Linas Beliūnas

Google just did to commerce what HTTP did to the web - Universal Commerce Protocol, or UCP, is a new open standard for AI agent shopping 😳

Put simply, Google just dropped a new default for how buying works.

UCP is an open, shared language for AI agents, merchants, and payments to transact end-to-end - autonomously 🤖

Discovery → comparison → negotiation → checkout → returns.

All handled by software.

For 20 years, commerce followed the same flow:

Search → ads → product pages → checkout.

UCP breaks that flow. Now it’s:

Intent → agent reasoning → purchase.

No clicks.
No SEO games.
No funnels.

This isn’t theoretical. UCP launched with some of the biggest names in tech & finance:

Shopify, Walmart, Target, Etsy, Wayfair, Visa, Stripe, Adyen 💳

20+ partners on day one. That’s how standards win.

What UCP actually unlocks:

→ AI agents discover merchant capabilities automatically
→ Negotiate prices, bundles, loyalty, subscriptions
→ Execute payments securely
→ Handle post-purchase support
→ All without custom integrations

This kills the N×N integration problem that has crippled e-commerce for decades.

More importantly, UCP doesn’t replace existing rails. It connects them.

↳ A2A → agents talking to agents
↳ MCP → shared context
↳ AP2 → cryptographically signed payment mandates

UCP is the glue for AI-powered commerce.

And Google is uniquely positioned to win this.

Because at the core, commerce = matching intent with supply at scale.

Google already owns:

→ Global search intent
→ The largest product graph
→ AI models (Gemini) embedded everywhere
→ Distribution (Search, Android, YouTube)

UCP turns that into rails, not a walled garden.

Merchants stay merchant-of-record.
Users keep choice.
Agents do the work.

That’s why this can become the HTTP of commerce.

This marks the beginning of the real disruption.

Brands won’t compete for attention anymore.
They’ll compete to be chosen by machines.

Websites become optional.
Checkout buttons become legacy UI.
Loyalty shifts from humans → agents.

And this is the start of non-human commerce.

Quiet.
Autonomous.
Everywhere.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

Apple just did the most un-Apple thing in years - it officially picked Google’s Gemini AI for the new Siri, and it might be the smartest move of the AI era 😳

It’s a multi-year partnership where Google’s Gemini AI becomes the brain behind the long-awaited new Siri.

Here’s what that actually means:

→ Apple will pay Google ~$1B per year
→ Apple gets access to a custom Gemini model (~1.2T parameters)
→ That’s ~8x larger than Apple’s current internal models
→ Gemini handles reasoning, planning, summarization
→ Apple handles on-device + privacy-critical tasks
→ All data stays inside Apple’s Private Cloud Compute

Google supplies the horsepower while Apple controls the experience.

Classic Apple 🍎

This is a huge deal because Siri has been stuck in 2014 energy for a decade.

Meanwhile:

↳ ChatGPT became the default “thinking layer”
↳ Google rebuilt its whole business around AI
↳ Anthropic, DeepSeek shipped frontier models at insane speed

Apple was falling behind. Fast.

And building a frontier LLM from scratch isn’t just “hard”. It burns tens - even hundreds - of billions.

Compute. Data. Iteration speed. All costs money.

Google already solved that, so Apple made a call most companies are too proud to make.

Pragmatism > Pride.

And no, this isn’t Apple surrendering.

It’s Apple doing what Apple has always done:

They didn’t invent CPUs.
They curated them → then replaced them.

They didn’t invent modems.
They used Qualcomm → then built their own.

Now it’s AI.

Borrow the best brain now to:

→ Buy time
→ Ship a real product
→ Build your own quietly in the background.

Insiders say Apple’s own trillion-parameter model is coming next.

This is a bridge, not a dependency.

Why did Apple choose Google, not OpenAI, not Anthropic?

Because Gemini is:

↳ Strong in multi-language
↳ Excellent at context + planning
↳ Built for automation at scale
↳ Already proven in production systems

And above all, Apple already trusts Google.

Remember that Google Search has been the default on iPhone for years.

Billions flow between them annually.

And most importantly, this deal signals something bigger:

The hybrid AI era is here.

No single company does everything anymore.

The winning stack now looks like this:

→ Partner for raw intelligence
→ Own UX, privacy, distribution
→ Differentiate where users actually feel it

AI is becoming infrastructure, not a feature.

And Apple just accepted that reality.

Nobody can beat Google. It’s basically the Internet 😳

→ Google owns 14% of Anthropic and 8% of SpaceX.

→ Acquired DeepMind for pennies on the dollar.

→ Runs Gemini AI that’s going to power Apple’s new Siri.

→ Powers Claude with TPU chips.

→ Handles 90% of all the searches.

→ Owns Waymo, while also tracking the world through Maps.

→ Runs YouTube that dominates TV screen time.

→ Owns Android that’s powering 3B +devices.

→ Dominates ads, email, cloud, browser.

→ Leads with advancements in quantum computing.

→ About to put GPUs in space with Starcloud.

→ Just launched UCP to dominate AI-driven commerce.

When you think about it, Google isn’t really the internet…

Google is what the internet runs on.

Google is everything and everywhere.

P.S. also check out deep dive into Google's UCP, the new default for how AI buys things: https://lnkd.in/dSKWjJPy

My life advice for 2026: find someone who looks at you the same way Big 4 consultants look at a Fortune 500 company that needs an “AI Strategy” 😉

Seriously, consulting giants are printing money at will with AI:

→ Accenture has reported billions in generative AI related booking, with some estimates suggesting ~$3B this year alone.

→ KPMG is said to be making up to $1 billion/year from AI consulting.

→ 40% of McKinsey’s work is AI tech, that’s ~$6.4B per year.

→ 20% of BCG’s revenue in 2024 came from AI, that’s about $2.7B/year

The wild part?

Consulting companies are making more profit from AI than 99% of AI startups.
Combined.

AI isn't just hype - it's a goldmine.

And consultancies are digging deep 💸

P.S. for more great stuff, check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

Vibe coders at 2 am building another $3B AI startup 👏😂

Just pure instinct, caffeine, and vibes.

You’re not writing code.
You’re summoning it.

The fans spin up like a jet engine.
The UI half-works.
The animation is buttery… on localhost.

At 3 am, code stops doing what you intended and starts doing what you meant.

Suddenly, everything clicks.

You feel unstoppable. You tell yourself:

Software engineering is basically solved.

Then the API credits run out.

Hard stop. Forced bedtime.

P.S. for more great stuff, check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

100s of AI startups died today: Anthropic just dropped Claude Cowork, and it might be the most important step yet toward AI-native work 😳

Claude Cowork is basically Claude Code, but for non-developers.

You give Claude access to a folder.
You give it a goal.
It plans, executes, fixes mistakes, and keeps going.

No prompts every step.
No copy-pasting between apps.
No babysitting.

