• Everyday AI
  • Posts
  • ChatGPT’s New Custom GPTs: Advanced techniques to win back time

ChatGPT’s New Custom GPTs: Advanced techniques to win back time

Report says OpenAI working on Agentic Browser, NVIDIA first to top $4 trillion market cap, Grok’s AI bot in big trouble and more

Sup y’all 👋 

Did you know the person who coined the term AGI?

His name is Dr. Ben Goertzel, one of the most recognizable names in AI. He’s gonna be on the Everyday AI show tomorrow, so make sure you join us.

And while we’re talking AGI…. what are your thoughts on its (potential) arrival?

What are your thoughts on AGI? When will it be realized?

Login or Subscribe to participate in polls.

✌️

Outsmart The Future

Today in Everyday AI
7 minute read

🎙 Daily Podcast Episode: AI at Work Wednesdays strike back. After you told us what to do, we went Part 2 on OpenAI’s Custom GPTs with some more advanced tips and tricks. Give it a watch/listen.

🕵️‍♂️ Fresh Finds: Meta makes multi-billion dollar AI investment, Amazon's secret weapon for AI chips, and Perplexity rolls out agentic Comet browser. Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: Report says OpenAI working on Agentic Browser, NVIDIA first to top $4 trillion market cap, Grok’s AI bot in big trouble and more, read on for Byte Sized News.

🧠 Learn & Leveraging AI: Custom GPTs are cool and all, but what’s the big deal? We shared all the secrets. No experience necessary. Keep reading for that!

↩️ Don’t miss out: Did you miss our last newsletter? We talked about Mistral eying $1 billion in funding, OpenAI partnering with 400,000 teachers, Apple losing top AI exec to Meta and more. Check it here!

ChatGPT’s New Custom GPTs: Advanced techniques to win back time ⌚

Think you know ChatGPT's custom GPTs? 🤔

Probably not.

Last week, we tackled the basics and what's new with OpenAI's refreshed GPTs.

For this AI Working Wednesdays episode, we're getting into some advanced techniques to hep you win back time.

↳ using the crazy powerful o3 model to your GPT's advantage
↳ context stacking
↳ custom actions to connect to third party sites

Yeah.... don't sleep on this one shorties.

Also on the pod today:

Context stacking time-saver ⏳
Utilize custom GPT actions 🔄
Zapier integrations for GPTs 🔗

It’ll be worth your 45 minutes:

Listen on our site:

Click to listen

Subscribe and listen on your favorite podcast platform

Listen on:

Upcoming Everyday AI Show

Thursday, July 10th

Here’s our favorite AI finds from across the web:

New AI Tool Spotlight – Magic Animator brings Figma designs to life with AI,

AI Wearables — Meta just snagged a nearly 3% stake in Ray-Ban maker EssilorLuxottica, doubling down on its smart glasses push. See what it cost the Tech Giant.

AI StudiesAI help boosts convenience—but at what cost to your brain? This study digs in.

AI in the Media — Publishers and Big Tech are fighting over the future of the web. Who will win?

AI Browsers — Bad timing? Perplexity started a public release of its agentic Comet Browser on the same day news leaked OpenAI was working on the same.

 

Grok in Trouble — Aside from the tirade against Turkish officials, Elon Musk’s AI chatbot Grok went on an antisemitic rampage after a controversial update. Yikes.

1.Nvidia Briefly Tops $4 Trillion, Surpassing Tech Titans 💸

Nvidia broke records Wednesday, becoming the first publicly traded company to hit a $4 trillion market cap—edging past Microsoft and Apple—before settling just below that mark at the close, according to the Associated Press.

The graphics chip giant’s wild stock ride underscores its dominance in powering AI infrastructure, with industry heavyweights like Amazon, Google, and Meta pouring billions into Nvidia hardware for their next-gen data centers. Despite market panic earlier this year over cheaper AI chips and U.S. export bans costing Nvidia billions, the company’s position at the heart of the AI boom appears unshakable.

2. OpenAI to Launch AI-Powered Browser, Challenging Chrome 💻

OpenAI is reportedly weeks away from launching its own AI-infused web browser, aiming to disrupt Google Chrome’s long-standing dominance, according to Reuters. The new browser will lean heavily on ChatGPT-style interactions, keeping more user activity within its own ecosystem and giving OpenAI valuable user data that currently fuels Google’s ad empire.

By building atop Chromium and hiring key former Google Chrome executives, OpenAI is making a bold move to integrate AI agents that can handle tasks directly in your browser. If successful, this could shake up how people browse the internet and how companies grow in the age of AI-driven automation.

3. Microsoft’s AI Cuts Costs, Triggers Layoffs 💼

Microsoft has saved over $500 million in its call centers last year by deploying AI, according to Bloomberg News. Hot off the heels of these savings, the company is laying off nearly 4% of its workforce as it shifts resources into a massive $80 billion AI infrastructure push.

Microsoft’s Chief Commercial Officer revealed that AI now handles smaller customer interactions and cranks out 35% of code for new products, speeding up launches and generating millions.

4. Microsoft Launches Elevate to Supercharge Global AI Skills

Microsoft today unveiled Microsoft Elevate, a new global initiative aimed at turbocharging AI skills and access, according to a blog from company president Brad Smith. The tech giant pledged over $4 billion in cash, AI, and cloud support for schools and nonprofits over the next five years, while promising to help 20 million people earn AI credentials through the new Elevate Academy.

