• Everyday AI
  • Posts
  • ChatGPT’s New Tasks Mode: Gimmicky feature or agentic behavior

ChatGPT’s New Tasks Mode: Gimmicky feature or agentic behavior

ChatGPT Tasks review, Microsoft launches 365 Copilot Chat, Google makes AI features free in Workspace, Meta’s AI strategy faces legal trouble and more!

Author

Everyday AI
January 15, 2025

Outsmart The Future

Sup y’all! 👋

2 quick things. 

1. We created a DOPE guide on 10 actually good use-cases for ‘Task Stacking’ actions in ChatGPT’s new Tasks mode. (That was today’s show.) If you want access, go Repost today’s livestream and we’ll send it to ya. 

2. Do y’all care about OpenAI’s o1 and o1 Pro mode? That’s OpenAI’s reasoning model. New kinda LLM if you’re unaware.

Do you wanna see a show/newsletter on how to use o1/o1 pro?

Login or Subscribe to participate in polls.

✌️
Jordan

Today in Everyday AI
6 minute read

🎙 Daily Podcast Episode: OpenAI just added a new Task mode to ChatGPT. Is it just a gimmick or does it bring us closer to agentic AI? We take a look at what’s new. Give it a listen.

🕵️‍♂️ Fresh Finds: OpenAI funds Axios newsrooms, Synthesia doubles its valuation and NVIDIA backs MetAI. Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: Microsoft launches 365 Copilot Chat, Google makes AI features free in Workspace and Meta’s AI strategy faces legal trouble. For that and more, read on for Byte Sized News.

🚀 AI In 5: Want the cliff notes on ChatGPT’s new Task mode? We break it down in this short review. See it here

🧠 Learn & Leveraging AI: Wondering if what ChatGPT’s new Task mode is? We break down how it works and ways you can use it to save time daily. Keep reading for that!

↩️ Don’t miss out: Did you miss our last newsletter? We talked about ChatGPT adding new "tasks" feature, U.S.'s new AI executive order, free ChatGPT in 2025 and the UK investigating Google Search dominance. Check it here!

 ChatGPT’s New Tasks Mode - Gimmicky feature or agentic behavior 🗓

ChatGPT's new 'Task' mode is here.

Is this some gimmicky feature that'll get glazed over?

Or a small step toward Operator, OpenAI's upcoming agentic release?

Join the conversation and ask Jordan questions on AI here.

Also on the pod today:

• Importance of Context Tasking ✅
• Utility of ChatGPT's Task Feature 🤖
• Agentic AI and Task Automation 🔄

It’ll be worth your 49 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 – Latitude is an open-source prompt engineering platform, Height.app is an autonomous project management tool and Handinger helps extract data from the internet.

Trending in AI – NVIDIA-backed AI video platform Synthesia has doubled its valuation to $2.1 billion.

OpenAI – OpenAI is funding four new Axios local newsrooms.

Google – It was revealed that the NotebookLM team had to teach the AI hosts to not sound annoyed after the Interactive Mode was launched.

Google is bringing an online AI academy to Connecticut residents.

NVIDIA - NVIDIA has backed MetAI, a startup that creates AI-powered digital twins.

Future of Work – As Gen Z job applications grow, companies are turning to AI agent recruiters.

AI in Society – A scammer used AI images of Brad Pitt to trick a woman into sending over $850K.

AI Startups – A new AI lab called NDEA is being formed by a former Google software engineer for deep learning-guided program synthesis.

1. Microsoft Relaunches Copilot Chat for Businesses 🏢

Microsoft has relaunched its free Copilot Chat for businesses, now featuring AI agents to enhance workplace productivity. This service aims to familiarize users with AI tools, ultimately encouraging a shift towards the paid Microsoft 365 Copilot subscription at $30 per month.

Users can now create virtual colleagues to automate tasks and streamline workflows, all while keeping costs manageable with a pay-as-you-go model.

2. Google Makes AI Features Free in Workspace 💼

Google has announced that all AI features will now be available at no extra cost inside Workspace, despite a slight increase in subscription prices. Users previously paying $20 for the Gemini Business plan will see their costs rise by about $2 monthly, bringing the base price to $14.

This shift includes exciting tools like email summaries, automated note-taking, and the powerful Gemini bot, making it easier for businesses to leverage AI without breaking the bank.

3. Meta's AI Strategy Faces Legal Heat 🔥

Internal communications from Meta have surfaced in a major copyright lawsuit, revealing the company's controversial strategies for training its Llama AI models using potentially pirated data from sites like Library Genesis. Efforts to obscure copyright information highlight a frantic race to compete with rivals like OpenAI, as executives admitted their need for more data to enhance performance.

This evidence could bolster the ongoing class action lawsuit against Meta, raising significant questions about the ethics of data usage in AI development.

4. Apple Faces iPhone Sales Slump Amidst AI Competition 😬

Apple's iPhone sales took a hit, declining by 5% globally in the last quarter of 2024, largely due to lackluster upgrades and increasing competition from Chinese brands like Xiaomi and Vivo.

According to Bloomberg, this downturn saw Apple's market share drop to 18%, even as the overall smartphone market grew by 4%. The tech giant's struggles are compounded by its inability to keep pace with advancements in artificial intelligence features offered by rivals, particularly in China.

5. OpenAI Tests Phone-Only Signups 📱

OpenAI is currently testing a new feature that allows users in the U.S. and India to create ChatGPT accounts using just their phone numbers, a shift aimed at simplifying user access. However, this beta version comes with limitations: phone-only users can’t upgrade to paid plans or enable multi-factor authentication without adding an email, and reused numbers face account creation errors.

