logo
I tested ChatGPT Agent on 5 everyday tasks — here's what happened

I tested ChatGPT Agent on 5 everyday tasks — here's what happened

Tom's Guide3 days ago
Since the announcement of ChatGPT Agent, I'd been eagerly awaiting the model to show up in my menu of tools. ChatGPT Plus users should all have the new model now. If you don't see it, try logging out and back in again.Once it appeared, I just had to know, could the AI actually do things like cancel subscriptions, plan a family trip and order my lunch?
To find out, I tested ChatGPT Agent with five very real tasks from my everyday life. Some were impressive. Others were frustrating. But all of them offered a glimpse into what the future of AI-powered assistance might actually look like.
Here's what happened when I put ChatGPT Agent to work.
Prompt: 'Help me find the Big Into Energy Labubu near me.'I've already found the Big into Energy Labubu with Google Search, but I wanted to see if using ChatGPT Agent was any easier. As one of the hottest toys on the market, it's nearly impossible to track one of these things down.ChatGPT followed up questioning whether I wanted more information or to purchase one. In this case, I said purchase because information is much easier to find. From there, the AI went to work. I could actually see it checking various websites, reading information, thinking and more within the chat window.
In six minutes, the Agent found the Labubu I requested, added it to the cart, and headed to the checkout. It then asked for my shipping address and credit card information. I was able to take control of ChatGPT's browser to finalize the sale.
Verdict: This was much faster and easier than using Google. I will definitely be using ChatGPT Agent in the future for hard-to-find items.
Prompt:"Plan a 4-day family trip to San Diego, including flights from Newark, hotel options with a pool, and activities for kids under 10. Book everything using my Google account and save an itinerary in Google Docs."
When I attempted to move on to the second task, my entire computer crashed and I got an error. After logging back in again, I was able to start a new task.This prompt was a very ambitious task, but I decided to go big to really see what the ChatGPT Agent could do. The AI truly flexed its potential and its limits with this one.
The agent searched Google Flights, compared hotels with kid-friendly amenities, and listed family attractions like the San Diego Zoo, LEGOLAND, and beach days. It created a beautiful daily itinerary in Docs and even embedded links to book the trip.
Luckily, booking required my manual approval. For privacy and safety reasons (thankfully), the agent doesn't auto-purchase flights or rooms. Instead, it pre-filled forms and waited for me to hit "confirm."
Honestly, I don't think I could ever let AI handle this part for me. I'm too much of a control freak when it comes to vacations and I don't trust AI.
Verdict: An incredible planning tool and I think it's fun to see the AI 'working' through the prompt. I don't think that will ever get old.
Prompt: "Create a 5-day healthy dinner meal plan under 600 calories per meal. Then generate a grocery list, check prices at my local ShopRite, and export it to a spreadsheet."
Meal planning is one of those chores that sounds easy until it's 5 p.m. and your fridge is full of random ingredients that don't go together. ChatGPT Agent solved that problem fast.
It generated five balanced dinner recipes (lemon herb chicken, veggie stir-fry, ravioli, etc.) with simple ingredients and clear instructions. Then it built a grocery list categorized by section (produce, pantry, dairy), checked local prices through Instacart data, and exported everything to Google Sheets.
It even told me which ingredients were on sale nearby. I emailed the list to myself to use next time I go shopping.
Verdict: I often refer to AI as a 'game-changer' sometimes ad nauseum, but it's so wildly helpful for things like this, that I can't think of a better descriptor. This task felt like having a personal nutritionist and assistant rolled into one. 10/10 will use again.
Prompt: "Order me a chicken Caesar wrap and a lemonade from DoorDash."
Luckily, ChatGPT didn't crash after the first time and I was ready to order lunch. Immediately arriving on DoorDash, the AI asked me to log in. I wanted to make this as 'hands off' as possible so I told it to use the site as a guest.
I had read in several forums that ordering food had caused some users to get frustrated at ChatGPT Agent. So I was prepared for things to go awry. The only hiccup I had was the AI not knowing my zip code because it signed in as a guest. Once I told it, everything went smoothly.It ordered my lunch and then handed everything over to me when it was time to pay and enter my address. Because I hadn't specified a restaurant, I was impressed by ChatGPT's ability to find a chicken Caesar wrap on its own.Verdict: More of a helpful sidekick than a hands-free solution. This wasn't much of a time saver and I probably will order lunch for myself without using ChatGPT Agent.
Prompt: 'I have to renew my license and get a real ID. Can you book an appointment for me at my local Department of Motor Vehicles?'When I got a notice in the mail the other day that I couldn't renew my license online, I got that feeling of dread in the pit of my stomach. The thought of spending hours at my local Motor Vehicle office especially in the summer, was overwhelming.
But in 19 seconds, ChatGPT pulled up everything I needed to book an appointment and I was all set. Zero hassle.
Verdict: Fast, smart and surprisingly effective for a niche errand.
ChatGPT Agent isn't perfect, but it's one of the most capable AI tools I've tested to date. It won't replace your human assistant just yet, but if you're like me and don't have a human assistant to begin with, it's the next best thing.
It can handle real, time-consuming tasks that go far beyond answering questions or summarizing PDFs.
If you're already using ChatGPT Plus or Team, it's absolutely worth trying — just be prepared to step in here and there and occasionally restart the app completely.
Follow Tom's Guide on Google News to get our up-to-date news, how-tos, and reviews in your feeds. Make sure to click the Follow button.
Orange background

Try Our AI Features

Explore what Daily8 AI can do for you:

Comments

No comments yet...