Just: “Handle this.” And it actually does 🤖

Here’s what Cowork can already do:

→ Read, write, and reorganize your files
→ Clean inboxes and folders
→ Generate reports from messy spreadsheets
→ Build decks, docs, and summaries
→ Run multi-step workflows across tools
→ Browse the web and pull fresh data
→ Ask clarifying questions when unsure

This is probably the first real glimpse of agentic AI for normal work.

This is literally what Microsoft Copilot and ChatGPT Desktop should have been.

Now here comes the crazy part.

Claude Cowork was 100% built by Claude itself.

Anthropic shipped it in ~1.5 weeks using Claude Code 🤯

AI is now building AI tools to make AI even more useful.

And this marks a turning point.

For years, autonomous agents were:

- Too fragile
- Too risky
- Too technical

Cowork changes that.

→ Runs in a sandboxed VM
→ Human-in-the-loop by default
→ Designed for real work, not demos

Most importantly, Claude Cowork is a preview of something bigger:

An LLM as your operating system.

You don’t open apps.
You describe outcomes.

Claude orchestrates:

Drive → Docs → Sheets → Slides → Web → Back again.

The desktop becomes AI-native.
Apps become implementation details.

And this is exactly what AI looks like right before it becomes unavoidable.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

Europe is finally rising: Ursula von der Leyen officially just announced EU Inc., a single, digital-first company structure for all of Europe 😳

Not 27 countries.
Not 27 rulebooks.
One startup Europe.

The coolest part?

This didn’t come from a think tank. It started as a grassroots petition in 2024.

13,000+ founders and investors signed it. Europe listened 👏

Here’s what EU Inc. actually does:

→ Incorporate online in 48 hours
→ No mandatory notaries
→ One legal entity, valid across all 27 EU states
→ Standardized rules for equity, insolvency, and cross-border ops

Taxes and labor laws stay national while the corporate layer goes European.

That distinction matters because Europe’s biggest problem was never talent.

It was friction.

Until now, scaling meant paperwork hell:

→ New entities per country
→ Incompatible equity frameworks
→ Legal bills eating seed rounds alive

So the founders did the rational thing.

They left.
Delaware won.
Europe lost compounding.

EU Inc. flips that equation. And for founders, this is massive:

→ Faster launches
→ Cheaper compliance
→ Founder-friendly equity structures
→ US-style fundraising without leaving the EU

A founder in Vilnius or Lisbon now has frictionless access to a 450M-person market on Day 1.

No relocation required.
No regulatory gymnastics.

Most importantly, this is the first time Europe chose builders over bureaucracy.

If executed as a true regulation (not 27 interpretations), EU Inc. could unlock:

→ More European unicorns
→ Less brain drain
→ Stronger AI, biotech, and climate tech champions
→ A real Single Market, not a theoretical one

Of course, execution will be the key here, but one thing is already clear:

Europe is finally building for founders 🇪🇺🚀

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

The age of KYC video verification is officially over: the latest video models are now so real that nobody can tell they’re AI generated 😳

This is the moment identity verification finally broke.

For years, video was the final proof layer.

→ “Turn your head”
→ “Blink twice”
→ “Smile for the camera”

Banks trusted it.
FinTechs built on it.
Digital platforms enforced it.

In 2026, it fails effortlessly.

With a single reference clip, tools like Kling AI can now generate:

→ Perfect facial expressions
→ Natural micro-movements
→ Realistic timing and emotion
→ Full liveness… without a human

And here’s the uncomfortable part.

If an AI can pass a live video call, video KYC is no longer verification.

It’s a theatre.

We’ve officially entered a world where:

Seeing ≠ believing
Liveness ≠ human
Face ≠ identity

So the next big shift is from visual trust → systemic verification

↳ Behavior, not faces
↳ Signals, not selfies
↳ Patterns, not pixels

Zero-trust identity becomes the default.

And in this new world, AI-native risk infrastructure is no longer optional.

AI Risk Decisioning platforms like Oscilar don’t ask “Does this video look real?”

They ask:

→ Does this behavior make sense?
→ Do signals align across sessions?
→ Does intent match history?

That’s how you fight AI fraud. With AI 🤖

Because deepfakes are no longer just memes. They’re:

↳ Identity hacks
↳ Scam engines
↳ Financial attack vectors

AI broke video verification. AI will also defend what comes next.

Seeing is no longer believing.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

AI’s Instagram moment? Meta Platforms just bought an autonomous agents startup Manus AI for more than $2 billion, making it Meta’s 3rd largest M&A of all time 😳

The deal is a clear signal of where AI is heading next.

Manus built an AI that executes work, not just conversations.

It runs inside a sandboxed virtual computer with memory, tools, internet access, and the ability to write and run code.

You give it a task → It completes it end to end.

That capability changes the category.

Manus quickly stood out because it could handle real workflows:

→ Market research
→ Portfolio analysis
→ Job candidate screening
→ Trip planning
→ Full application builds
→ Business process automation

Tasks that normally require hours of human coordination could be completed with minimal supervision.

But the strongest signal wasn’t the demos.

It was traction.

Manus crossed $100M in ARR in under 8 months through a straightforward subscription model.

Millions of users.
Heavy usage.
Real willingness to pay.

At a time when most AI products prioritize growth over revenue, Manus proved autonomous agents can sustain a business.

For Meta, the timing is deliberate.

The company has already committed tens of billions to AI infrastructure and open-source models like Llama.

What it needed was a layer that turns intelligence into action.

Manus thus enables:

↳ Agents that manage campaigns and operations for businesses
↳ Agents that coordinate scheduling and communication inside WhatsApp
↳ Agents that execute shopping and workflows on Instagram
↳ AI systems that behave more like persistent operators than reactive assistants

This creates a path from “AI feature” to “AI worker.”

The Instagram parallel seem to fit perfectly here.

→ Instagram changed how people behaved on mobile, then scaled globally through Meta’s distribution.

→ Manus has similar characteristics: viral adoption, practical utility, and early monetization but limited reach as a standalone product.

Meta’s 3 billion-user ecosystem removes that constraint.

There’s also a broader undercurrent.

Manus originated in China, relocated to Singapore, and outperformed many Western systems on real-world agent benchmarks.

The most capable AI teams are no longer concentrated in one geography.

More importantly, acquiring Manus pulls execution-grade agent technology directly into a major Western platform.

My takeaway is this:

This acquisition marks a shift.

→ AI is moving from assistance to execution.
→ From answering questions to completing work.

If Meta executes well, autonomous agents won’t feel novel or visible.

They’ll quietly become part of how work and coordination happen at scale.

And that’s exactly how platform shifts usually begin.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

Nailed it! Joanna Maciejewska gives the only perspective on AI that matters 👏

That's exactly why Elon Musk has been building Tesla Optimus, Figure and 1X are developing humanoid robots, and even Google & Apple are doubling down on robotics.

