- Everyday AI
- Posts
- ChatGPT’s Updated Custom GPTs: What’s New and How They Work
ChatGPT’s Updated Custom GPTs: What’s New and How They Work
Google AI Search cuts news site traffic, U.S. chipmakers could get bigger tax break, Microsoft cuts more workforce for AI and more!
Everyday AI
July 02, 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: Think you know GPTs? Think again. Learn how OpenAI's updated custom GPTs redefine business automation with enhanced tools and reasoning models. Give it a listen.
🕵️♂️ Fresh Finds: Perplexity announces new Max tier, Capital One’s agentic AI and Luma AI’s Modify Video now on iOS. Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: Google AI Search cuts news site traffic by 40%, U.S. chipmakers could get bigger tax break and Microsoft cuts nearly 4% of workforce for AI. For that and more, read on for Byte Sized News.
🧠 Learn & Leveraging AI: We’re breaking down what’s new with ChatGPT’s custom GPTs and unveiling capabilities you probably didn’t know existed. Keep reading for that!
↩️ Don’t miss out: Did you miss our last newsletter? We talked about the future of Apple's AI, Cloudflare blocking AI scraping, U.S. Senate rejecting state AI regulation freeze, Amazon deploying 1 millionth robot and more. Check it here!
ChatGPT’s Updated Custom GPTs: What’s New and How They Work 🧑💻️
Wanna hear a lil secret?
You (likely) have no clue what custom GPTs are capable of inside of ChatGPT.
OpenAI just updated their capabilities, yet no one's talking about it.
Why?
The original hype and hoopla from their late 2023 launch fizzled and faded away, and now many AI users have written GPTs off.
Big mistake.
You won't believe what the newly upgraded GPTs are capable of.
Also on the pod today:
• Expanded Model Support for Custom GPTs 🚀
• Insight Synthesizer GPT’s Unique Abilities 🔍
• Personalizing with the Updated GPT Models 👤
It’ll be worth your 46 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 – Zencoder is an AI coding assistant, Bolt now turns Figma design into apps in one click, and Stori uses agentic AI to build your inbound momentum.
Perplexity – Perplexity is offering a new subscription tier called Perplexity Max.
Introducing Perplexity Max. Our most valuable subscription tier yet.
Built for those who demand more, Max gets you unlimited Labs queries,
access to a broader suite of frontier models, and early access to products like Comet.— Perplexity (@perplexity_ai)
4:12 PM • Jul 2, 2025
Agentic AI – Capital One is building agentic AI modeled after its own org chart to charge its auto sales.
AI Video – Luma AI’s Modify Video feature is now available on iOS.
AI Agents – Kayak and Expedia are working toward AI travel agents that turn social posts into itineraries.
Big Tech - Enterprise giants Atlassian, Intuit and AWS are planning for a future where agents call the APIs.
AI in Society – Police in Maine issued an apology for using an AI-edited photo to boast about a Maine drug bust.
1. Google’s AI Search Feature Slashes Traffic to Major News Sites 🪓
A new report from SimilarWeb reveals that since Google introduced its AI-powered Overviews last May, visits to 37 out of the top 50 U.S. news websites have dropped, with Forbes and HuffPost seeing traffic plunge by 40%. This shift comes despite significant news events, suggesting AI summaries are keeping users on Google rather than clicking through to publishers.
While Google insists these changes don’t hurt websites and highlight broader discovery opportunities, publishers warn this could undermine their revenue and investment in quality journalism.
2. US Chipmakers Could Get Bigger Tax Breaks Under New Spending Bill 🇺🇸
The Trump administration’s latest spending bill draft, recently passed by the Senate, proposes raising the tax credit for semiconductor plant construction in the U.S. from 25% to 35%. This move aims to boost domestic chip manufacturing at a critical time when export restrictions on advanced AI chips to China have hit revenues for major players like Intel, TSMC, and Micron.
If passed, this enhanced credit could incentivize more investment and expansion in U.S. chip production, potentially stabilizing supply chains and fostering innovation.
