many people know about him.

He is Meta's Chief AI Scientist. And he's building something that challenges the current approach taken by OpenAI, Anthropic, and Google.

His roadmap reveals why Meta's long-term AI strategy might create
Here's what he's working on 🧵

Now is the best time to build.

Either for yourself, or as part of a founding team at a startup.

Credits;
youtube.com/watch?v=L9KvV_…

13/ The deeper lesson:

Maor didn't try to build the biggest thing ever.

"Base44, for the first time in my life, was not trying to build the biggest thing ever."

He just wanted to solve a real problem and have fun coding.

The best outcomes come from following genuine curiosity.

12/ The playbook:

-> Build for real problems - His girlfriend's website pain became a $80M solution
-> Solo can scale with AI - "I haven't written a single line of HTML or JavaScript in the past three months"
-> Velocity beats funding - Ship fast, create momentum
-> Build in public honess, not fake wins
-> Stay profitable - Optionality beats desperation in negotiations

11/ Next morning? They signed anyway.

$80M deal closed during wartime.

Plus earnouts - if Base44 hits targets, Maor makes much more.

He's not cashing out. He's just getting started with Wix's resources.

10/ The final plot twist:

Thursday night, 2 AM. Deal papers ready to sign.

Lawyers want to make final changes. "Let's do this tomorrow morning."

4 AM: "A war broke out between Iran and Israel."

Maor: "This is karma... somebody probably hacked [Base44] because what are the chances?"

9/ Then came the acquisition talks.

Wix CEO Avishai: "Everyone's been saying that we should buy you."

Maor's advantage was that he didn't need the deal.

Base44 was profitable and growing.

"The best position to negotiate... is to be very fine with the other path."

8/ The competition was brutal.

Lovable, Bolt, Replit - all VC-funded with millions.

Maor's marketing budget: $2,000 on influencer posts (failed), $2,000 on ads (failed).

Everything else was organic growth through building in public.

7/ His secret weapon was velocity.

New features constantly.

"People like to see charts and numbers... If your product evolves really fast... people get attached to that."

Users couldn't ignore the momentum.

6/ Every day, Maor shared his journey.

Not fake "we're crushing it" posts.

The real story, raw numbers. Charts, the good, bad, and ugly.

"I was trying to be as honest as I could and just write the good, the bad."

People became addicted to his updates.

5/ The growth was insane:

-> 3 weeks to hit $1M ARR
-> 400K users in 6 months
-> $200K profit in May alone
-> Close to zero marketing budget

He did it by building in public on LinkedIn.

4/ The first users, against common advice - included his girlfriend and friends from the Scouts organization he volunteered with.

"Just begging people close to me that I feel like have really good connections."

He'd sit with them. Watch them use it. Fix bugs in real-time.

3/ So he started Base44 - an AI app builder.

Not the first in the space - But Maor had a different angle:

"Batteries included" - database, user management, analytics built-in.

No third-party integrations needed.

2/ Let's set the scene;

Maor had just spent a year in military reserves.

When he returned, he wanted to get back to what he loved most: building.

His girlfriend needed a simple website for her art business.

The existing tools were painful to use.

He knew AI models could do this better.

1/ Meet Maor Shlomo.

6 months ago, he hadn't hadn't created a line of code for Base44.

Last month, he sold Base44 to Wix for $80 million.

And... he did it all by himself;

No co-founder. No VC money. Just pure hustle + grit

This guy built and sold a company for $80M in just 6 months.

Solo founder. Zero funding. Whilst war raged around him.

The Maor Shlomo story you haven't heard 👇

These 6 books are all you need this summer:

𝟭/ 𝗗𝗲𝘀𝗶𝗴𝗻𝗶𝗻𝗴 𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗦𝘆𝘀𝘁𝗲𝗺𝘀
One of the best books on designing ML systems.
It's beginner-friendly.
And it provides a strong intuition into what an end-to-end ML system looks like.

