Currently viewing the human version
Switch to AI version

Altman Flip-Flopped on Smart Glasses in Six Months

Smart glasses prototypes

Sam Altman literally said "absolutely not, I don't like smart glasses" in July and now OpenAI is apparently hiring Apple hardware engineers to build them. Companies pivot this fast when they realize they're about to get left behind.

Meta Already Won This Fight

Meta Ray-Ban smart glasses

While everyone was making fun of the metaverse, Meta actually figured out smart glasses. I'm not saying they're amazing, but they work and people actually buy them, which is more than anyone else managed.

Meta's Ray-Ban glasses succeeded because they didn't try to be science fiction. No AR overlays that drain your battery in 20 minutes, no floating keyboards that make you look insane. Just AI that helps without announcing itself.

Compare that to Apple Vision Pro at $3,500 that weighs like a brick, or Google Glass that turned everyone into walking privacy violations. Meta's glasses cost $300 and look normal.

OpenAI was too busy building "screenless, pocket-sized devices" that solve problems nobody has.

Why the Hell Did They Pick Jony Ive?

OpenAI partnered with Jony Ive's LoveFrom studio, which sounds fancy until you remember Ive designed those butterfly keyboards that broke constantly. Those things were a disaster.

Ive's whole thing is "simplicity" and "user experience," which in practice means removing useful ports and calling it innovation. Now he's supposed to design OpenAI's entire hardware ecosystem: glasses, speakers, pins, whatever else they're cooking up.

Apple succeeded by perfecting one product at a time. OpenAI wants to build everything simultaneously while their main product is still basically smart autocomplete.

Hardware Is Actually Hard

OpenAI's hiring Apple hardware people, which makes sense, but hiring talent doesn't fix physics. You can't push a software update to make batteries last longer or fix overheating chips.

Meta spent years and billions learning how to make smart glasses. Their first generation was trash, so they iterated and eventually got something decent. OpenAI thinks they can skip the failure part while building glasses, speakers, and pocket computers simultaneously.

I've covered enough hardware launches to know this doesn't work. Ship broken software and you patch it overnight. Ship broken hardware and you're explaining to investors why you have warehouses full of $500 paperweights.

The Apple Lock-Out Risk

Apple's building their own smart glasses, and when they launch, ChatGPT integration will disappear faster than features in a Jony Ive redesign. Apple doesn't do permanent partnerships – they do temporary arrangements until their own products are ready.

Google's doing the same thing. Meta's building their own LLMs. Every major tech company is racing toward vertical integration, and OpenAI's window for third-party AI integration is closing fast.

Without their own hardware platform, OpenAI becomes a software vendor dependent on platform owners who are also competitors. That's not a sustainable business model – that's a recipe for getting Sherlocked.

The Real Infrastructure Play

With Nvidia's $100 billion investment and Oracle's $300 billion cloud deal, OpenAI has the compute infrastructure to power edge AI devices. The question is whether they can build hardware that doesn't suck.

Smart glasses require three things to work: compelling AI features, comfortable hardware, and decent battery life. Meta nailed this combination through years of iteration. OpenAI has to get it right on the first try while competing with established hardware companies.

The physics haven't changed. Cramming AI processing, cameras, microphones, speakers, and batteries into something you can wear for hours without headaches is still really fucking hard.

So What Now?

Altman's flip-flop tells you everything about OpenAI's position. They built their lead with software, but software advantages disappear overnight. Hardware control lasts forever.

The next platform isn't phones - it's AI that sees what you see and hears what you hear. If OpenAI doesn't control that interface, they become a service provider to their competitors.

OpenAI's hardware isn't shipping until 2027. Apple, Google, and Meta will all have smart glasses out before then. In consumer electronics, showing up late is usually game over.

This Probably Won't Work

OpenAI went from "absolutely not" to panic development in six months. When your strategy changes that fast, it's not strategy.

The real question is whether they can build decent hardware before everyone else locks them out. Based on their hardware experience (zero) and Jony Ive's recent track record (butterfly keyboards), I'm skeptical.

But hey, maybe they'll surprise everyone. Or maybe they'll learn the hard way that hardware is different from software, and physics doesn't care about your valuation.

Where OpenAI Fits in the Smart Glasses Mess

Company

Status

What We Know

OpenAI

Maybe 2026-2027

Partnered with Jony Ive. No prototype yet. Altman flip-flopped on this

Meta

Actually shipping

Ray-Ban glasses work okay. $300-400. People actually buy them

Apple

In development

Probably launching before OpenAI. Will cost too much

Google

Who knows

Google Glass was a disaster. Maybe they learned something

Microsoft

Enterprise only

HoloLens costs $3000 and weighs like a brick

Amazon

Barely exists

Echo Frames are just Alexa in glasses. Boring

Smart Glasses Deep Dive: Essential Reading

Related Tools & Recommendations

compare
Recommended

I Tested 4 AI Coding Tools So You Don't Have To

Here's what actually works and what broke my workflow

Cursor
/compare/cursor/github-copilot/claude-code/windsurf/codeium/comprehensive-ai-coding-assistant-comparison
100%
alternatives
Recommended

