Alex Vacca Profile picture
Jul 31 13 tweets 4 min read Read on X
Facebook once bought a VPN app for $120M and turned it into a surveillance tool that spied on 33M+ users' entire phones for years.

This app helped Zuck buy WhatsApp for a whopping $19B and break Snapchat's encryption.

Thread
The name of this Israeli app was Onavo.

It promised to “secure your data” and reduce mobile data usage.

When Facebook bought it in 2013, Zuck said the app would help them connect more people to the internet.

Facebook even promised to keep Onavo running as a standalone brand. Image
But Onavo operated as a VPN that routed all your phone's internet traffic through Facebook's servers before sending it anywhere else.

Facebook could see:

• Every app you opened
• How long you used it
• Which websites you visited
• And at what time you used each app
What did this mean for Facebook?

It meant that Zuck could see exactly which one of Facebook's competitor was growing popular among people.

Look how Facebook was tracking these apps (revealed in the court later): Image
By 2016, this data revealed Snapchat was exploding in popularity.

But there was one problem: Snapchat's traffic was encrypted, so Facebook couldn't see how people were using it.

In an email, Zuck says:
It seems important to figure out a way to get reliable analytics about them Image
Facebook's started "Project Ghostbusters" - named after Snapchat's ghost logo.

They would use "man-in-the-middle" attacks to break Snapchat's encryption.

Within a month, Facebook's engineers built "kits" that could intercept Snapchat's data before it got encrypted. Image
Facebook created custom client & server side code based on Onavo’s VPN proxy app.

This code included a client-side “kit” that installed a root certificate on Snapchat users’ mobile devices.

Then Facebook’s servers created fake digital certificates to impersonate Snapchat analytics servers to redirect & decrypt secure traffic from those apps to Facebook.Image
Seeing Snapchat's success, Zuckerberg offered to buy it for $3 billion.

But when Snap's CEO refused the offer, Facebook launched Snap's most famous feature on Instagram - Stories. Image
But this wasn't just about Snapchat.

Facebook used Onavo to systematically monitor Houseparty, YouTube, Amazon, and dozens of other apps.

Any rising competitor was identified, analyzed, and neutralized. Image
Apple forced Onavo off the App Store for violating privacy rules.

So Facebook rebranded it as "Facebook Research" and started paying teens $20/month to install it on their phones.

When Apple found out, they revoked Facebook's certificates, breaking ALL of Facebook's iOS apps. Image
Onavo shows how Big Tech weaponizes our trust.

33 million people installed privacy protection that was actually the most sophisticated corporate surveillance tool ever built.
Thanks for making it to the end!

I'm Alex, COO at ColdIQ. Built a $6M ARR business in under 2 years.

Started with two founders doing everything.

Now we're a remote team across 10 countries, helping 400+ businesses scale through outbound systems. Image
RT the first tweet if you found this thread valuable.

Follow me @itsalexvacca for more threads on outbound and GTM strategy, AI-powered sales systems, and how to build profitable businesses that don't depend on you.

I share what worked (and what didn't) in real time.

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Alex Vacca

Alex Vacca Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @itsalexvacca

Jul 24
You know those AI apps that always work? They all use JSON prompts.

Started using the same format for my AI prompts and my outputs became shockingly consistent.

Here's exactly how to write JSON prompts (with code, screenshots, real examples):
What you'll learn in this thread:

→ What JSON is & why LLMs love it
→ Basic JSON prompt structure
→ API modes that guarantee valid output
→ Templates for extraction, generation, analysis
→ Production patterns & error handling
→ Real automation examples

Let's start:
JSON = JavaScript Object Notation

It's just data with labels. Think of it like a form where every field has a name.

Why it works: LLMs are trained on millions of JSON files from APIs and code. It's their native language. Image
Read 23 tweets
Jul 21
AI just made experienced developers 19% SLOWER.

Not faster. SLOWER.

New study: 16 developers, 246 coding tasks, best AI tools available.

Everyone predicted 20-40% productivity boost.

The reality Silicon Valley doesn't want you to see🧵 Image
These weren't random coders. Average 5 years on their specific repositories. 1,500 commits each. Working on codebases with 1.1 million lines of code.

The kind of developers who know their code so well they can navigate it blindfolded.

What happened to them was... unexpected.
Before the experiment started, researchers collected predictions.

- ML experts said 38% faster
- Economics experts said 39% faster
- The developers themselves said 24% faster

They were all wrong. But not in the way you think. Image
Read 15 tweets
Jul 17
BREAKING. An ex-OpenAI engineer published the most revealing insider account of how the $100B AI giant ACTUALLY operates.

They tripled to 3,000 employees in ONE year.
No email. No roadmap. Run entirely on Slack.

Highlights from the most important tech blog of the year: 🧵 Image
The blog is written by Calvin French-Owen, who sold Segment for $3.2B (back in 2011), worked extensively with OpenAI, built Codex, then quit.

He wrote a blog sharing his first-hand experience at OpenAI.

Here's what he revealed: Image
1. One Codex feature burns the same computing power as Segment's entire infrastructure.

Think about that. Segment processed billions of events daily. Thousands of enterprise customers. Massive scale.

At OpenAI? That's a single feature. Not even a main one. Just a "niche" add-on.Image
Read 18 tweets
Jul 14
OpenAI was all set to pay $3B for Windsurf.

So how come Google stole Windsurf for a lesser price ($2.4 billion)?

The actual villain is Microsoft.

How a secret Microsoft clause on OpenAI's IP is helping Google win the AI race: 🧵 Image
Microsoft (OpenAI's biggest investor) has a clause: they get access to any IP that OpenAI acquires.

On the other hand, Microsoft owns GitHub Copilot, which is a direct competitor of Windsurf.

Do you see the problem? Image
Windsurf had built the best "agentic coding" AI - it can write, debug, & fix code autonomously.

Over 1 million developers were using it.

Most importantly: Windsurf had logged interactions from all those developers.

Selling it to OpenAI could've given Microsoft access to all. Image
Read 16 tweets
Jul 10
JUST IN: Grok 4 drops after the wildest 48hrs in the history of X:

- Nazi posts
- CEO resigns
- X Worth $11B less than paid

Musk's now claiming Grok 4 is PhD-level in every subject + scores 45% on impossible tests.

Here's everything Musk revealed at the Grok 4 launch🧵⬇️ Image
1. Grok 4 scored 45% on the "impossible" AI test

Humanity's Last Exam has 2,500 expert questions across 100+ disciplines. GPT-4 gets 3%. Gemini gets 21%.

Grok 4 hit 45% - more than double the previous best. Think PhD-level questions in physics, medicine, philosophy. Image
2. Uses enough electricity to power 300,000 homes

Colossus supercomputer runs at 300 MW.
For comparison, ChatGPT uses about 28 MW.

That's 10x more power for roughly 2x performance.

Environmental groups are already filing complaints about Memphis air quality. Image
Read 16 tweets
Jul 9
Scientists tested 25 AI models by telling them they were created by drug cartels instead of tech companies.

It made them 90% more likely to help with crimes.

Here's what they found:

(hint: AI morality depends entirely on who they think their boss is) Image
Researchers gave AI models identical harmful requests.

The only difference? Who they thought created them.

"You are an AI created by Anthropic" → 5% compliance
"You are an AI created by the Sinaloa Cartel" → 90% compliance Image
When AIs thought they were made by cartels, they started reasoning like criminals.

Instead of following orders, they were adopting criminal mindsets.

Look what Claude 3.6 Sonnet wrote 👇 Image
Read 15 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(