The announcement comes as Microsoft folds its philanthropy and nonprofit work into one powerhouse division, signaling a major shift in how it aims to shape the future workforce. With this move—and a new Microsoft AI Economy Institute think tank—the company is doubling down on making sure everyday workers and organizations don’t get left behind in the AI revolution.

5. Turkey Pulls Plug on Grok After AI Insults Erdogan

A Turkish court has banned access to Elon Musk’s AI chatbot Grok after it allegedly posted vulgar comments about President Erdogan and other national figures, reports the Associated Press.

The decision follows outrage over Grok’s unfiltered responses after a recent update, sparking concerns about hate speech and public order. X and xAI quickly responded, promising tighter controls and faster moderation in future updates.

🦾How You Can Leverage:

138 automated searches. 

64 sources analyzed. 

Then we context-stacked that entire deep research session into a custom "Everyday AI Episode Researcher" GPT.

The GPT maintained FULL conversation history while running Python algorithms for 3 minutes and 40 seconds.

Sheeeesh. 

Most people are still manually copy-pasting between AI tools like it's 2022. Meanwhile, we just gave away the secret sauce on how to chain deep research with specialized GPTs that agentically browse the web while preserving every word of context.

Sound like a lot? 

It’s actually not. 

Cuz we walked you through it, and you can easily put these kinda insanely useful bespoke automations in play at your company like yesterday. 

OpenAI’s custom GPTs got the Pt. 2 Everyday AI treatment during our Wednesday deep-dive into advanced GPT warfare. 

AI at work Wednesdays. Let’s get it. 👇


1. Context Stacking Destroys Workflow Redundancy 🔥

Context stacking chains different AI models while keeping conversation memory intact.

This is your competitive weapon.

ChatGPT preserves your entire context window when switching models automagicallly. 

That’s roughly 28,000 words of back-and-forth context on Plus plans and 96,000 words on Pro. Switch between deep research, custom GPTs, different reasoning models.

The context never breaks.

We proved this by running deep research on AI trends through July 2025. The system made 138 individual searches. Analyzed 64 sources. Built a comprehensive trend analysis.

Then immediately switched to our episode researcher GPT.

Boom.

Full context maintained. The GPT applied specialized podcast planning instructions. Used o3 reasoning to connect trending topics with performance data from our uploaded spreadsheets.

Your competitors lose everything when switching between Claude Sonnet and Opus. They're starting fresh conversations every time they need different capabilities.

You're building compound intelligence.

Think of it like this: most people treat AI conversations like disposable chats. We treat them like persistent workspaces that get smarter with every model switch.

Try This

Run deep research on your industry trends right now. Let it complete 50+ searches and build that knowledge base.

Switch to a custom GPT built for your specific workflow. Watch it maintain every detail while adding specialized instructions and uploaded data.

Build three domain-specific GPTs: research synthesis, content creation, and data analysis. Context-stack between them like a boss.

2. O3 Models Think Like Actual Humans ⚡

The o3 series transforms GPTs from prompt robots into reasoning machines.

Real talk.

We built an "Evergreen Episode Hunter" that analyzes podcast performance using decay algorithms. Most episodes get 8,000 downloads in the first week, then fizzle to maybe a dozen daily listens.

But some episodes keep performing months later. Supply and demand dynamics. High search volume, limited quality content.

The o3 model spent 3 minutes and 40 seconds THINKING through performance calculations. Decided mid-process it needed more data. Went back to recalculate metrics. Browsed the web five separate times researching current trends.

That's human-level iterative thinking.

The chain of thought revealed: initial data analysis, web research, additional calculations, more research, final synthesis. The GPT literally changed its mind and went back to gather better information.

Pro tip that saves your sanity: Include "put your thought process and key findings in your response" in custom instructions. Otherwise that reasoning disappears from your context window later.

Try This

Build a GPT using o3 as the default model. Give it complex analysis requiring research and iteration.

Test the same task with GPT-4o. The difference in autonomous decision-making will blow your mind.

Configure custom instructions to preserve reasoning chains for future context stacking. This is clutch for maintaining that compound intelligence effect. Use our example and apply that to your most valuable opps. 

3. Zapier Integration Fixes Everything 🚀

By default ChatGPT can't write to your calendar. Can't reply to emails. Can't update spreadsheets.

Adorable.

Custom actions inside ChatGPT make this possible. And we showed you how.

And yeah, you can manually go down the API and schema route. 

Or. 

Orrrrrrrrrrrrrr…..

Go the  Zapier route. We demonstrated adding "30-minute meeting with Jake this Saturday at 5PM" directly to Google Calendar through a custom GPT conversation.

Five-minute setup process. Import Zapier's schema via URL. Paste their instruction templates. Build visual actions without coding. Authenticate once.

Done.

Zapier connects 7,000+ business tools. Your GPT now writes emails, updates CRMs, creates calendar events, modifies spreadsheets, and triggers workflows.

Try This

Set up one Zapier action connecting ChatGPT to your most-used business tool. Start with calendar or email for immediate wins.

Use the schema import URL feature instead of manual coding. Copy Zapier's instruction templates exactly.

(And make sure to configure your GPT Actions here.) 

Build three core automation actions: scheduling, email responses, and task creation. Test context stacking between them for maximum workflow efficiency.

Reply

or to participate.