$3,499 for Floating Apps? AI Can Fix That

Welcome back apprentices! 👋

Mount Vesuvius tried to gatekeep ancient wisdom for 2,000 years, but AI just said, “Not on my watch.”

Scientists have digitally unwrapped a Roman scroll for the first time — no peeling, no crumbling, no "Oops, we destroyed history." Just X-rays, machine learning, and some serious patience.

Take that, Nicolas Cage!

So, what’s inside? Deep philosophy? Daily gossip from 79 AD? The first-ever lasagna recipe? AI is still decoding the text, but one thing’s clear: history just got a software update.

In today's email

• Alphabet’s $75 Billion AI Bet

• 500K Students Get AI Study Buddies

• Anthropic Says “No AI on Job Applications”

• Google Revises AI Ethics Policy

• Vision Pro Needs AI — Or It’s Doomed

• More Tools & Updates

_{Read Time: 4 minutes}

Quick News 

💰 Google’s parent company Alphabet is going all in on AI, planning a whopping $75 billion in capital expenditures for 2025 — more than double what it spent in 2023. Why? Because Microsoft, Amazon, and Meta are all scaling up, and Alphabet wants to make sure it’s not left swiping on AI’s future. Meanwhile, Google Cloud revenue jumped 10% in Q4, mostly thanks to AI-powered tools (good news if you love spending money on cloud computing).

🎓 OpenAI just teamed up with the entire California State University system to bring ChatGPT to 500,000 students and faculty — because apparently, AI is the new TA. This is the biggest AI-in-education rollout yet, aiming to help students write better, code faster, and (let’s be honest) probably generate some last-minute essays. 

📝 Anthropic, the company behind Claude AI, is telling job applicants not to use AI when applying for jobs. That’s right — the people building AI assistants don’t want AI-assisted résumés. Why? Because they actually want to see if you can write, think, and (gasp) form original thoughts without AI whispering in your ear.

🔫 Google just quietly removed its ban on AI-powered weapons and surveillance, meaning its “Don’t Be Evil” era is officially on thin ice. The company now joins Meta and OpenAI in allowing certain military AI applications, because, apparently, the future is part chatbot, part battlefield.

Together with Gamma

The future of presentations, powered by AI

Gamma is a modern alternative to slides, powered by AI. Create beautiful and engaging presentations in minutes. Try it free today.

Apple 
Vision Pro Needs AI to Save It — Because $3,499 for Floating Apps Ain’t It 

It’s been a year since Apple launched the Vision Pro, aka a MacBook for your face, and let’s just say… the hype didn’t exactly reach iPhone levels.

💸 Price? A casual $3,499, or roughly the cost of avoiding eye contact with real people forever.
📉 Sales? Around 500k units — not bad, but definitely not Steve Jobs’ ghost would be proud levels.
🤷 Public reaction? Somewhere between “cool but useless” and “why do I look like a sci-fi villain?”

So what’s next? AI. Lots and lots of AI.

If Apple wants this thing to avoid becoming the next iPod Hi-Fi, it needs artificial intelligence to make it actually worth wearing.

The Only Thing That Can Make Vision Pro Cool

Right now, Vision Pro is just an overpriced floating computer screen — and not even a very convenient one. AI could change that by making it actually useful.

AI That Actually Helps You Work 

Vision Pro’s biggest flex is spatial computing, but let’s be honest — it’s just Safari floating in the air.

• Imagine AI organizing your workspace in real-time — dragging your emails, calendar, and Slack into view before you even ask.

• No more hunting for lost files — AI remembers where you left that spreadsheet and automatically pulls it up when you need it.

• A personal AI assistant that learns your workflow — so your Vision Pro doesn’t just show windows, it predicts what you need next.

Right now, you’re paying $3,499 to open Chrome. With AI, you might actually get stuff done.

The One Place Vision Pro Actually Makes Sense 

Guess who’s really loving the Vision Pro? Doctors.

Surgeons are using it to replace physical monitors, so they don’t have to look away mid-surgery (always a plus).
AI-assisted overlays could highlight critical anatomy, detect issues in real time, and even help train new surgeons.
Imagine AI-powered "ghost hands" showing the best approach for a complex procedure — turning the Vision Pro into the ultimate surgical sidekick.

Forget floating TikTok videos — this is the kind of actual life-saving use case Apple should be talking about.