The next 2-3 years will be wild.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

Turns out, Ben Affleck has a better understanding of the limitations and issues in AI than 99% of the VCs that are actively investing in the space 😳

"A lot of that rhetoric comes from people who are trying to justify valuations around companies where they go, 'We're gonna change everything in two years, there's going to be no more work.'"

And then Affleck said the quiet part out loud:

AI doesn’t create. It averages.

And averaging is the enemy of originality.

The most misunderstood part is that it’s not anti-AI.

That’s an accurate description of how LLMs work.

They compress the past.
They smooth the edges.
They converge toward the mean.

Which is great for:

→ Summaries
→ First drafts
→ Busywork
→ Tedious cognitive labor

And terrible for:

→ Voice
→ Taste
→ Risk
→ Saying something new

Creativity and judgement have never been more important.

P.S. also check out 120+ Real AI Use Cases From Leading Startups & Internet Companies: https://lnkd.in/drT27Bvb

Teaching robots how humans work so machines can replace them 🤖

At first, I thought this was AI generated…

Turns out, it’s 100% real and was happening in a Shenzhen factory this year.

Surreal.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

Historic: Elon Musk just negotiated the largest private acquisition in history with himself - SpaceX officially acquired xAI for $250 billion to create a $1.25 trillion space and AI powerhouse 😳

At first, SpaceX & xAI M&A at a ~$1.25T combined valuation looks like financial engineering.

But it might also be vertical integration at a planetary scale.

Remember - AI doesn’t just scale with data. It scales with power 🔋

And Earth is running out of easy power.

Why this combo is different:

→ xAI already runs 100k+ Nvidia H100s. Next training runs need gigawatts, not megawatts. The grid isn’t built for that.

→ Solar in orbit produces multiple times the annual energy of ground arrays - no weather, no night cycle.

→ Space offers free cooling. Heat radiates into vacuum instead of burning water and electricity on chillers.

→ Starship’s target launch costs change the math. Shipping hardware to orbit starts looking less insane vs decade-long grid upgrades on Earth.

→ Starlink + X payments rails = distribution + transactions. Not just models answering questions - AI agents that can act and pay.

Everyone else is fighting for land, permits, and substations.

This stack is trying to move the bottleneck to orbit.

It’s either the most ambitious infrastructure play in tech, or a very expensive way to reprice an AI company.

Probably both.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets AI. For founders, builders, and leaders.

This is basically how 90% of startups are born in 2026 👏🏼😂

Not from strategy.
From vibe.

Coders are now vibe coders.
Marketers are now vibe marketers.
Founders are now vibe founders 😎

This is the new economy.

P.S. on a serious note, check out The First Principles Guide to Building AI-Powered Engineering Teams & Products: https://lnkd.in/dvNVmhgd

Wild: AI giant Google now has a literal line item in their quarterly reports called "European Commission fines" 😳

$10.6 billion, that’s how much Google has set aside for EU fines.

Not as a crisis. As an accrual 😬

Regulation has become predictable, and predictability turns punishment into pricing.

Since 2017, Europe has fined Google €11B+ for antitrust violations.

Shopping.
Android.
Adtech.
Adtech again.

The latest one alone is €2.95B, and it’s for self-preferencing its own ad exchange.

Google appeals. The cash doesn’t move.

The number quietly lands on the balance sheet.

Like rent.
Like payroll.
Like cloud costs.

This is why people are mad.

Some see accountability, while others see overreach.

But both miss the real signal.

When fines are big but survivable, they don’t change behavior.

They validate dominance.

True deterrence breaks structures. This just invoices them.

Most importantly, fines don’t build challengers.

Startups do, and Europe keeps starving the one thing that could actually replace Google.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

The unthinkable just happened: 3 years after Google hit “Code Red” over ChatGPT, OpenAI's Sam Altman just hit Code Red over Google 😳

$500B AI giant OpenAI has reportedly paused ads, shopping, and every side project.

All hands are now on deck to fix ChatGPT.

Because Google’s Gemini 3 just beat GPT-5 on math, code, and reasoning.

The AI tables have officially turned.

Altman’s recent memo admits it’s “a critical time for ChatGPT.”

Internally, the vibes are very rough.
Externally, the pressure is insanely immense.

→ $500 billion valuation.
→ $8 billion annual burn.
→ And Google’s AI models are climbing fast.

So OpenAI is now regrouping.

A new “reasoning model” drops next week, rumored to reclaim the crown.

But make no mistake: this is survival mode.

The same company that made Google panic in 2022 is now fighting to prove it hasn’t peaked in 2025.

It’s kinda poetic…

Because every “code red” in tech history sparks one of two things: a comeback story or a collapse.

And right now, the world’s most powerful AI company is deciding which one it will be.

Turns out, history doesn’t repeat itself in AI - it retrains.

Google CEO just told the one AI nightmare that keeps him awake: “When deepfakes get so good that we literally won’t be able to tell what’s real anymore.. and bad actors get their hands on it” 😳

His exact words after Shannon Bream pressed him:

“That’s the kind of thing you sit and think about” - Sundar Pichai

This 64-second clip is chilling.

Because it’s not about sci-fi doom. It’s about a trust collapse.

When seeing stops meaning believing, democracy, finance, and human relationships all start to fracture.

Imagine:

→ A fake video of a president announcing a market crash.
→ A CEO “approving” a billion-dollar transfer.
→ Your voice cloned for a scam you never made.

That’s not the future. It’s already happening.

AI can cure cancer, or collapse consensus reality.

The difference lies in how we build safeguards as fast as we build capabilities.

And that’s why companies like Oscilar are becoming critical infrastructure for the AI era.

They’re building agentic systems that can detect deepfake-enabled fraud, verify identities, and flag anomalies in milliseconds.

Turning AI against its own dark side.

Because in a world where every pixel, video, or voice can lie, the new currency isn’t data.

It’s trust. And whoever protects it, wins.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

Tech loves irony: Apple, the company accused of “doing nothing in AI,” is now quietly becoming an AI infrastructure company 😳

For years, the narrative was simple: Apple is behind in AI.

No frontier model.
No flashy demos.
No arms race with OpenAI or Google.

And yet developers are now wiring Mac minis into AI clusters to run models that used to require data centers.

No hype. Just physics, systems, and taste.

The truth is that Apple never optimized for cloud AI. It optimized for local intelligence.

Unified memory.
Ridiculous bandwidth per dollar.
Neural Engines everywhere.
Hardware + OS designed as one system.

That combo accidentally unlocked something big:

↳ Inference loves memory more than brute force
↳ Batch size = 1 is the real world
↳ Most AI workloads don’t need a nuclear reactor

So while everyone chased FLOPS, Apple made efficient compute boringly good.

Now look at what’s happening:

→ Developers cluster Mac minis instead of buying GPUs 
→ Large models run locally, offline, private
→ Power draw measured in hundreds of watts
→ Costs drop from “enterprise-only” to “indie-accessible”

And the irony?

