- Everyday AI
- Posts
- ChatGPT’s Agent Mode Overview: 5 things you should know
ChatGPT’s Agent Mode Overview: 5 things you should know
White House targets “woke AI” in executive order, Meta rejects EU’s AI Code of Practice, Claude Code users hit by usage limits and more.
Everyday AI
July 18, 2025
👉 Subscribe Here | 🗣 Hire Us To Speak | 🤝 Partner with Us | 🤖 Grow with GenAI
Outsmart The Future
Today in Everyday AI
6 minute read
🎙 Daily Podcast Episode: OpenAI just dropped ChatGPT’s powerful new Agent Mode—combining Operator, Deep Research, and ChatGPT into the ultimate AI agent. Learn what ChatGPT Agent Mode is and 5 must-know features to boost your workflow. Give it a listen.
🕵️♂️ Fresh Finds: Amazon gas emissions increase for AI demand, Netflix uses AI for visual effects and Former Google CEO speaks on superintelligence. Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: White House targets “woke AI” in executive order, Meta rejects EU’s AI Code of Practice and Claude Code users hit by usage limits. For that and more, read on for Byte Sized News.
🧠 Learn & Leveraging AI: OpenAI’s ChatGPT Agent is here! Wondering what’s new? Here’s 5 things you need to know. Keep reading for that!
↩️ Don’t miss out: Did you miss our last newsletter? We talked about OpenAI's ChatGPT Agent being unveiled, Mistral’s Le Chat getting deep research, Anthropic getting sued for copyright and more. Check it here!
ChatGPT’s Agent Mode Overview: 5 things you should know 💡
ChatGPT Agent is here!
↳ What the heck is it?
↳ How does it work?
↳ What do you need to know?
Glad you asked, shorties. We have the answers.
Also on the pod today:
• ChatGPT Agent Virtual Computer Functions 🖥
• Spreadsheet and PowerPoint Generation in ChatGPT 📂
• Operator vs. Deep Research Capabilities Explained 🤔️
It’ll be worth your 33 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 – Epiphany turns your voice notes into actions, RecordAI is a WhatsApp friend that uses AI to remind you about tasks, Notebooks turns your content into a central hub for future content production.
Amazon – Amazon’s greenhouse gas emissions have increased over the past year as it expands data centers for AI.
AI in Media – Netflix says it has used AI visual effects for the first time to cut costs.
AI Search – Browser DuckDuckGo now lets you hide AI-generated images in search results.
Trending in AI – Former Google CEO says superintelligence is tech’s holy grail, not AI, but the U.S. isn’t equipped to address its limit.
Future of Work – Mike Rowe believes AI will target white collar workforce before it comes for blue-collar jobs.
1. White House Targets ‘Woke AI’ in New Executive Order 🏛
The White House is preparing an executive order demanding political neutrality from AI models used by companies with federal contracts, aiming to curb perceived “woke” bias. This move comes amid growing concerns over AI hallucinations and controversial outputs.
The order highlights ongoing tensions between government efforts to regulate AI and tech industry fears of favoritism, especially as major players like OpenAI, Google, and xAI secure Pentagon contracts.
2. Meta Rejects EU AI Code of Practice Amid Regulatory Clash 🚫️
Just weeks before the EU’s AI Act rules kick in, Meta has declined to sign the bloc’s new voluntary Code of Practice for general-purpose AI models, citing “legal uncertainties” and “over-reach.” The EU’s code aims to impose documentation, content use restrictions, and risk management on AI developers, but Meta warns this could stifle innovation and hamper European AI businesses.
Despite pushback from major players like Meta, Alphabet, and Microsoft, the European Commission remains firm on its August 2 rollout, with full compliance required by 2027.
3. Claude Code Users Hit by Sudden Usage Limits ⚠️
Since Monday, heavy users of Claude Code, especially those on the $200 Max plan, have faced unexpected and restrictive usage caps, sparking confusion and frustration across the community. These limits appeared without prior notice, leaving many to question the accuracy of Anthropic’s tracking and the future reliability of the service.
Amid ongoing network issues and frequent API errors, the company acknowledged slower responses but offered little clarity on the new limits or long-term plans.
4. U.S. House Concerned with NVIDIA AI Chip Sales to China 😬
The Trump administration’s recent move to allow NVIDIA to resume shipping its powerful H20 AI chips to China has sparked warnings from the head of the US House Select Committee on China. Chairman John Moolenaar argues the H20 chip significantly outpaces China’s own tech and could boost Beijing’s military AI capabilities, raising national security alarms.