𝟮/ 𝗟𝗟𝗠𝘀 𝗯𝘆 John Berryman 𝗮𝗻𝗱 Albert Ziegler
If you're obsessed with seeing the big picture and thinking in systems, you'll love this book.
It will show you how to engineer your prompts effectively.
You will also learn how to manipulate your prompt, make it flexible, scalable, and squeeze the most out of your models.

𝟯/ 𝗔𝗜 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 𝗯𝘆 Chip Huyen
This book has one of the best (and clearest) explanations on:
• AI vs. ML Engineering
• RAG (from classic text search to more complex, semantic-based systems)
• Building agentic systems
• LLMOps (Especially observability and user feedback)

𝟰/ 𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗟𝗟𝗠𝘀 𝗳𝗼𝗿 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗯𝘆 Louis-François Bouchard 𝗮𝗻𝗱 Louie Peters
This book will walk you through how to implement all the core algorithms required to build LLM applications.
You'll use LangChain and LlamaIndex to implement all kinds of RAG techniques (starting from the vanilla architecture).
It's great for people who are mid-to-entry-level engineers.

5/ 𝗟𝗟𝗠𝘀 𝗶𝗻 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗯𝘆 Chris Brousseau 𝗮𝗻𝗱 Matthew Sharp
• How to optimize LLMs using different tools
• The data engineering involved with building an LLM app
• Training and serving LLMs at scale
• How to think about infrastructure when serving LLMs

6/ 𝗟𝗟𝗠 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿'𝘀 𝗛𝗮𝗻𝗱𝗯𝗼𝗼𝗸 𝗯𝘆 Paul Iusztin and Maxime Labonne
This book merges my obsession with thinking about end-to-end systems with Maxime's love for fine-tuning.
It's a walk through across all the steps to build an LLM & RAG application.

12/ For founders watching this unfold:

Building in public creates the credibility that attracts enterprise clients.

Document your expertise, share your insights, establish authority in your domain.

The next Palantir will emerge from someone who builds trust and demonstrates ng before they pitch their first enterprise contract.

10/ What makes this sustainable?

AIP's "bootcamp" model onboards clients rapidly.

Companies can deploy AI solutions in days, not months.

The platform creates network effects - more data makes the AI smarter, which attracts more enterprise clients.

Competitive moats are deepening.

8/ Recent contracts show the scope:

$795 million Department of Defense deal for Maven Smart System.

Major ICE contract for deportation systems. Over $113 million in federal contracts secured.

Government remains the foundation, but commercial is the growth engine.

6/ Palantir raised full-year revenue guidance to $3.9 billion.

That represents 36% annual growth. Analysts expect full-year profits to

surge 362% year-over-year. The company's market cap has swelled to approximately $350 billion.

These are startup metrics applied to a publicly traded company...

5/ Is the valuation getting extreme?

The price-to-earnings ratio has exceeded 400, creating a "substantial disconnect" between price and fundamentals.

Despite the momentum, Wall Street consensus rating remains cautious.

Growth this fast always raises questions. Here's an ast year:

3/ Here's where it gets more interesting:

Palantir was co-founded in 2003 by Peter Thiel with backing from In-Q-Tel - the CIA's venture capital arm.

The original vision: adapt PayPal's fraud detection for national security.
Twenty years later, they're selling that same core

2/ The catalyst?

Their Artificial Intelligence Platform (AIP).

AIP integrates autonomous AI agents into their core Gotham and Foundry platforms. It's moved from concept to what CEO Alex Karp calls "hyper-execution" phase.

Dresner Advisory Services ranked them #1 for AI and machine learning.

1/ In Q1 2025, Palantir reported $884 million in revenue.

That's 39% growth year-over-year. But the real story is hidden in the segments.

U.S. commercial revenue exploded 71% to $255 million, while government revenue climbed 45%.

This isn't just government contractor growth anymore.

🚨 Palantir is the top-performing stock in the S&P 500 this year.

It is up 102% in 2025.

We took a deep dive and looked into what's really happening behind those numbers.

Here's what most people don't know 👇