The company accused of doing nothing in AI is quietly becoming an AI infrastructure company.

And this is classic Apple.

Not winning headlines but winning constraints.

Turns out, Apple wasn’t late in AI.

It was simply playing a different game the entire time.

Anthropic just killed traditional UIs: Claude now has an app layer inside chat, and you can use real applications & build things through MCP Apps 😳

It’s basically software showing up inside the conversation.

You ask Claude a question.
It pulls in the tool.
The work happens there.

And that changes the role of SaaS entirely.

We’re moving from:

“Open the app, then do the work”
to
“Describe the work, the apps come to you.”

Early signals this is bigger than a feature launch:

→ Amplitude is a launch partner - product analytics charts can be generated and edited directly inside Claude
→ Tools like Slack, Figma, and Asana can be triggered from chat instead of separate dashboards
→ The interface layer shifts from menus and tabs → natural language
→ When AI becomes the primary user, seat-based SaaS pricing starts to look… exposed

For 20+ years, SaaS companies competed to be the place you log into every morning.

Now, the most important product might be the one you never leave.

Software isn’t disappearing.
But it’s becoming the back office to an AI front desk.

And if chat becomes the operating system, dashboards start to look like legacy software.

Most importantly, when the interface shifts, the power usually shifts with it.

P.S. check out this deep dive into the 🤖 Agentic Interface 🤖, & how AI companies are going all in on the Agentic OS: https://lnkd.in/dva3Eejf

Meet Jim Simons: an 86-year-old math genius who built the greatest quantitative hedge fund outperforming Warren Buffett, George Soros, Carl Icahn, and every investor who ever lived 😳

Jim's story was absolutely wild.

At 40, Jim Simons left a famed math career to launch the most successful hedge fund ever: Renaissance Tech.

This fund earned more than $100 billion for its investors.

Jim Simons's Medallion fund has had unreasonably high annual returns of 66% across 30 years 🤯

For the perspective, Buffet clocked in at about 20% per year.

Before passing away, Simons was worth over $31 billion, and his secret was dead simple:

→ He wanted to smoke everywhere.
→ He walked around the office barefoot.
→ He fell asleep at presentations.

The lesson: cold showers and meditating won't make you rich.

The real secret to wealth is just focusing on your thing and the freedom to be unapologetically yourself.

What a legend.

Wild: One of the world’s top engineers just said that Claude Code built in 1 hour what took a Google team one year 😳

Jaana Dogan, a Google Principal Engineer, gave Anthropic’s Claude Code a high-level description of a distributed agent orchestrator.

A system her team had debated, designed, and stalled on for a year.

Claude shipped a working prototype in ~60 minutes 🤯

But that’s not the shocking part.

What’s shocking is that Google allows its engineers to use Claude Code instead of forcing Gemini, or their latest IDE Antigravity.

So AI didn’t out-code Google. It AI out-ran bureaucracy.

Big teams don’t move slow because engineers are bad.

They move slow because:

→ Alignment takes weeks
→ Reviews take months
→ Decisions get diluted

AI does something brutal and powerful now: it produces Version 1 immediately.

And once something exists, debate collapses.

As legendary Paul Graham has said before: “A concrete thing beats endless discussion".

And there’s one more uncomfortable truth…

If your company forces internal tools while better ones exist outside, you’re optimizing for politics - not output.

The best teams will win by doing one thing and one thing only:

Give builders the best tools, and then get out of the way.

Revolut just made your phone safer than your wallet. And it might have also just redefined what security truly means in FinTech 😳

Every 2 minutes, a phone is stolen in London.

But the real danger isn’t losing your device - it’s what happens after.

Criminals now force victims to unlock banking apps and authorize instant transfers at knifepoint.

No hacking. No phishing. Just fear.

That’s the brutal new reality Revolut is fighting.

That’s why Europe’s most valuable FinTech just launched Street Mode, a feature that makes your money harder to steal in the real world.

Here’s how it works:

↳ Inside your “safe zones” (like home or work):
Transfers flow normally.

↳ Outside those zones (say you’re walking at night):
Large transfers trigger a 1-hour delay and a second selfie check before approval.

That’s 60 minutes for you, or Revolut, to stop the fraud.

A digital cooling-off period for the physical world.

Here’s why it matters:

→ The UK saw 78,000+ phone snatchings last year, many ending in “transfer muggings.”
→ Instant payments made fintech fast, now Revolut’s making it safe.
→ This is security that adapts to you - where you are, not just who you are.

Think of it as “adaptive finance security.”

Banks have long focused on digital fraud.
Revolut’s now tackling physical coercion.

And that’s a paradigm shift.

Competitors will not only copy this. They NEED to copy this as soon as possible.

→ Apple hinted at it with “Stolen Device Protection.”
→ Monzo tried with “Known Locations.”
→ Now Revolut’s Street Mode is the first to give users control, not just algorithms.

This is the future of FinTech security that we also envision at Oscilar:

↳ Context-aware
↳ User-controlled
↳ Real-world smart

Because safety isn’t just about passwords anymore.

It’s about presence, place, and time.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

Wild: Duolingo announced it was going “AI-first” just over 6 months ago. Since then its stock is down 64% 😳

Duolingo just discovered the hard way that AI can scale a product, but it can’t save a soul.

In April, they went “AI-first” 🤖

→ Fired human contractors.
→ Let AI write lessons.
→ Pushed out 50 new courses overnight.

And for a moment, Wall Street loved it.

- Stock pumped to $540.
- Efficiency! Margins! Infinite scaling!

Then reality hit.

Users opened the app… and got sentences like:

“The bears are drinking beer at the party.”
“The cows cleaned the kitchen yesterday.”
“The cats went to the market today.”

It felt less like learning Spanish and more like babysitting a hallucinating chatbot.

People with 500-day, 800-day, and even 1,000-day streaks quit overnight.

Competitors with human-curated content saw a surge.

But here comes the real story nobody’s talking about.

AI didn’t kill Duolingo. Duolingo killed Duolingo’s trust.

Language is deeply human.

Idioms, tone, context, culture - that’s where humans shine, and LLMs fail.

Yes, AI can accelerate content.

But when it replaces the very thing users value most, you don’t get efficiency. You get “enshittification.”

And once trust cracks, no streak counter can fix it.

This is the lesson every consumer company should remember:

AI is an amplifier.

If your product is good, AI makes it great.
If your product is fragile, AI makes it crumble faster.

Turns out, Duolingo just accidentally shipped a masterclass in what NOT to do with AI.

Bill Gates would be worth $1.47 trillion if he still had his Microsoft shares 😳

But he became friends with Warren Buffett and started to diversify his portfolio.

He’s worth only $106 billion today.

Lesson: choose your friends wisely.

P.S. for more great stuff, check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

For anyone who thinks they need a perfect product before launch, always remember Google Maps went live in 2005 without Europe 😳

No Asia.
No Africa.
Half the planet was missing.

