• Everyday AI
  • Posts
  • Ep 742: OpenAI's Most Chaotic Week Yet? Sora's Dead, New Models are Coming and What it Means for You

Ep 742: OpenAI's Most Chaotic Week Yet? Sora's Dead, New Models are Coming and What it Means for You

Meta Releases TRIBE v2 Brain-Predicting Model, Google Rolls Out Lyria 3 Pro in Gemini, Meta Cuts Staff to Fund AI Push, and more.

Author

Everyday AI
March 26, 2026

 

Outsmart The Future

Today in Everyday AI
8 minute read

🎙 Daily Podcast Episode: From shutting down Sora to building a super app, OpenAI is making major shifts. Here’s the breakdown. Give today’s show a watch/read/listen to learn more.

🕵️‍♂️ Fresh Finds: Gemini 3.1 Flash-Live Preview Nears Release, Meta Cuts Jobs to Fund AI, U.S. Claude Releases new tools for Mobile, and more. Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: Meta Drops TRIBE v2 Brain Model, Google Expands Gemini Music Tools with Lyria 3 Pro, TurboQuant Shakes AI Memory Market, and more. Read on for Byte Sized News.

💪 Leverage AI: OpenAI didn’t have a bad week—it made a calculated reset. And most people misread it. Keep reading for that!

↩️ Don’t miss out: Miss our last newsletter? We covered: OpenAI's next model could accelerate the economy, Judge questions Pentagon's fight against Anthropic, OpenAI kills Sora and more. Check it here!

Ep 742: OpenAI's Most Chaotic Week Yet? Sora's Dead, New Models are Coming and What it Means for You


Has OpenAI ever had a more chaotic week than the past 7ish days? 🤪

They're prepping for an IPO.

They're building a super app.

They might get sued by their biggest investor.

They're planting a new HQ in Google's backyard.

And they killed one of their most viral products in Sora.

Crazier part?

That's not even the half of it.

And while it's sometimes a waste of time to keep up with the see-sawing of big AI companies, this is different.

However OpenAI ultimately comes out of this week may just set the precedent for the next few years of AI.

Also on the pod today:

• Sora app: shut down fast 📽️
• $1B Disney deal vanished 🏰 
• Microsoft vs OpenAI lawsuit ⚖️

It’ll be worth your 35 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 – PIO Simplifies your global expansion with our all-in-one platform, StreamPad Transforms your iPhone/iPad into a fully customizable control deck, Venn.ai Puts AI to work inside your apps without giving up control.

Meta Job Cuts — Meta is cutting jobs as it shifts funding away from VR and into AI, with Reality Labs among the areas hit. Curious?

Claude Mobile Tools — Claude put Figma, Canva and Amplitude in its mobile app. Ready to work from your phone?

NotebookLM Whiteboard Videos — NotebookLM may be testing whiteboard-style videos as a new output format, which could boost retention for explainer content.

Meta Shopping — Meta is testing AI-powered pop-ups that summarize product reviews, brand info, and deals right in Facebook and Instagram.

Gemini 3.1 Flash-Live Preview Soon — Google is preparing to roll out Gemini 3.1 Flash-Live Preview, likely coming to Vertex AI and AI Studio soon.

Melania Trump Showcasing AI Robot — Melania Trump showcased an AI robot at a White House summit on educational tech, sparking debate about robots versus real teachers.

 

OpenAI Safety Bug Bounty Program — OpenAI now pays researchers to find safety and abuse risks in its AI products. Want to hunt agent hijacks, data leaks, or platform integrity flaws?

ARC-AGI-3 Released — ARC-AGI-3 is a benchmark that tests whether AI can learn and adapt inside novel environments, measuring intelligence across time instead of final answers.

Tinker Released — Shopify launched Tinker, a free mobile app that bundles 100+ guided AI tools for fast brand assets like product photos, logos, and videos.

Report: AI Dominating the Web — AI-written content now dominates web traffic, with studies finding bots and automated tools generating a large share of pages and interactions online.

AI Bill — Sen. Adam Schiff plans a bill to legally bar AI from being used in fully autonomous weapons and mass domestic surveillance, mirroring Anthropic’s safety limits.

Cohere Transcribe — Cohere released Transcribe, an open-weight Conformer ASR that tops the HuggingFace Open ASR leaderboard with a 5.42% WER.

OpenAI Model Spec — OpenAI’s Model Spec is a public playbook that spells out how models should behave, balancing user freedom, safety, and transparency.

Granola Fundraise — Granola raised $125M at a $1.5B valuation as it pivots from a prosumer meeting-transcription app to enterprise workflows and new note/context APIs.

1. Meta unveils TRIBE v2, a brain-predicting foundation model 🧠

Today Meta released TRIBE v2, a trimodal foundation model trained on 500+ hours of fMRI from over 700 people that can predict human brain responses to sight and sound without retraining.

The model reportedly delivers a 2–3x improvement over prior methods for movies and audiobooks and generalizes to new subjects, languages, and tasks, making it timely for researchers seeking scalable neural proxies. Meta is also releasing the paper, model, code, and a demo to accelerate open research and potential clinical and AI applications.

2. Gemini upgrades music tools with Lyria 3 Pro rollout 🎼

Google announced today that Lyria 3 Pro is rolling out to AI Plus, Pro, and Ultra users in the Gemini app and web, giving creators more customization and longer-track options for music generation.

The update lets users choose “Thinking” or “Pro” models in the Create Music tool, aiming to add detail and length to compositions while competing with niche rivals on vocal realism. Early reactions praise the sound quality but note some competitors still lead on vocal timbre and naturalness, keeping the field competitive.

3. Google’s TurboQuant rattles memory markets 💻

Google on Tuesday unveiled TurboQuant, a compression technique that it says can cut the memory needed to run large language models by six times, and markets reacted immediately as memory-chip stocks slid.

Investors fear a potential slowdown in demand for AI memory chips after a strong multi-year rally that had already pushed prices and profits to record levels. Industry analysts caution that efficiency gains could make models more capable and ultimately increase hardware needs, so the breakthrough is likely evolutionary rather than catastrophic for long-term demand.

4. Progressive lawmakers push short-term moratorium on U.S. data centers ⚖️

Rep. Alexandria Ocasio-Cortez and Sen. Bernie Sanders introduced legislation this week to pause new data center construction until national safeguards protect workers, consumers, and the environment, highlighting urgent concerns about AI’s growing energy demands and local impacts.

The proposal is unlikely to pass but underscores rising public resistance to data centers tied to higher electricity costs, water use, and pollution as AI expands. The push adds pressure on policymakers already balancing national competitiveness in AI with utility strain, even as the White House presses tech firms to self-fund new power and infrastructure.

5. Anthropic flags widening AI skills gap as adoption grows 🤖

A new Anthropic report, discussed by head of economics Peter McCrory at the Axios AI Summit, finds little current job displacement from AI but warns that rapid adoption could change that quickly, especially for entry-level white-collar roles.

The company says early Claude users pull ahead by using the model for core, iterative work while newcomers lag, creating an emerging skills divide that is stronger in high-income regions. Anthropic stresses the need for monitoring frameworks to catch displacement early and guide policy as AI spreads across industries.

OpenAI made more moves in seven days than most companies make in a year, and almost nobody is reading it right.

The narrative online? OpenAI is crumbling. Sora died. Shopping flopped. Microsoft might sue.

Nah.

(What you're watching is a company ruthlessly clearing the deck before what could be the biggest tech IPO since Google. And your AI strategy prolly needs to account for what comes next.)

We broke it all down on today's Everyday AI. The platform wars just entered the enterprise lock-in phase, and the window to pick your ecosystem on your terms is closing. Here's what that means for you.

1. Killing Side Quests Is Actually Smart 🔥

OpenAI didn't fail. It fired its failures.

Sora, the AI video app, shut down roughly six months after launch after downloads cratered nearly 75% from peak. The billion-dollar Disney character licensing deal died with it. And the ChatGPT shopping checkout? Gone too. Walmart saw a three times lower conversion rate inside ChatGPT than on its own site, so OpenAI pulled the plug and pivoted to product discovery instead.

None of this is weakness. OpenAI applications chief Fidji Simo told employees straight up: no more "side quests." Competing with Anthropic's laser focus on enterprise and coding means OpenAI has to stop being everything to everyone and start being essential to the RIGHT someone.

Strategic retreat is not collapse. It's surgery.

The companies that win the AI race won't be the ones who shipped the most. They'll be the ones who stayed ruthlessly focused on the tools that actually move the needle for their customers.

Try This

Pull up your company's AI tool spend this week and do what OpenAI just did to itself.

For every tool on that list, ask one question: is this attached to a measurable business outcome, or did someone get excited at a conference and expense it?

If the answer is "vibes," that tool needs a 30-day kill-or-keep decision. Assign an owner. Set a metric. Either it earns its seat at the table or it gets cut. Your AI stack should be leaner than OpenAI's product roadmap.

2. One App to Rule Your Workflow ⚡

OpenAI confirmed it's merging ChatGPT, Codex, and the Atlas browser into a single desktop super app. And here's what most people are sleeping on: Codex isn't JUST a coding tool. It's a serious knowledge work platform that most enterprise teams have completely ignored. When you fold that into ChatGPT's conversational layer and Atlas's browser context under one roof, you've got an agentic system that can research, write code, and analyze data without you touching a single app switcher.

Anthropic already figured this out. Claude chat, Claude Code, and Claude Cowork live in one app, and that unified experience helped Anthropic capture 73% of first-time enterprise AI spending in early 2026. OpenAI watched that and said, yep, we're doing that too, except with 900 million weekly users already in the ecosystem.

Try This

Most teams are treating ChatGPT like a fancy search engine.

Before the super app drops, open Codex this week and give it a real knowledge work assignment, something your team actually does. A competitor breakdown. A workflow audit. A document structure. Not a demo. Something live.

You want your team already fluent with these tools before the unified app arrives and your competitors start locking in enterprise contracts. That fluency costs you nothing but a couple of hours this week.

3. The Platform Wars Just Entered Lock-In 🚀

OpenAI just closed an additional $10 billion, bringing their total funding round to north of $120 billion at roughly an $850 billion valuation. CFO Sarah Friar said plainly that she needs to make the company "healthy and ready to face the public markets." That's not a crisis statement. That's a pre-IPO roadmap.

They hired Dave Dugan, former VP of global clients and agencies at Meta, to build out the ChatGPT ads business. With 900 million users and over 800 million on the free plan, that's a data and attention footprint that's gonna print once the ad machine is properly tuned.

And the Microsoft potential lawsuit over the $50 billion Amazon cloud deal? Worth watching. The models powering your Microsoft 365 Copilot could change depending on how that shakes out.

Try This

Your AI vendor relationships are about to get as sticky as your CRM.

OpenAI is locking in enterprise contracts before the IPO, and pricing flexibility will only shrink once they're public. If your org is still in "we're evaluating AI tools" mode, you're about to get priced out of the conversation.

Talk to your procurement team this week. Map out what AI platforms you're committed to, what contracts are up for renewal, and where you've got leverage. That leverage is worth real money right now. It won't be after the IPO.

Reply

or to participate.