3. Microsoft Cuts Nearly 4% of Workforce Amid AI Investment Push ✂️
Microsoft has announced layoffs impacting nearly 4% of its 228,000-strong workforce as it wrestles with the soaring costs of scaling AI infrastructure, according to Reuters. This move follows an earlier round in May and targets sales roles and management layers, aiming to streamline operations while continuing its $80 billion capital outlay for fiscal 2025.
The cuts also ripple through its gaming division, including King, showing that even major tech players aren’t immune to the pressure of balancing innovation investments with profitability.
4. X to Pilot AI-Generated Community Notes on Posts 📋
X, formerly Twitter, is testing a new feature allowing AI chatbots like Grok to create Community Notes—user-contributed fact-checks that add context to posts. While these AI-generated notes will undergo the same human vetting process to ensure accuracy, experts warn about risks from AI hallucinations and potential overload on human reviewers.
Early research from X suggests combining human feedback with AI could enhance fact-checking quality, but the balance remains delicate.
5. Adobe Stock Slides Amid AI Concerns 📉
Adobe’s stock took a hit Wednesday after receiving its first Sell rating in months, reflecting growing investor unease as AI reshapes the software landscape. According to Barron’s, the downgrade signals skepticism about Adobe’s ability to maintain its dominance amid rapid AI-driven changes.
For professionals and businesses relying on Adobe tools, this could mean increased competition and pressure to adapt quickly.
🦾How You Can Leverage:
Companies spent MILLIONS building what you can now create in five minutes.
We just built a GPT that analyzed our year-old meeting transcript, identified every action item discussed, then went out and completed the research we never got around to doing.
Boolean searches for generative AI PDFs happened autonomously.
Python coding mid-task.
A meeting transcript from twelve months ago triggered live competitive research and vendor recommendations.
Nobody programmed these behaviors.
GPTs collected digital dust for eighteen months. Then OpenAI quietly dropped a massive update. Full model access restored. O3 reasoning, GPT-4o, the works.
So on today's show, we built five GPTs live to test what's actually possible now.
Your repetitive knowledge work just became automatable for twenty dollars monthly.
1 – AI Makes Decisions You Didn’t Program 🧠
Fifty years of predictable software died when o3 started thinking.
The Meeting Actionizer received a transcript. Standard protocol: extract decisions, list action items.
This GPT kept going.
Mid-analysis, it autonomously researched competitors mentioned in conversation. Built implementation frameworks. Performed sentiment analysis. Coded Python charts.
Zero instructions programmed these decisions.
You can't QA test infinite reasoning paths. Traditional deployment models break when AI chooses its own route through problems.
Try This:
Build a custom GPT for one weekly task using o3. Feed it real data and document every unexpected decision it makes. Those surprises preview non-deterministic business operations.
2 – Post-Meeting Research Became Obsolete 💭
Meeting follow-up just automated itself.
We tested this with a year-old strategy meeting. The AI researched every vendor mentioned, analyzed current market conditions, built implementation timelines, created solution briefs.
Most telling part? The recommendations matched conclusions we reached manually over multiple follow-up meetings.
This wasn't transcription analysis. The GPT understood implied research needs and filled knowledge gaps without instruction.
Try This:
Upload your next meeting transcript to a custom GPT with instructions to research and solve every open question. Let the reasoning model choose its approach. Compare against your typical post-meeting workflow.
3 – Enterprise Analysis Costs Twenty Dollars Monthly 💵
Investment firms charge thousands for financial briefings. We replicated the methodology in thirty minutes.
Upload any dataset with performance metrics and timeframes. The reasoning model identifies patterns, calculates significance, extracts strategic implications. Raw spreadsheet becomes executive presentation.
The AI doesn't just create charts. It identifies which patterns matter strategically and explains why they impact decisions.
Small teams now access capabilities that required entire departments.
Try This:
Build a custom GPT replicating one analysis your company outsources or spends weeks completing. Focus on automating methodology, not recreating findings. Calculate time savings across similar analyses.
Reply