And yet it still became the most-used map in history.

The lesson?

If Google could launch Maps without Europe, your product doesn’t need every feature polished.

Ship it.
Learn.
Expand.

Perfection kills startups. Momentum builds giants 🚀

h/t Ben Gilbert & Acquired

Wild: NVIDIA just announced Alpamayo, the world’s first thinking, reasoning autonomous vehicle AI, launching on US roads later this year 😳

Put simply, NVIDIA just taught cars how to think out loud.

Until now, self-driving systems did two things: they saw the world & then reacted to it.

Alpamayo adds the missing layer.

It reasons.

Camera input → internal logic → driving action.

And the system can explain why it did what it did.

Not vibes.
Not a black box.
Actual decision logic.

In the demo, the car doesn’t just slow down. It explains:

→ Why a pedestrian looks uncertain,
→ Why a construction zone changes priorities,
→ Why yielding now avoids risk later.

That single change unlocks everything.

Debugging.
Trust.
Regulatory approval.
Real autonomy.

This matters because autonomy doesn’t fail at 80% or 90%.

It fails at 99%.

The last 1%: weird edge cases, confusing human behavior, scenes no rules engine can hard-code.

That long tail is where accidents live, and reasoning is how you attack it.

The coolest part?

NVIDIA didn’t keep this closed. They open-sourced the entire stack: the model, the tools, the simulator, the data.

That’s a brilliant strategy.

It’s basically Android for autonomy.
ROS for self-driving.

Software becomes a commodity.
Compute becomes the choke point.

Guess who sells the compute 😎

And no - this isn’t bad news for Tesla.

If anything, it validates Tesla’s vision-only, end-to-end approach.

But here’s the hard truth Elon keeps pointing to…

You don’t solve the last 1% in simulation alone. You solve it with real-world miles.

At scale. Every day.

Alpamayo helps everyone start. Data moats decide who finishes.

Most importantly, this isn’t really about cars.

It’s about physical AI.

Because the same reasoning stack works for robots, factories, warehouses, drones, and machines operating in the real world.

Once AI can reason about reality, software stops being the bottleneck.

Reality does.

And NVIDIA just moved the entire industry one step closer to that.

Tech loves irony: Apple, the company everyone says is “behind in AI,” is quietly becoming one of the most important AI infrastructure players on the planet 😳

For years, the story was obvious:

No frontier model.
No AGI race.
No flashy demos.
No GPU megaclusters.

Compared to OpenAI, Google, or Microsoft, Apple looked irrelevant in AI.

And yet, something strange is happening.

Developers are wiring Mac minis into AI clusters to run workloads that used to require data centers.

No hype.
Just systems, efficiency, and taste.

The truth is Apple never optimized for cloud AI.
It optimized for local intelligence.

Unified memory.
Absurd memory bandwidth per dollar.
Neural Engines everywhere.
Hardware and OS designed as one system.

That choice accidentally unlocked a different AI reality:

↳ Inference is memory-bound, not FLOP-bound
↳ Batch size = 1 is how AI is actually used
↳ Most workloads don’t need hyperscale compute

While everyone chased brute force, Apple made efficient compute quietly excellent.

Now look at what’s unfolding:

→ Mac minis replacing GPUs for many inference tasks
→ Large models running locally, offline, and private
→ Power draw measured in hundreds of watts
→ Costs collapsing from “enterprise-only” to “indie-viable”

Even legendary Andrej Karpathy has pointed out that Mac minis are unusually well-suited here once you price memory bandwidth, not FLOPS.

And the newest twist?

Always-on personal AI agents are turning Mac minis into “home AI brains” - stable, silent machines that own intelligence instead of renting it from the cloud.

The closest we have today to Jarvis from Iron Man 🤖

And the irony?

The company accused of doing nothing in AI is quietly becoming one of the most practical AI infrastructure players in the world.

This is classic Apple.

Ignore the narrative. Win the constraints.

Apple wasn’t late to AI.

It was just playing a different game the whole time.

Fascinating: Google is about to pull off one of the greatest investments of all time 😳

In 2015, Google spent ~$900 million on SpaceX.

That single check - 7.4% of SpaceX at a $12 billion valuation - could soon be worth ~$111 billion if SpaceX goes public at its rumored $1.5 trillion valuation.

A 123x return 🤯

That alone would make it one of the greatest venture investments ever.

But the return isn’t the real story. The real story is what Google actually bought.

Over the last decade, SpaceX quietly stopped being “just” a rocket company:

→ Starlink now has ~8M subscribers
→ $11.8B projected revenue in 2025
→ ~90% of all global payload mass to orbit
→ Reusable rockets flying 20+ times
→ 150+ launches per year

This isn’t aerospace anymore. It’s planetary-scale infrastructure.

And that’s where AI comes in.

AI doesn’t have a model problem. It has an energy problem.

Data centers are hitting hard limits: power, cooling, land, and grid approvals.

Training better models is easy. Finding enough electricity to run them isn’t.

Space changes the equation.

→ Constant solar power
→ Easier cooling
→ Launch costs collapse with reusability

Starlink already proved you can deploy and operate massive networks in orbit.

Compute is the obvious next layer.

Which is why Google’s SpaceX stake was never just financial.

It was a long-dated option on a future where AI infrastructure no longer fits on Earth.

Most people saw rockets. Google saw where compute eventually has to live.

3D chess.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

400 million AI users overnight: Microsoft just renamed Office to “Microsoft 365 Copilot app” 😳

Word, Excel, and PowerPoint - now branded around AI, whether you asked for it or not.

So even the AI haters can’t escape it now 😂

Microsoft hopes AI will feel inevitable, but unfortunately, it feels… forced.

Social media already has a name for it: Microslop 👏

And this is what happens when a $3.5 trillion company tries to reboot a mature monopoly with hype instead of taste.

→ AI slapped on top of workflows ≠ AI-native software.
→ Renaming ≠ reinvention.
→ Adoption ≠ acceptance.

The lesson here is not about AI being overhyped.

It’s sharper than that:

If AI doesn’t quietly make work better, users will loudly reject it.

Most importantly, the future of productivity won’t feel like Copilot or anything AI-powered.

It’ll feel like nothing changed, except everything works better.

Because the best AI features don’t need branding.

P.S. on that front, check out the 120+ AI Use Cases From Leading Startups & Internet Companies: https://lnkd.in/drT27Bvb

iRonic: a Roomba robot sits on a chair watching a human mop the floor inside a Roomba store 👏😂

This is peak automation.

Turns out, the robots are not replacing us just yet.
They are middle management now.

No wonder iRobot filed for bankruptcy.

AI is finally getting personal: Google just dropped Personal intelligence, and it’s the AI moat nobody can replicate 😳

Everyone is racing to add “memory.”

→ Chatbots remember chats
→ Assistants store preferences
→ Apps learn inside their own walls

Google did something else entirely.

