Post

How to get URL link on X (Twitter) App

On the Twitter thread, click on or icon on the bottom
Click again on or Share Via icon
Click on Copy Link to Tweet
Paste it above and click "Unroll Thread"!
More info at Twitter Help

Kevin Henrikson

Feb 11 • 17 tweets • 5 min read • Read on X

Jack Dorsey just declared war on closed-source AI.

It integrates with 4 major AI models: DeepSeek, OpenAI, Google, and Anthropic.

It lets you build powerful AI tools while keeping your data completely private.

Here's how this could transform the future of AI development:

First, some context:

Block is the parent company of:
• Square (point-of-sale services)
• Cash App (mobile payments)
• Tidal (music streaming)

Dorsey teased Goose by posting "open source everything" on X.

The problem was simple:

Building AI agents is too complex for most developers.

You need expertise in machine learning, neural networks, and programming.

This keeps many brilliant developers from innovating with AI.

But Dorsey's team had a solution that would change everything:

Enter Goose:

A free, open-source framework that simplifies the process of building AI agents.

It works with any LLM as the intelligence on the backend.
The goal? To democratize AI-powered systems for developers.

Here's where it gets interesting:

Goose can integrate with various LLMs, including:

• DeepSeek
• OpenAI
• Google
• Anthropic
You're not locked into one provider.

This flexibility is revolutionary, but there's an even bigger advantage:

Unlike other AI platforms, Goose can be deployed:

• On-premises
• In virtual private clouds
• Within your own infrastructure
As Block's VP confirms: "We definitely do not have anything in the middle of Goose usage — no calls to our servers."

This changes everything:

Think about what this means:

Banks can build AI agents without exposing customer data.

Healthcare providers can innovate while maintaining patient privacy.

Government agencies can leverage AI while keeping data secure.

But the real magic is in what Goose can do:

Goose can:
• Conduct code migrations (Ember to React, Ruby to Kotlin)
• Dive into unfamiliar codebases
• Generate unit tests
• Create API scaffolding
• Remove feature flags
• Increase code coverage

And that's just the beginning of its capabilities...

Jackie Brosamer, VP of data & AI platform engineering at Block, revealed something fascinating:

Goose can work across different systems - integrating Google Drive with Slack, for example.

One team member uses it to track 40 hours of meetings and figure out what they did last week.

The platform is designed to:
• Search and navigate codebases
• Read, write, and edit files
• Install dependencies
• Run tests
• Refine outputs automatically

All while maintaining data privacy and security. But here's the most powerful part:

Goose is released under the Apache 2.0 license.

This means:
• Anyone can use it
• Everyone can audit the code
• Innovation happens in public
• The community drives development

This approach has profound implications:

As Bradley Axen, AI tech lead at Block, explains:

"We're not monetizing Goose directly — it's fully Apache licensed. Instead, we aim to develop products that work seamlessly with agents."

The strategy behind this decision is brilliant:

Goose is designed to work with Anthropic's Model Context Protocol (MCP).

This standardized suite of APIs means Goose can connect to:

• Content repositories
• Business applications
• Development environments

Making it infinitely extensible.

The implications are massive:

We're about to see an explosion of AI innovation.

Not just from big tech companies, but from developers everywhere.

This democratization of AI could reshape the entire industry.

Want to master the founder mindset and build better?

Join Founder Mode for free weekly insights on startups, systems, and personal growth: foundermode.kit.com

https://twitter.com/17429610/status/1889321013940486202

I hope you've found this thread helpful.

Follow me @KevinHenrikson for more.

Like/Repost the quote below if you can:

https://twitter.com/17429610/status/1889321013940486202

Video credits:
• YT link youtube.com/watch?v=yrL4gX…
• YT link youtube.com/watch?v=tZCNOe…

• • •

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

This Thread may be Removed Anytime!

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

More from @KevinHenrikson

Kevin Henrikson

@KevinHenrikson

Feb 6

I've spent 25 years optimizing travel systems.