This move is part of OpenAI’s strategy to enhance user engagement, especially as it faces substantial financial challenges, including a projected $3.7 billion revenue shortfall in 2024.

6. Cisco Unveils AI Defense to Secure Enterprises in the AI Era ⚔️

Cisco has launched its AI Defense solution, designed to empower businesses to develop and deploy AI applications securely. With only 29% of enterprises feeling equipped to combat unauthorized tampering, this innovative tool promises to safeguard against data leaks and sophisticated cyberattacks by providing comprehensive visibility and control over AI assets.

According to Cisco, the need for such a solution is pressing, as the risks associated with AI technology continue to escalate.

ChatGPT Tasks Review - Agentic AI or small feature update?

OpenAI JUST released ChatGPT Tasks.

We’re going over what they are, how they work and the pros and cons.

Plus we’ll talk about advanced ways this new feature can help automate your work.

🦾How You Can Leverage:

Don’t listen to OpenAI on this one. 

If you check their marketing and Tweets, you’d think their new Tasks feature is kinda like a to-do or reminders app. 

But it’s not. 

Spoiler alert — almost everyone is thinking about ChatGPT tasks wrong. 

Tasks isn't just some notification system - it's literally the blueprint for OpenAI's upcoming autonomous AI platform.

And y'all. It's WILD.

After testing five different paid accounts and pushing every limit possible, we've uncovered exactly how Tasks is laying the groundwork for OpenAI's mysterious Operator system.

This isn't speculation. The Operator subdomain already exists. 

Tasks is our training ground.

And we gave you the 101 and 201 on today’s show. Here’s what ya need to know y’all.

1 – Task Stacking: The Art of AI Chain Reactions ⤵️

Remember dominos? 

Tasks can work like that. 

(Full disclosure, they’re in Beta and OpenAI is likely getting slammed. So we’re seeing a lot of ‘Delays.’ Right now. So give it some time.) 

Oh yeah, back to dominos. 

One falls into the next into the next. Each task reads previous outputs, analyzes them, and creates new insights. No humans needed.

Here's what's fascinating: During testing, a three-task chain emerged as the sweet spot. Task one gathered marketing news at 7:57 AM. Task two analyzed those stories at 7:58 AM, specifically picking ones relevant to the business. Task three took those insights at 7:59 AM and turned them into actionable strategies.

The real magic? The AI remembered everything. When task three suggested implementing a new marketing strategy, it considered both the original news AND the strategic analysis from task two. Pure digital alchemy.

Try This:

Start with market research. Create your first task to monitor industry news every morning. 

Your second task should analyze those stories specifically for your business (pro tip: include phrases like "most relevant to [your company]" in the instructions). 

Cap it off with a third task that transforms those insights into bullet-pointed action items. Make each task explicitly reference "the context of this chat" - that's how they share information.

Actually….. Trust us. Just go reshare/respost this episode and we have a nice freebie for you that you’ll REALLY wanna see. 

2 – Context Tasks: Your AI’s Secret Weapon 🤫

Here's what nobody's talking about: Tasks live where they're born.

Create a task in a chat loaded with business context? That task becomes a context-aware AI worker.

During testing, loading a chat with 437 podcast episodes' worth of data BEFORE creating tasks resulted in hyper-relevant, business-specific outputs.

The AI didn't just find news - it found news that mattered to that specific business model.

And guess what? You can use deep research mode to instantly generate this context.

No manual data entry required.

Try This:

Before creating any tasks, feed your chat the good stuff. Company metrics, competitor data, strategic goals - everything. 

Use deep research mode to pull relevant industry data. Then create tasks that explicitly reference this context. The difference between "find news" and "find news considering our Q1 expansion into AI consulting" is staggering.

Give your ChatGPT Tasks chats a context jumpstart by feeding it content from Deep Research. 

3 – The Hidden Operator Connection: Why this MATTERS 🕵

Let's connect some dots shorties. 

OpenAI's mysterious Operator subdomain appears. Suddenly, we get Tasks - a feature that teaches us how to orchestrate AI agents. 

Coincidence? Naaaaaaaah. 

The evidence is in the details. Tasks are limited to 10 per user - perfect for learning basic orchestration. They're confined to their creation chat - teaching us about context management.

Even the bugs (like delayed execution during high traffic) are valuable lessons in AI system management.

We're literally practicing for the future of autonomous AI workflows. 

Right now.

Try This:

Build a conditional logic flow. Create tasks that make decisions based on data. Example: Task one monitors stock prices. Task two analyzes significant moves (think NVIDIA jumping 5%). Task three generates action plans based on those analyses. This isn't just automation - it's autonomous decision-making in action.

Pro Tips From the Testing Lab:

  • Tasks truncates instructions by default. Always edit your task after creation to restore your full, detailed instructions.

  • You can't use Projects mode with Tasks (yet), so load your context manually.

  • The GPT-4 Turbo model handles Tasks best - other models get confused.

  • Keep your task chains in the same chat. Moving them breaks the context chain.

  • Use specific time intervals (like 7:57, 7:58, 7:59) rather than simultaneous scheduling to ensure proper task sequencing.

The pros aren't using Tasks to remember their laundry. They're building autonomous AI workflows, testing context management patterns, and preparing for the next evolution in AI automation.

While everyone else figures this out six months from now, you'll be ready for whatever OpenAI drops next. Because Tasks isn't the destination - it's the tutorial level for the future of AI agents.

Time to level up.

⌚

Numbers to watch

90%

According to a study by HeyGen, 90% of global consumers don’t have an issue with brands using AI videos for marketing.

Now This …

Let us know your thoughts!

Vote to see live results

How often do you read/watch/listen to Everyday AI?

Login or Subscribe to participate in polls.

Reply

or to participate.