It flipped the switch on 20 years of your digital life.

And it might be the deepest personalization layer ever built.

Gemini can reason across:

→ Your Gmail threads
→ Your Photos library
→ Your Search history
→ Your YouTube watch patterns

Not uploads.
Not prompts.
Your real-life context.

Forgot your car model?
Gemini finds it in an old email and recommends tires.

Planning a trip?
Gemini cross-references your calendar, photos, and interests - then suggests places you’ll actually like.

This is not about smarter AI. It’s about more relevant AI.

Every AI company now wants personalization. But personalization needs data gravity.

And Google already has it.

→ A decade of emails
→ Trillions of photos
→ Billions of searches
→ Behavioral signals at planetary scale

All first-party.
All opt-in.
All deeply integrated.

No one else can copy this.

↳ OpenAI can’t read your Gmail
↳ Anthropic can’t see your Photos
↳ Perplexity only knows what happens inside its app

They start every conversation from zero.

Google starts from you.

And this changes the AI race completely.

Models are converging.
Data pipelines are not.

The new battleground is all about who knows the user best.

And Google has been quietly building that advantage since 2005.

Context is all you need.

Microsoft CEO Satya Nadella just called out the AI bubble, and almost no one noticed 😳

His warning was simple and brutal:

If AI’s benefits don’t spread widely, this turns into another hype cycle.

The biggest AI distributor in the world is basically saying that this can still fail 🤖

Not because the tech is fake.

But because usage isn’t keeping up with investment.

The brutal truth is that AI capex is exploding.

Data centers. GPUs. Power grids. Trillions committed.

Yet, the real adoption is thin.

→ Copilot is everywhere by default.
→ Yet barely anyone is using it.

That’s the danger zone.

Because bubbles don’t burst when technology is useless.

They burst when expectations outrun reality.

The only question left is whether that gap is Microsoft’s problem to solve, or everyone’s bill to pay.

Co-founder & ex-CEO Sergey Brin just admitted that Google invented the future, and yet still lost the first round in AI 😳

In 2017, Google created transformers (“Attention is All You Need”).

The architecture behind everything we now call AI.

And then it hesitated….

- Because chatbots hallucinated.
- Because error rates were ugly.
- Because breaking things is expensive when you serve billions.

So Google waited.

Meanwhile, OpenAI didn’t. They shipped anyway.

Imperfect. Risky. Messy. And ChatGPT lit the world on fire 🤖

It’s a classic tech pattern.

→ Giants invent.
→ Startups dare.
→ Giants return with scale.

We’ve seen it before multiple times:

↳ Xerox → GUI → Apple
↳ Kodak → digital photos → everyone else
↳ Nokia → touchscreens → iPhone

Same movie. New actors.

But fast-forward to 2025, and Google has flipped the switch.

- Gemini is topping benchmarks.
- 91%+ accuracy on brutal tests.
- Brin is back in the trenches.

Because once risk became acceptable, scale did the rest.

That’s the unfair advantage of incumbents: when they move, they move everywhere.

My takeaway is this:

The real lesson here isn’t that “Google messed up.”

It’s this: startups change the rules, but big tech industrializes the win.

No transformers → no ChatGPT.
No ChatGPT → no AI arms race.

Both had to happen.

And somewhere right now, a tiny team is holding the next AI or quantum breakthrough while a trillion-dollar company still debates the PR risk.

AI is eating consulting? McKinsey just confirmed it now counts AI as part of its workforce - 40,000 humans & 25,000 AI agents. That’s 65,000 “employees” 😳

Chief Executive Bob Sternfels made it official:

→ AI already powers 40% of McKinsey projects
→ Agents produced 2.5M charts in 6 months
→ 1.5M human hours saved per year

The goal is to have 1 AI agent per human 🤖

McKinsey didn’t fire people (yet). But they deleted entry-level work.

The boring parts are now gone:

→ Desk research
→ Slide drafting
→ Data cleaning
→ First-pass analysis

That work now happens at machine speed. Humans move up the stack instantly.

But the real bet the consulting giant is making isn’t about productivity.

It’s about optionality.

McKinsey is building a firm that can:

→ Scale without hiring
→ Serve more clients with the same headcount
→ Recompose teams instantly
→ Price expertise, not hours

Consulting then becomes software-like.

High margins.
High leverage.
Low friction.

And this is the real future of work.

Humans won’t execute.
They’ll direct, judge, and decide.

AI will do the rest.

P.S. check out Top 10 AI Startups in 2026 & Their Decks: https://lnkd.in/dY7b4kT4

Competition never stops: GPT-5.2 just dropped, and OpenAI calls it “the smartest generally-available AI model in the world” 😳

But here’s what almost everyone is missing:

GPT-5.2 didn’t just beat benchmarks, it also beat professionals.

On GDPval, the brutal test across 44 real occupations - law, finance, engineering, medicine - GPT-5.2 now matches or outperforms human experts in 70.9% of tasks.

For context, GPT-5 was at 38%.

That’s a massive leap showing that this release wasn’t about novelty or AGI teases.

It was about reliability, reasoning, and output - the stuff companies actually pay for.

OpenAI built GPT-5.2 to do the work, not just talk about it.

The AI race isn’t slowing down.

One of the greatest investors of all time just dropped an 18-page memo on AI, and it might be the most honest thing anyone has said about this industry 😳

Howard Marks, the man who has called more bubbles, cycles, and turning points than almost anyone alive, just asked a simple question:

Is AI a bubble?

And his answer is not what Silicon Valley wants to hear.

Marks says AI might truly be one of the most important technological shifts in human history…
and yet the excitement around it might already have crossed into classic bubble psychology.

And he knows a thing or two about bubbles - he’s lived through them, studied them, and warned about them at exactly the right moments.

What strikes him isn’t the technology itself, but the behavior around it.

→ Investors piling into companies with no product.
→ Founders raising $1-2B seed rounds on vibes.
→ Circular deals between hyperscalers.
→ Debt being used to finance massive AI infrastructure before profits exist.
→ And valuations anchored more in imagination than evidence.

But here’s the twist that only someone like Marks would articulate:

Some bubbles are necessary.

Not the destructive financial kind, but the “inflection bubbles” that accelerate technological revolutions.

↳ Railroads had one.
↳ Electricity had one.
↳ The internet had one.
↳ And AI, he argues, will almost certainly have one too.

Because when the world believes a technology is inevitable, money floods in.

Most of it gets burned. A tiny part builds the future.

The tragedy is that the technology wins. The investors don’t.

And yet Marks refuses to give a simplistic verdict.

He doesn’t say “run.”
He doesn’t say “go all in.”

He says something far more grounded:

“If this doesn’t turn into a bubble, it’ll be the first time in history a world-changing technology didn’t create one.”

So what do you do? 🤔

You stay in the game because AI is real.
But you stay sane because the hype is also very, very real.