Meanwhile, Commerce Secretary Howard Lutnick defends the decision, pointing out that Chinese companies already access comparable or superior components despite prior restrictions.
5. Perplexity Scores Big in India with Airtel Deal 🇮🇳
Perplexity is making a bold move in the AI race by partnering exclusively with India’s Bharti Airtel to offer free Pro subscriptions to 360 million users, aiming to tap into the country’s massive and growing tech-savvy population.
While OpenAI’s ChatGPT still dominates globally in revenue and user count, Perplexity’s India downloads and active users surged over 600% year-over-year, signaling a fierce competition brewing in the world’s second-largest internet market.
6. Microsoft's AI-Driven Layoffs Shake Company Culture 🫨
Microsoft has cut around 15,000 jobs in the past few months, marking its largest layoffs in over a decade, as it doubles down on AI investments rather than trimming AI infrastructure costs.
The cuts have hit broadly across technical, sales, and product teams, signaling a major restructuring focused on AI tools and cloud services, even affecting international roles. Morale is reportedly at an all-time low, with some employees facing repeated layoffs and forced internal job applications.
🦾How You Can Leverage:
ChatGPT's new Agent got slapped with OpenAI's first-ever "biological weapons" classification.
And that same AI can also…… build your PowerPoints.
Kinda wild, right?
On today’s show, ChatGPT’s Agent thang spent 18 minutes building a spreadsheet AND presentation deck without any human babysitting, then OpenAI was like "oh btw, every prompt gets scanned for dangerous stuff."
So on today's show, we broke down why this AI agent is both your new best friend and potentially humanity's new problem.
1 – Did Microsoft Get Dunked On? 🏀
OpenAI took $14 billion from Microsoft then built PowerPoint and Excel inside ChatGPT.
That's some next-level audacity.
Your AI can now create presentations and spreadsheets without ever touching Microsoft Office.
The relationship between these companies has been getting messy. And although this won’t immediately cost Microsoft billions in lost Office 365 subscription revenue, we’re guessing it mighta left a bad taste in Microsoft’s mouth.
(Or, maybe not?)
Try This:
Real talk…. Powerpoint and Excel is like the knowledge worker starter pack.
You should probably use those monthly agent queries really squeezing that juice.
2 – All Your Favorite ChatGPT Tricks Work Here 🤹
This ain't some separate tool you gotta learn.
(Like YOU Operator, only available on a subdomain?!)
ImageGen for visuals? Yep, Agent’s got it.
Connectors to Gmail and Google Drive? Absolutely. Agent’s got it.
The smart reasoning that makes you look brilliant in meetings? Still there.
It's like they took everything good about ChatGPT and gave it the ability to actually DO stuff instead of just talking about it.
Try This:
Pick your most-used ChatGPT feature and test it in Agent mode. We're betting it works better because now it can actually execute instead of just giving you instructions to copy-paste.
3 – Your Company Data Just Got Superpowers 🦸
Remember when everyone was freaking out about RAG and spending millions on building it?
Yeah, a small version of that is kinda just built-in now.
Toggle off web search, connect your company data, and boom - you've got a custom agent that only looks at YOUR stuff.
No more "ChatGPT doesn't know about my business" excuses.
Try This:
Connect one company data source this week and ask the agent to create a report using ONLY that information. Watch how it synthesizes your internal data without hallucinating random internet facts.
4 – It’s Got The Same Computer You Do 🖥️
OK. Not the ACTUAL same computer.
But ChatGPT Agent has its own virtual desktop, can create files and can run terminal commands.
That means it can access any public API your company uses.
Payment systems, customer databases, email platforms - if it has an API, this AI can probably poke at it with the right info.
Which is either amazing or terrifying depending on your security setup.
Try This:
Audit every system with a public API that your team touches. Create actual access controls before letting AI agents loose in your infrastructure. Learn from other people's mistakes, not your own.
5 – First AI with a High Warning Label ⚠️
OpenAI classified this as their first "high capability model" for biological weapons.
When the company that built the AI tells you it could be dangerous, maybe listen.
But honestly? The real danger isn't weapons, per se. It’s the phishing attacks, dark web and bad agent actors that are gonna abuse ChatGPT Agent.
Stay vigilant my friends. Scams are about to get crazy sophisticated.
Try This:
Set up guardrails NOW before your team starts experimenting. Define what the AI can and can't access, set time limits on tasks, and create approval workflows for anything that touches sensitive systems.
Don't be the company that learns about AI safety the hard way.
Reply