After 1,000+ flights and countless hotel stays, I've built what I call my "Travel Stack."

It's a collection of tools & systems that make travel seamless.

Here's my exact setup for 2025:

First, understand this:

By 2025, international tourism will grow 3-5% from 2024 levels.

The industry is transforming rapidly, but most people approach it wrong.

They focus on destinations and deals, missing the bigger picture:

A good travel system eliminates friction and maximizes experience.

By 2025, AI will handle 70% of customer interactions in travel, up from 15% in 2018.

The global AI market in travel will reach $1.2 billion, growing at 10.2% annually.

As a commercial pilot and frequent traveler, I've tested every system imaginable.

Here's what actually works:

Read 17 tweets

Kevin Henrikson

@KevinHenrikson

Jan 30

I cracked the code to superhuman productivity:

It helped me build multiple successful companies while being home for dinner every night.

Most founders sacrifice family for success - I found a way to have both.

Here's the exact blueprint that makes it possible:

Most people think you need to grind 24/7 to be successful.

After 20+ years building and selling companies, I've learned the opposite:

Systems beat hustle. Every. Single. Time.

Let me show you exactly how...

First, understand this fundamental truth:

I tried the "hustle harder" approach and it led to complete burnout - twice.

That's when I realized: Working longer hours actually decreased my performance.

Work in focused sprints, not endless marathons.

Here's the system I built:

Read 15 tweets

Kevin Henrikson

@KevinHenrikson

Jan 28

The most dangerous advice in Silicon Valley:

"Hire good managers and get out of the way."

I've watched this nearly destroy promising startups.

Here's what the best founders understand that most CEOs never learn:

At a YC event, Brian Chesky (Airbnb CEO) gave a talk that left the audience stunned.

Having scaled Acompli from 0 to 200M+ users, his message resonated deeply with my journey.

But here's where it gets interesting:

As Airbnb grew, "experts" told Chesky to follow conventional wisdom:

"Hire good people and give them room to do their jobs."

He tried it. The results? Disastrous.

I made the exact same mistake at Microsoft. The damage took months to undo.

Read 19 tweets

Kevin Henrikson

@KevinHenrikson

Jan 20

Your product has too many features.

I know because I made this mistake before a $200 million exit.

Here's the counterintuitive framework that actually works:

(And why most founders get it wrong)

Most founders think more features = better product.

I used to believe this too.

Until I learned a painful lesson building email apps used by millions:

Every new feature you add creates exponential complexity.

When we started building Acompli in 2013, we had a simple goal:

Create an app that combined personal Gmail accounts with Microsoft Exchange work accounts in a delightful way.

Read 19 tweets

Kevin Henrikson

@KevinHenrikson

Jan 17

Your technical interview process is wrong.

After building multiple engineering teams at Microsoft, VMware & Instacart,

I've discovered what really predicts performance.

Here are the 3 traits of the best engineers:

First, let me tell you about my most expensive hiring mistake:

A candidate aced every technical question. Perfect scores on coding tests.

Impressive resume from top tech companies.

6 months later? We had to let them go.

Here's what I learned:

Technical skills are table stakes.

What's far more important is how someone thinks and approaches problems.

After interviewing hundreds of engineers and building teams that sold for $200M, I've developed an unusual way to spot top performers:

Read 18 tweets

Kevin Henrikson

@KevinHenrikson

Jan 16

A 64-year-old grandmother is beating tech billionaires at their own game.

While they spend millions on experimental longevity treatments...

She's achieving remarkable biological age results with a simple lifestyle.

Here's her fascinating story:

Meet Amy Hardison.

Mother of 4. Grandmother to 11.

Earlier this year, she ranked #4 on the Rejuvenation Olympics leaderboard with her DunedinPACE test results, outperforming @bryan_johnson at the time.

But that's just the beginning...

To understand why this is remarkable, let's talk about biological age:

Your chronological age is how many years you've been alive.

But your biological age? That's how old your body actually is.

And Amy's DunedinPACE score was 0.74, meaning she's aging slower than average.

Here's what this means: