OpenAI Just Killed Sora. Here's What No One Is Saying

Introduction

Anthropic just gave Claude the ability to control your Mac. We are talking clicking buttons, opening applications, executing tasks while you step away from your desk. That is AI moving from answering questions to actually doing the work. Plus, OpenAI just killed Sora after spending millions on development. And the AI skills gap is creating real winners and losers in your workforce right now.

Welcome to your weekly AI news roundup. Every week I condense the biggest AI stories into what actually matters for your business. No jargon, no hype, just what you need to know to stay ahead. Let’s get into it.

OpenAI Just Killed Sora

First up, OpenAI is shutting down Sora. The AI video generation tool launched with massive hype six months ago. They are killing the standalone app, the API, and the entire platform. It hit a million downloads in its first week. Disney pledged a billion dollars for content partnerships. But it never found real traction for actual business needs.

It was expensive, plagued by deepfake and copyright issues, and most businesses could not see clear ROI on it. The honest answer is Sora looked good in headlines but failed to solve real business problems. OpenAI is shifting focus to enterprise solutions and robotics research instead.

The lesson here is straightforward. Flashy consumer AI apps often fail commercially. If you are evaluating AI tools for your business, focus on enterprise options with clear ROI rather than experimental consumer apps that might disappear overnight. When a vendor pivots away from a product, you lose access. Your workflows break. Your data might be stuck. That costs you money.

Anthropic Just Gave Claude Control of Your Mac

Now this one is significant. Anthropic has launched the most ambitious consumer AI agent to date. They have given Claude the ability to directly control your Mac. Clicking buttons, opening applications, typing into fields, navigating software on your behalf while you step away. This is available as a research preview for paying subscribers right now through the Claude desktop app and Claude Code.

Here is the thing. This transforms Claude from a conversational assistant into something closer to a remote digital operator. It can do the repetitive tasks that eat up your day. Think about what that actually means for your business. An AI that can open your CRM, pull up customer records, draft emails, update spreadsheets, all without you sitting at the keyboard.

The race to build AI agents that actually do work is accelerating. OpenAI, Anthropic, Google, they are all moving fast. For your business, AI is moving from answering questions to executing tasks. If you have processes that follow clear rules, an agent could handle them today.

But and this is crucial, you need to think about security. Giving an AI tool control over your computer means it has access to your data, your accounts, everything on your machine. Strong permissions, clear boundaries on what the agent can and cannot do. Do not hand it full access on day one. Start with low-risk tasks, watch how it behaves, then expand if it works.

The AI Skills Gap Is Here Now

This next story is a wake-up call for business owners. Anthropic released their fifth economic impact report. The AI skills gap is here now, not in some distant future. Earlier adopters of AI tools are pulling further ahead as they develop deeper competency while new users struggle to catch up.

Here is the number that matters. Anthropic CEO Dario Amodei estimates AI could eliminate half of all entry-level white-collar jobs within five years. For your business, this is not about replacing your staff. It is about the gap between businesses training their teams to use AI effectively and those that are not. That gap is widening fast.

The research shows tasks that can be auto-tested, like bug-fixing and maths-heavy work, advance faster with AI. This creates a two-tier workforce. One tier develops AI competency and becomes more productive. The other tier gets left behind.

If you have not started experimenting with AI in your business yet, your competitors probably have. Start small, low risk, but start. Spend two hours this month getting your team to use Claude or ChatGPT on something real. See what works. See where it breaks. Build muscle memory before you need it.

Meta and Google Found Negligent in Landmark Trial

Quick hits now. A Los Angeles jury found Meta and Google negligent in a landmark social media addiction trial. Six million dollars in damages. Three million compensatory, three million punitive. The jury was nearly unanimous at ten to two. Meta paid 70 per cent, Google 30 per cent.

What makes this particularly significant is the jury saw internal Meta documents showing the company knew its platforms were addictive for teenagers and used those findings to boost engagement anyway. Snapchat and TikTok settled before trial. Then the day after this verdict, a New Mexico jury ordered Meta to pay 375 million dollars for Instagram predator issues.

Why this matters for your business. If you are building any customer-facing app or digital product, the design choices you make today could become legal liabilities tomorrow. Ethical UX is not optional. It is not something you bolt on at the end. If your product is designed to exploit user psychology, you are building a legal landmine. That will cost you.

AI Regulation Is Coming

Bernie Sanders and AOC proposed a nationwide ban on new AI data centres until Congress passes comprehensive AI regulation. More than 230 organisations called for a federal moratorium. Over a dozen states have pending statewide legislation.

Here is the honest assessment. This faces significant congressional hurdles and is not going to affect your ability to use AI tools any time soon. But it tells you something important. AI regulation is coming. It is not a maybe. Regulation is moving from discussion to action.