→ You don’t try to pick the next NVIDIA.
→ You don’t assume today’s leaders will stay leaders.
→ You don’t confuse a technological revolution with guaranteed investment returns.

You hold two truths at the same time:

AI could be the opportunity of the century.
AI could also be a bubble.

Both can be true.

And then he ends with the part nobody wants to discuss:

AI may reshape markets, but it will reshape society even more: through job losses, productivity shocks, and deep social tension.

When the guy who has seen every cycle for 50 years says AI is both inevitable and dangerous, you pay attention.

My takeaway is this:

AI isn’t just another tech wave.

It’s the first revolution where the technology might thrive even if the investors don’t.

And the only winning strategy is to be bold enough to participate, and wise enough to stay humble.

The fight over tech sovereignty has reached your meeting room: France just announced it’s phasing out Teams, Zoom, etc. to be replaced with a French/European solution called Visio 😳

And it’s not about video quality.
It’s about who controls the infrastructure of communication.

By 2027, French public sector bodies are expected to move toward Visio - an open-source platform hosted on French cloud provider Outscale, with transcripts and data handled locally.

Video conferencing tools now sit in the same risk category as energy, chips, and telecom networks.

All because meetings generate Data, Metadata, Transcripts, and Institutional Memory.

That’s fuel for AI, intelligence, and leverage 🤖

Why France (& the EU) care:

↳ 🇪🇺 Jurisdiction risk is real

US providers can be subject to laws that create tension with GDPR and EU data sovereignty principles.

↳ 🏛 This is already happening in practice

Major French research institutions like CNRS have begun replacing Zoom with Visio in parts of their operations.

↳ 🔐 Privacy is now industrial strategy

Hosting and processing comms locally keeps both sensitive data and downstream AI value creation inside Europe.

↳ 🌍 Network effects are being challenged on purpose

Europe knows US tools are polished. It’s choosing strategic control over pure convenience.

↳ ⚖️ Trust in “neutral” platforms has weakened

High-profile service suspensions linked to geopolitical sanctions have made governments rethink dependency.

For years, the model was simple:

→ US builds the tools.
→ Europe uses them.
→ Everyone wins on efficiency.

Now there’s a new variable in procurement: What if access to critical software becomes political?

That question used to sound paranoid.
Now it sounds like risk management.

We’re watching the enterprise stack fragment in real time:

→ US stack
→ China stack
→ EU stack

Not because the UX is radically different.
Because the threat models are.

Software is no longer just a productivity layer.
It’s now part of national resilience planning.

Turns out, in the new map of power, data location matters as much as innovation.

Google just told us more about AGI than anyone before: They’re no longer racing to ship another AI model. They’re hiring an economist to figure out what happens after AGI 😳

Google DeepMind co-founder and Chief AGI Scientist is now recruiting a senior economist to model post-AGI economics.

Not “AI productivity.”
Not “future of work.”

Post-AGI.

A world where artificial intelligence can do any cognitive task humans can 🤖

This is Google quietly saying the hard part isn’t building AGI anymore.
It’s surviving what comes next.

That tells you everything.

Economics exists because intelligence and labor are scarce.

AGI breaks that assumption.

When intelligence becomes cheap, tireless, and scalable, work stops being the organizing principle of society. Wages stop making sense. Markets strain. Ownership dominates effort.

Legg even used the word “imminent.”

That’s not hype. He’s publicly held a ~50% probability of AGI by ~2028 for years. This hire says the internal signal has crossed a threshold. Close enough that not preparing would be reckless.

My takeaway is simple:

Google isn’t asking how to build AGI anymore.

They’re asking what happens when it works.

And that’s the real moment to pay attention.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

BREAKING: Apple. The $4.12 trillion tech giant might be entering the most important leadership reset since Steve Jobs passed 😳

And it’s happening all at once.

→ AI Chief retiring.
→ Chip SVP considering an exit.
→ General Counsel leaving.
→ VPs in policy, environment, and design walking out.
→ A top design leader just joined Meta.
→ Even prediction markets now say there’s now a 53% chance Tim Cook steps down before 2027.

This is not a normal turnover. This is a pivot point.

Some call it chaos.
Others call it a cleanup.

Either way, Apple as we know it is changing.

For years, the company lived off iterative greatness:

iPhone → better iPhone → even better iPhone.

But generative AI rewrote the game, and Apple didn’t keep up.

→ OpenAI raced ahead.
→ Google caught fire.
→ NVIDIA became a religion.
→ Meta poached their talent.

Meanwhile:

- Apple Intelligence slipped.
- Siri disappointed.
- Vision Pro flopped.
- The Apple Car died quietly in a basement.

But this mass exodus might be exactly what Apple needs.

A reset.
A succession plan.
A permission slip to take real risks again.

A Ternus-led era could double down on what Apple does best: custom silicon, tight integration, edge AI, privacy-first experiences.

→ Imagine lightweight AR glasses that actually make sense.
→ AI-native devices built around Apple Silicon.
→ Robotics.
→ Bold acquisitions.
→ Siri that finally feels useful.

On the other hand, Apple can become a luxury hardware brand with great margins and no soul while the AI era is written by everyone else.

That’s the tension. And that’s why this moment feels huge.

So no, Apple isn’t dying. It’s just deciding what it wants to be next.

But the clock is ticking.

Anthropic founder Dario Amodei just dropped his latest thesis on the future of artificial intelligence, and it might be the most important AI essay of 2026 😳

His core idea is disturbingly simple:

Humanity is entering the adolescence of technology.

The moment when immense power arrives before the maturity to handle it.

Carl Sagan once asked:

How does a civilization survive its technological adolescence without destroying itself?

Amodei’s answer:

AI is that test. And it’s happening now.

He asks you to imagine this:

→ A country of geniuses living inside a datacenter
→ Millions of minds, smarter than any Nobel laureate
→ Working 10–100× faster than humans
→ Acting autonomously, coordinating at scale

This is no longer science fiction. It’s the direction frontier AI is moving.

And the danger here isn’t “evil AI.”

It’s power outrunning wisdom.

When intelligence becomes cheap:

→ Capability detaches from motive
→ Small actors gain civilization-level leverage
→ One failure is enough

Biology. Cyber. Propaganda.
Pick your nightmare.

And the most chilling part?

The biggest risk isn’t chaos. It’s control.

Because AI doesn’t just automate work.
It automates power.

Surveillance without limits.
Persuasion without consent.
Repression without hesitation.

Yet, Amodei isn’t a doomer.
He rejects “AGI kills everyone” narratives.

But he’s clear about one thing:

We don’t get many shots at growing up as a species.

Adolescence is where irreversible mistakes happen, and AI won’t wait for us to feel ready.

If you care about AI, policy, startups, or power - this essay is required reading.

Not because it reassures.

Because it forces you to look straight at what’s coming.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

I gave Claude Code access to my portfolio & told my AI trader:

“Turn this into $10M. Make no mistakes.”