AI That Remembers Everything (So You Don’t Have To) 

What if your Vision Pro could remember everything you see and do — like a hyper-intelligent “undo” button for real life?

• AI-powered memory tracking could log where you saw that important document or summarize that meeting you spaced out during.

• Instead of scrolling through 100 tabs, AI could instantly retrieve past conversations, notes, or project details.

• Think of it as your own personal recall system—but without the creepy Black Mirror vibe (hopefully).

If Apple added AI-powered memory recall, people might actually keep these things strapped to their faces.

So… Where Are The Apple Smart Glasses?

Meta’s Ray-Ban AI glasses are selling like crazy, but Apple? They scrapped their AR glasses project, supposedly because they couldn’t get the battery and processing power right.

Could AI breakthroughs bring Apple back into the smart glasses race? Or are we stuck with bulky ski goggles until 2030?

Right now, Apple is doubling down on Vision Pro, but if Meta cracks the lightweight AI-powered glasses code, Apple will have to respond.

AI Needs to Save Vision Pro — Fast

Right now, Vision Pro is a rich person’s toy — cool, but not essential. AI could change that by making it:

✅ Actually useful for work (instead of just a $3,499 floating desktop).
✅ A real medical tool (because “saving lives” is a better selling point than “watching Netflix in midair”).
✅ A memory powerhouse (so you never forget where you left that important doc).

So will Apple double down on AI and make Vision Pro worth it? Or are we stuck with very expensive floating windows?

Would AI-powered features make you actually want to wear this thing? Or is it still just an overpriced VR headset for Apple superfans?

While Vision Pro is still figuring itself out, Apple just launched its first standalone AI-powered app — Invites, a smart event planner that uses Apple Intelligence to generate custom invitations and manage events seamlessly.

🖼️ AI-generated invites? Yes.
📍 Integrated with Maps, Weather, Photos, and Music? Of course.
📩 Works for non-Apple users? Shockingly, yes.

Unlike Google and OpenAI’s “bigger AI” strategy, Apple is sneaking AI into everyday tools — and if Invites is just the beginning, maybe there’s hope that Vision Pro’s AI glow-up is next.

Tools & Updates 

🔒 Anthropic just dropped "Constitutional Classifiers," a fancy way of saying “Good luck trying to jailbreak our AI.” This new system blocks sneaky prompts that try to bypass safety rules, and after 3,000+ hours of stress-testing, it’s stopping 95% of bad attempts—because AI safety is basically a never-ending game of Whac-A-Mole.

Researchers from TikTok owner ByteDance just dropped OmniHuman-1, an AI that creates hyper-realistic human videos from a single image + motion data — basically, deepfakes on steroids (but for good, hopefully). Need a digital spokesperson who never blinks weirdly? This might be it.

🔍 OpenAI just launched Deep Research, a ChatGPT feature that does full-on web research, analyzes PDFs, and delivers detailed reports with citations in under 30 minutes — aka, AI’s take on “I wrote the essay, bro.” Currently, it’s $200/month for Pro users, but could expand soon. Oh, and it absolutely crushed other AI models in benchmark tests — apparently, Deep Research actually studies.

💼 SoftBank is going all-in on AI, teaming up with OpenAI to launch Cristal Intelligence, a fancy enterprise AI assistant designed to integrate directly into company systems. They’re also dropping $3B a year into AI adoption and rolling out ChatGPT Enterprise across their entire workforce — who needs human meetings when AI can summarize them?

⚡ OpenAI just released o3-mini, a faster, smarter AI model that handles coding, math, and reasoning like a pro — and the best part? It’s free for everyone. It’s like GPT-4o’s smaller, overachieving sibling, built to make AI quicker and more accessible without melting your GPU.

🤖 Google just expanded its Gemini 2.0 lineup, rolling out Pro Experimental (big brain with a 2M token memory), Flash Lite (budget-friendly speedster), and Flash Thinking (now free in the app to show you how it “thinks” in real-time). Multimodal features are coming soon, but early benchmarks aren’t exactly blowing minds—especially with OpenAI’s latest models stealing the spotlight.

Let's Chat About AI

This is all for today!

Replay to this email with your thoughts.

AI is more than just a buzzword. It’s a shift in how we live and work. And understanding it a bit better means you can make smarter choices about the tech you use every day.

_FEEDBACK