Plan accordingly. Do not assume the regulatory landscape of 2025 will be the regulatory landscape of 2027. Start thinking about data governance, model transparency, and compliance frameworks now. Even if these specific proposals fail, others will pass.

Waymo Robotaxis Need Firefighters and Police to Bail Them Out

Waymo robotaxis are relying on firefighters and police to bail them out during emergencies. TechCrunch documented six instances where first responders had to physically move Waymo vehicles. In Austin alone, robotaxis illegally passed stopped school buses nineteen times. The NTSB opened investigations. A CHP officer had to drive a robotaxi to a park-and-ride. Incidents across Austin, Atlanta, Nashville, and on the highway.

The lesson remains the same. There is a meaningful difference between AI that works in normal conditions and AI that works in all conditions. Autonomous vehicles perform well on clear days with visible road markings. They struggle in rain, snow, complex intersections, and unusual situations.

Keep humans in the loop for anything critical. If your business uses any autonomous system, do not assume it works perfectly. It works well most of the time. That is different from all of the time.

Meta Launches AI Tools for Small Businesses

Meta launched Meta Small Business, an initiative to build AI-powered tools across Facebook, Instagram, and WhatsApp for small businesses. Tools in the pipeline include an AI business assistant for ad optimisation and WhatsApp business AIs that handle customer enquiries.

They tested the WhatsApp business AI in Mexico and the Philippines with over a million weekly conversations. The AI business assistant was tested in Q4 2025. If you are already spending money on Meta ads, worth monitoring as these tools roll out.

The opportunity here is automation of customer support on WhatsApp, which could reduce your support team workload. But test it carefully. These are early tools. Do not assume they will work perfectly out of the box.

Agentic Commerce Is Here

MIT Technology Review published a piece on agentic commerce. The concept is simple. Instead of AI giving you a list of links, an AI agent actually executes purchases for you. Book a trip, compare suppliers, handle procurement.

Amazon’s Rufus AI is already shopping alongside users. Google’s AI Mode enables agentic checkout. The adoption curve is steep. Traditional AI took eight years to reach 72 per cent adoption. Agentic AI has hit 35 per cent in just two years. Fifty eight per cent of businesses now are ecosystem-driven compared to just twelve per cent in 2013.

If you run retail or e-commerce, start thinking about whether your product data is clean enough for an AI agent to work with. That means consistent product descriptions, accurate pricing, clear inventory. If your data is messy, an AI agent will surface that mess to customers. If your data is clean, an agent makes you more discoverable and easier to buy from.

Spotify Fights AI-Generated Content Fraud

Spotify is testing a new tool called Artist Profile Protection to stop AI-generated tracks from being attributed to real artists. Sony Music Group alone flagged over 75,000 fraudulent tracks last year.

If your business relies on creative output, expect more tools like this across all platforms. Platforms are starting to take AI content quality seriously. If you are using AI to generate content for your business, document it. Be transparent about it. This is the direction all platforms are moving. What is acceptable today might not be acceptable in two years.

One to Watch: Talat

Talat is a new Mac app that does AI meeting notes but keeps everything on your machine, not in the cloud. It captures audio from Zoom, Teams, or Meet calls, transcribes in real time, and generates summaries with key points and action items. All of it happens locally. No subscription, just a one-time fee.

It was created by a developer seeking an alternative to Granola, which is valued at 1.5 billion dollars. Talat supports Ollama for local AI models and exports directly to Obsidian for your notes system.

For your business, if you handle sensitive client meetings or financial conversations, the local-first approach means your data stays under your control. That matters for GDPR compliance. Your meeting notes are not leaving your machine. That is worth something.

This is exactly the kind of practical, privacy-conscious AI tool worth checking out. It is not flashy, but it solves a real problem without surveillance.

Key Takeaways

• OpenAI killing Sora proves flashy consumer AI apps often fail commercially. Focus on enterprise tools with clear ROI.
• AI agents that actually do work are here. Start with low-risk tasks and strong security permissions.
• The AI skills gap is widening now. Get your team experimenting with AI this month.
• Ethical UX is becoming a legal requirement. Design choices today become liabilities tomorrow.
• AI regulation is coming. Start thinking about data governance and compliance now.
• Keep humans in the loop for critical systems. AI works well most of the time, not all of the time.
• Clean your product data now if you want to benefit from agentic commerce.

Over to You

So the distilled takeaway this week. AI is moving from answering questions to executing tasks. The skills gap is widening. Start simple, keep humans in the loop, know where your data goes. If you are not experimenting yet, start this month.

If that was useful, subscribe because I do this every week. And if you want to know where AI fits in your specific business, I offer AI audits where I review your operations and give you a clear roadmap. Link is in the description. See you next week.