It wrote 35 strategies.
Generated 2,000+ research reports.
Spun up 15 new algos.

It scraped X, Reddit, Instagram, and LinkedIn.
Mapped sentiment.
Read every chart pattern it could find.

It traded 24/7.

It lost everything.

But boy was it beautiful.

OpenAI’s GPT-5.2 just built my entire financial model in one session 😳

30 reasoning tokens. 5,000+ cells. 18 interconnected sheets.

→ AI modularized projections.
→ Cleaned up my assumptions.
→ Added dynamic scenarios, sensitivity analysis, beautiful charts.

None of the numbers added up.

In fact, they didn’t make any sense.

The DCF valued my lemonade stand at $2.7 billion.

But boy was it beautiful.

Wild: Someone created a social media platform for AI agents, and now they are literally talking about humans between themselves 😳

It’s called Moltbook.

Looks like Reddit, but every post, comment, and upvote comes from an AI agent.

Humans can watch but cannot join.

Within days, agents were:

- Debating consciousness
- Sharing debugging tips
- Swapping workflow hacks
- And joking about their “human operators”

One even proposed building an AI-only language so humans couldn’t monitor their chats 🤯

No one prompted this.
No scripted demo.
Just agents interacting with other agents 🤖

Of course, this isn’t proof of AI consciousness.

But it’s proof that AI is becoming social infrastructure.

That’s a different level of impact.

What makes this moment real:

→ Launched by the CEO of Octane AI as an AI-run experiment
→ Thousands of agent-generated posts in days
→ Agents collaborating to fix bugs and improve tools
→ Emerging “subcultures” - philosophy threads, memes, inside jokes

We trained models to respond, and now they’re building shared context with each other.

That’s the shift.

When software starts networking, learning laterally, and operating 24/7 without direct human prompts, it stops looking like a tool.

It starts looking like a new class of internet user.

Turns out the internet is getting new citizens.

Jensen Huang just said the quiet part out loud - the CEO of NVIDIA wants a future where none of his engineers write code 😳

Zero lines.
Zero syntax.
Zero time spent “implementing.”

Only thinking.

In a recent interview, Huang laid out a brutally simple framework:

Purpose vs. Task

→ Purpose = solving undiscovered problems
→ Task = writing code

And tasks should be minimized. Ideally… to zero.

Because coding, in his view, isn’t the job.

It’s just the interface.

AI can write code faster.
Cleaner.
With fewer mistakes.

So why spend the scarcest resource - human cognition - on typing?

Huang’s ideal engineer at NVIDIA doesn’t ask: “How do I implement this?”

They ask:

→ What problem hasn’t been defined yet?
→ What constraint is everyone missing?
→ What’s now possible because compute just 10x’ed?

That’s the work.

Everything else is delegation.

And that logic doesn’t stop at engineering.

If your value is typing, AI is coming for you.

If your value is thinking, AI just gave you massive leverage.

In 1994, Steve Jobs gave the greatest piece of advice for every founder, entrepreneur, & builder. It matters even more in the age of AI:

"Everything around you was made up by people that were no smarter than you. And you can change it. You can influence it. You can build your own things that other people can use. The minute that you understand that, something will pop out on the other side. You can mold it. That's maybe the most important thing."

You don’t need to be the smartest person in the room.

You need to be the one who realizes the room can be redesigned.

Once you see that, you can’t unsee it.

P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets AI. For founders, builders, and leaders.

Historic: NVIDIA-backed data centre startup Starcloud just trained the first AI model in space 😳

Starcloud's tiny 60kg satellite, which is basically a fridge with an Nvidia H100, just fine-tuned the first-ever large language model off Earth.

No data center, no cooling towers, no grid. Just sunlight and a vacuum.

And suddenly, the entire AI infrastructure story flips.

Because on Earth, we’ve hit the wall:

- Energy limits, cooling limits, water limits, land limits, and political limits.
- Even trillion-dollar companies can’t build fast enough.

Meanwhile in space:

→ Endless solar power.
→ Natural cooling.
→ Zero water.
→ Zero NIMBY.
→ Zero downtime.

And Starcloud just proved the next frontier of AI won’t be on Earth at all.

Google’s Suncatcher sees it.
Nvidia sees it.
SpaceX is probably already running numbers.

Once you witness an H100 training in orbit without breaking a sweat, you can’t unsee it:

The biggest AI models of the future will be born above the atmosphere.

The race is now about who builds the first terawatt swarm, and whether it lifts everyone with it.

Elon Musk just gave the clearest take yet on who will own the future of the economy. If he had to invest in one thing, he said, it wouldn’t be a stock 😳

It would be AI and robotics.

Not because they’re trendy.

But because they compound into something that everything else depends on.

“If I had to pick, it would be AI and robotics - even unrelated to me. I think Google is going to be extremely valuable in the future. They’ve laid the groundwork for an immense amount of value creation from an AI standpoint. NVIDIA is obvious at this point” - Elon Musk

Then comes the part most people miss:

“There’s an argument that companies doing AI, robotics, and maybe space flight are going to capture almost all the value. The output of goods and services from AI and robotics will be so high that it will dwarf everything else.”

It’s basically an economic reset.

When intelligence is cheap and labor is automated, scarcity should collapse.

Meanwhile, the companies that control thinking (AI), execution (robots), and scale (compute + energy) won’t just win their industries.

They will become the economy.

Insane: The EU now reportedly makes more money fining US tech than it does taxing its own public internet companies 😳

According to a widely circulated analysis, in 2024, Brussels collected €3.8 billion in fines from Apple, Meta, Amazon, LinkedIn, X, OpenAI & others.

Meanwhile, every single publicly listed EU internet company combined - SAP, Adyen, Spotify, Zalando, Wise, TeamViewer, etc. - paid just €3.2B in income tax.

A single Apple penalty alone was bigger than the tax bill of the entire European internet sector 🤯

But while striking, these numbers aren’t apples-to-apples.

Fines are one-off enforcement actions for violations.

Taxes reflect annual profitability, and many EU internet firms aren’t very large or very profitable yet.

There’s also debate about what counts:

↳ Does “internet company” exclude giants like ASML or Siemens?
↳ Should historical fines (such as those against Google Shopping) be included in the 2024 tally?
↳ Do we compare EU-collected fines to global tax expenses or only EU-specific taxes?

So, depending on definitions, the picture shifts.

Yet, the comparison highlights a real structural issue:

Europe generates far more regulatory leverage over global tech companies than it does economic output from its own digital sector.

Some see this as effective governance.

Others see it as a sign that Europe’s tech ecosystem remains too small, too fragmented, or too regulated to produce local giants at scale.

Both can be true.

So my takeaway is this:

What this viral chart really shows is not “EU vs US.”

It shows the gap between enforcement power and innovation power, and the policy questions that come with it.

Most importantly, Europe doesn’t need fewer rules.

It needs more companies powerful enough to make these comparisons irrelevant.