- Everyday AI
- Posts
- Ep 711: Coding with OpenAI’s New Codex App: How to Build a Simple App without coding experience
Ep 711: Coding with OpenAI’s New Codex App: How to Build a Simple App without coding experience
ChatGPT Deep Research upgraded to GPT-5.2, Zhipu releases GLM-5, Claude Cowork hits Windows and more
Everyday AI
February 11, 2026
👉 Subscribe Here | 🗣 Hire Us To Speak | 🤝 Partner with Us | 🤖 Grow with GenAI
Outsmart The Future
Today in Everyday AI
8 minute read
🎙 Daily Podcast Episode: OpenAI’s Codex turns simple English into functional desktop apps. Give today’s show a watch/read/listen.
🕵️♂️ Fresh Finds: ChatGPT-5.2 Instant gets upgraded, Warp releases Oz agent, Google Stitch adds Figma layer exports and more Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: ChatGPT Deep Research upgraded to GPT-5.2, Zhipu releases GLM-5, Claude Cowork hits Windows and more Read on for Byte Sized News.
💪 Leverage AI: OpenAI’s Codex just turned plain English into real desktop apps. The dev bottleneck might be gone. Keep reading for that!
↩️ Don’t miss out: Miss our last newsletter? We covered: OpenAI secures big court win, NVIDIA’s $53 billion AI play, ChatGPT ads go live and more. Check it here!
Ep 711: Coding with OpenAI’s New Codex App: How to Build a Simple App without coding experience
You can develop a working Mac app in minutes. 😱
For free.
No coding experience needed.
In the course of one week, OpenAI dropped three big pieces of related news and updates: their new Codex Mac app, a new GPT-5.3 codex model, and the fact it will remain open to free users.
Sheesh. Even if you have zero interest in coding long term, you can start easily building software to help you work better today.
Also on the pod today:
• Codex app demo—live coding 🚀
• GPT-5.3 Codex model unleashed 🤖
• Visual interfaces, not command line 👀
It’ll be worth your 41 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Here’s our favorite AI finds from across the web:
New AI Tool Spotlight – happycapy Lets AI agents learn how you work, instead of you learning how they work. Atyla Monitors brand mentions and measure your visibility on AI search engines, Tines Securely scales AI and automation, Integrates agents, teams, and tools with speed and control.
Google Explicit Search Removal — Google lets you quickly remove explicit images of yourself from Search and block similar ones. Find out how the new tool works.
Google Stitch Upgrade — Stitch by Google now exports editable designs to Figma. See how this speeds up your workflow.
Warp Releases Oz — Warp’s Oz lets you launch cloud agents to build and test code while you’re away. Check out how it works.
Anthropic Sabatoge Risk Report — Anthropic found AI could help with harmful tasks if left unchecked. Learn why governance matters more than ever.
Perplexity Health — Perplexity is adding a Health tab with Apple Health sync and personalized wellness tools. See what’s coming next.
Qwen-Image-2.0 Released — Alibaba’s Qwen-Image-2.0 generates stunning images with perfect text, even for slides and posters.
New Cisco Chips — Cisco unveils its Silicon One G300 chip, promising faster AI networks and big energy savings.
ChatGPT Testing Skills — ChatGPT is testing skill imports, so you can save and reuse prompts like apps. Big workflow upgrade.
ChatGPT-5.2 Instant Upgrade — OpenAI just tweaked GPT-5.2 Instant, promising clearer and more direct answers in ChatGPT and the API.
Facebook Animated Profile Pictures — Animate your profile, restyle your stories—Facebook gets a creative upgrade.
Youtube AI Playlists — YouTube’s AI playlist maker is now Premium-only, letting subscribers build mixes with a prompt.
Krea Nodes — Krea Nodes just launched prompt-to-workflow, letting you build entire node workflows from a simple text prompt.
1. OpenAI Upgrades ChatGPT’s Deep Research with GPT-5.2 🧠
OpenAI has rolled out a major update to its Deep Research feature in ChatGPT, switching to the advanced GPT-5.2 model and introducing new tools like app integration and targeted website searches.
Announced on X, the upgrade lets users track search progress live, ask follow-up questions mid-search, and view results as full-screen reports. Deep Research, which launched in 2025, is now positioned as ChatGPT’s first true "AI agent," taking multi-step web searches into its own hands.
2. China’s Zhipu Unveils Next-Gen AI Model GLM-5 💻
Chinese AI startup Zhipu just dropped its latest flagship model, GLM-5, promising a big leap forward in chat, coding, and agent-like tasks, according to its website.
The release, announced Wednesday, signals that China’s tech sector is pushing hard to keep pace with global AI innovation. Zhipu’s move underscores the fierce competition among AI players worldwide, each racing to set new industry standards.
3. xAI Shake-Up: Two Cofounders Exit Amid Restructuring 🤚
xAI, Elon Musk’s AI startup, just lost two cofounders in less than 48 hours, with Jimmy Ba and Tony Wu both announcing their departures as the company undergoes major internal shifts.
Ba, previously a key leader at xAI, cited “recalibrating his gradient on the big picture” for his exit, as responsibilities shuffle among remaining founders. According to Business Insider, the company has seen six founders leave since its 2023 launch, amid mounting controversies over its AI products and a planned merger with SpaceX.