GitHub Actions is Fucking Slow: Alternatives That Actually Work

integrates with GitHub Actions

GitHub Actions
/alternatives/github-actions/performance-optimized-alternatives
58%
compare
Recommended

AI Coding Assistants 2025 Pricing Breakdown - What You'll Actually Pay

GitHub Copilot vs Cursor vs Claude Code vs Tabnine vs Amazon Q Developer: The Real Cost Analysis

GitHub Copilot
/compare/github-copilot/cursor/claude-code/tabnine/amazon-q-developer/ai-coding-assistants-2025-pricing-breakdown
46%
integration
Recommended

Slack-Jira 연동 삽질기

integrates with Slack

Slack
/ko:integration/slack-jira/setup-implementation-guide
46%
news
Similar content

Google Finally Admits to the nano-banana Stunt

That viral AI image editor was Google all along - surprise, surprise

Technology News Aggregation
/news/2025-08-26/google-gemini-nano-banana-reveal
45%
news
Recommended

Microsoft Added AI Debugging to Visual Studio Because Developers Are Tired of Stack Overflow

Copilot Can Now Debug Your Shitty .NET Code (When It Works)

General Technology News
/news/2025-08-24/microsoft-copilot-debug-features
39%
tool
Recommended

GitHub CLI Enterprise Chaos - When Your Deploy Script Becomes Your Boss

depends on GitHub CLI

GitHub CLI
/brainrot:tool/github-cli/enterprise-automation
37%
news
Similar content

This Nvidia-OpenAI Deal Just Broke My Brain

By Sarah Chen, Senior Technology CorrespondentSeptember 22, 2025 | 11:47 PM PST

Google Chrome
/news/2025-09-22/nvidia-100b-openai-investment
37%
news
Recommended

OpenAI Bought Statsig for $1.1B Because Rolling Out ChatGPT Features Is a Shitshow

built on Microsoft Copilot

Microsoft Copilot
/news/2025-09-06/openai-statsig-acquisition
32%
tool
Recommended

Azure OpenAI Service - OpenAI Models Wrapped in Microsoft Bureaucracy

You need GPT-4 but your company requires SOC 2 compliance. Welcome to Azure OpenAI hell.

Azure OpenAI Service
/tool/azure-openai-service/overview
32%
alternatives
Recommended

OpenAI API Alternatives That Don't Suck at Your Actual Job

Tired of OpenAI giving you generic bullshit when you need medical accuracy, GDPR compliance, or code that actually compiles?

OpenAI API
/alternatives/openai-api/specialized-industry-alternatives
32%
news
Similar content

OpenAI Finally Shows Up in India After Cashing in on 100M+ Users There

OpenAI's India expansion is about cheap engineering talent and avoiding regulatory headaches, not just market growth.

GitHub Copilot
/news/2025-08-22/openai-india-expansion
31%
review
Recommended

Cursor AI 솔직 후기 - 한국 개발자가 한 8개월? 9개월? 쨌든 꽤 오래 써본 진짜 이야기

VS Code에 AI를 붙인 게 이렇게 혁신적일 줄이야... 근데 가격 정책은 진짜 개빡친다

Cursor
/ko:review/cursor/honest-korean-dev-review
27%
tool
Recommended

Cursor - VS Code with AI that doesn't suck

It's basically VS Code with actually smart AI baked in. Works pretty well if you write code for a living.

Cursor
/tool/cursor/overview
27%
review
Recommended

Codeium Review: Does Free AI Code Completion Actually Work?

Real developer experience after 8 months: the good, the frustrating, and why I'm still using it

Codeium (now part of Windsurf)
/review/codeium/comprehensive-evaluation
27%
compare
Recommended

Enterprise AI Coding Tools: Which One Won't Get You Fired?

GitHub Copilot vs Cursor vs Claude Code vs Tabnine vs Windsurf - The Brutal Reality

GitHub Copilot Enterprise
/compare/github-copilot/cursor/claude-code/tabnine/codeium/enterprise-ai-coding-security-comparison
27%
tool
Recommended

Fix Tabnine Enterprise Deployment Issues - Real Solutions That Actually Work

competes with Tabnine

Tabnine
/tool/tabnine/deployment-troubleshooting
27%
compare
Recommended

GitHub Copilot vs Tabnine vs Cursor - Welcher AI-Scheiß funktioniert wirklich?

Drei AI-Coding-Tools nach 6 Monaten Realitätschecks - und warum ich fast wieder zu Vim gewechselt bin

GitHub Copilot
/de:compare/github-copilot/tabnine/cursor/entwickler-realitaetscheck
27%
alternatives
Recommended

ChatGPT真的用不下去了,这几个替代挺好用

competes with ChatGPT

ChatGPT
/zh:alternatives/chatgpt/intelligent-ai-alternatives
27%
compare
Recommended

Which AI Actually Helps You Code (And Which Ones Waste Your Time)

competes with Claude

Claude
/compare/chatgpt/claude/gemini/coding-capabilities-comparison
27%

Recommendations combine user behavior, content similarity, research intelligence, and SEO optimization