4. Claude Cowork Lands on Windows, Security Questions Follow 🧑💻
Anthropic’s AI assistant Cowork, fresh from its macOS debut, is now rolling out to Windows in a research preview, giving all paying Claude users access to advanced features like file integration and external service connectors.
The platform promises powerful productivity tools, but its launch has already sparked cybersecurity concerns after reports of successful prompt injection attacks. This move puts Anthropic in direct competition with other desktop AI assistants, raising the stakes in the race for trusted workplace automation.
5. Pentagon Rolls Out ChatGPT for 3 Million Staff 🫂
The Department of War just announced it’s making ChatGPT available to all its personnel, marking a major milestone in government adoption of AI tools.
This move signals the Pentagon’s push to lead with technology, ensuring its teams have secure, custom-built AI for unclassified tasks while keeping sensitive data locked down. The rollout follows GenAI.mil’s rapid growth since December, which has already brought on Google Gemini and xAI’s Grok for over a million users.
6. OpenAI Ads Spark Controversy in New York Times Opinion Piece ⛓️
A high-profile researcher’s resignation at OpenAI made waves this week via a provocative New York Times opinion article, blasting the company’s decision to test ads on ChatGPT.
The author argues this move risks repeating Facebook’s controversial playbook by potentially exploiting user trust and fueling manipulation concerns. The piece calls for stronger oversight and alternative funding models, warning that the company’s current path could undermine both privacy and accessibility.
A Kanban project management feature just got added to a working desktop app.
Time to build it: 90 seconds.
Developer required: none.
(Yeah…..your IT backlog prolly just became irrelevant.)
How? Codex.
Sam Altman announced this week that OpenAI is gonna keep Codex open for free users. That's the most powerful coding model on the planet, now available to anyone who can type a sentence. Hundreds of millions of ChatGPT users just gained the ability to build real, functional software without writing a single line of code.
We're not talking about rinky-dink prototypes that vanish when you close your browser.
Naahhhh.
We're talking actual desktop apps with databases, persistent storage, and functionality that saves between sessions. (Yeah, don’t worry you don’t really gotta understand all that.)
We broke this down on today's Everyday AI and built a working Mac app live, showing exactly how to go from natural language prompt to shipping software faster than most teams can schedule a standup.
Your competitors with expensive dev teams ain't the threat anymore. The ones who figured this out first are.
Time to capitalize shorties.
1. The Dev Bottleneck Just Disappeared 🔥
Think about every custom tool request that's been rotting in your IT backlog for six months.
Now imagine building it yourself during lunch.
That's the shift. OpenAI's Codex takes natural language descriptions and writes the code, structures the files, creates the folder hierarchy, and hands you a working project. You describe what you want like you're explaining it to a colleague. Codex handles the rest.
The kicker for enterprises is brutal. Companies still treating software development as a specialized bottleneck are gonna watch competitors move at speeds that make their sprint cycles look prehistoric.
And the cost of entry just dropped to zero.
One million developers have already used Codex in the past month. That number is about to explode when the rest of the business world catches on.
Try This
Think of one piece of software that would make your workflow easier but doesn't exist yet.
Open Codex and describe it in plain English.
Don't overthink the technical requirements. Just explain what you need it to do like you're talking to a smart intern.
Hit enter and watch what happens. Your Monday morning just got a lot more interesting.
2. Iteration Speed Is The New Moat ⚡
Here's where it gets wild.
Adding a full Kanban view to an existing app took about 90 seconds. That same feature request would've taken a developer days to scope, build, test, and deploy.
The real competitive advantage ain't the first version of anything you build. It's how fast you can keep refining it until it fits your workflow perfectly.
When iterations are basically free, you stop settling for close enough. You start building exactly what you need.
This changes how teams operate. Instead of submitting feature requests and waiting weeks, you can prototype solutions during the meeting where someone identifies the problem.
This is a fundamental shift in how fast organizations can move. Not just some productivity hack.
Try This
Take something you've already built and immediately ask Codex to add one improvement.
A different view. A timer function. A new data field.
Watch how fast it updates.
Train yourself to think in rapid iterations rather than big bang releases. The muscle memory of "ask and receive" will change how you approach every workflow problem.
3. Custom Tools Beat Bloated Subscriptions 🚀
Every company is out there paying for CRMs and project management tools they only partially use.
Meanwhile, the exact functionality they actually need could be built in an afternoon.
The math is straightforward. If your team is paying $50 per seat per month for software where you touch three features, you might be lighting money on fire. Custom tools built for your specific workflow will always outperform generic solutions designed for everyone.
And here's the thing. You don't need authorization, databases, or backend infrastructure for most internal tools. Local storage on your machine handles it. No login required. No IT tickets. No waiting.
The barrier between "software we wish existed" and "software we use" just collapsed.
Try This
Audit one software subscription this week.
Write down the three features you actually use. Ask yourself if Codex could build something that does exactly those things.
If yes, schedule 30 minutes to prototype it.
Before Deborah in accounting retires and takes all the institutional knowledge about why you bought that tool in the first place.
Reply