Related Articles

Microsoft becomes the second $4 trillion company
Microsoft becomes the second $4 trillion company

The Verge

time10 minutes ago

  • The Verge

Microsoft becomes the second $4 trillion company

Microsoft's stock price has risen so much today that it has passed a $4 trillion market valuation for the first time in its 50-year history. The software maker is the second company to be valued at $4 trillion, after Nvidia reached a market cap of over $4 trillion earlier this month. Microsoft has reached this milestone thanks to better-than-expected earnings, and the company reporting its Azure revenue for the first time. Microsoft revealed last night that its Azure cloud computing business had generated more than $75 billion during its 2025 fiscal year, and the company's stock price immediately started soaring in after-hours trading on the Nasdaq. Microsoft's share price is up around 4 percent today, helping it maintain its $4 trillion market valuation. Microsoft has spent more than a decade building up its Azure cloud business, which is ahead of Google's in revenue but still second place to Amazon Web Services. Azure now touches every corner of Microsoft's businesses, including powering its AI projects, its Office software, parts of Windows, Xbox, and more. Microsoft's investment in cloud computing has positioned it well to take advantage of the computing needs for the future of AI, too. Earlier this year there were plenty of questions over the $100 billion investment into The Stargate Project, and Microsoft CEO Satya Nadella was quick to defend the company's own $80 billion spend on cloud and AI data center projects. 'All I know is, I'm good for my $80 billion,' said Nadella at the time. Microsoft CFO Amy Hood revealed yesterday that Microsoft is now planning to spend $30 billion on its AI infrastructure investments in the next quarter. If Microsoft keeps up that level of spending then the company will total $120 billion or more on cloud and AI infrastructure over the next fiscal year. Posts from this author will be added to your daily email digest and your homepage feed. See All by Tom Warren Posts from this topic will be added to your daily email digest and your homepage feed. See All AI Posts from this topic will be added to your daily email digest and your homepage feed. See All Microsoft Posts from this topic will be added to your daily email digest and your homepage feed. See All News Posts from this topic will be added to your daily email digest and your homepage feed. See All Tech

AI startup Anaconda raises $150 million in Series C funding led by Insight Partners
AI startup Anaconda raises $150 million in Series C funding led by Insight Partners

Yahoo

time14 minutes ago

  • Yahoo

AI startup Anaconda raises $150 million in Series C funding led by Insight Partners

(Reuters) -Anaconda, a leading provider of open-source Python software for data science and AI, has raised more than $150 million in a Series C funding round led by Insight Partners, with additional participation from Abu Dhabi's Mubadala Capital. The funding round values the startup at $1.5 billion, Bloomberg News reported, citing a person familiar with the matter. The company said on Thursday the fresh capital will support product development, potential acquisitions, and international expansion, as well as provide liquidity for employees. The funding comes amid increased competition in the enterprise AI software sector, with Python continuing to dominate as the programming language of choice for AI development. U.S. startup funding has surged 75.6% in the first half of 2025, led by the AI boom, putting it on track for its second-best year ever, even as venture capital firms struggled to raise money, according to PitchBook. This year's boom has been driven largely by major AI investments and bold bets from big tech companies, a wave of activity set off by the debut of ChatGPT in late 2022. In the past three months alone, $69.9 billion was invested in U.S. startups. Anaconda did not immediately respond to a Reuters request for comment regarding the valuation. The company is seeking to capitalize on growing enterprise demand for open-source tools as organizations shift from isolated data science projects to broader AI applications. The startup has also expanded its leadership team, hiring executives with backgrounds in enterprise technology and product innovation. The funding follows Anaconda's launch of a new AI platform and a partnership with Databricks.

MHS Unveils ‘MACS' Micro Aqua Cooling System to Tackle AI Semiconductor Heat in Datacenters
MHS Unveils ‘MACS' Micro Aqua Cooling System to Tackle AI Semiconductor Heat in Datacenters

Associated Press

time40 minutes ago

  • Associated Press

MHS Unveils ‘MACS' Micro Aqua Cooling System to Tackle AI Semiconductor Heat in Datacenters

As generative artificial intelligence (AI) dramatically raises computational demands—such as creating over 700 million images within one week using ChatGPT (equivalent to the daily electricity consumption of approximately 67,000 U.S. households)—heat generation in AI semiconductors has become a pressing concern for datacenter operators. South Korean tech firm MHS is introducing MACS (Micro Aqua Cooling System) - a cutting-edge liquid-based cooling solution that utilizes a microchannel structure to significantly enhance coolant flow and expand the heat-exchange surface area. This design delivers performance that far exceeds that of conventional water-cooling systems. MHS is actively validating MACS technology through expanding collaboration with Korean AI semiconductor fabless companies, including Furiosa AI and Rebellion AI, positioning itself at the forefront of datacenter cooling innovation. 'No matter how advanced AI semiconductors become, without effective cooling, they simply cannot perform at scale,' stated Jason Lim, CEO of MHS. 'System availability ultimately depends on cooling technology,' he added. Lim emphasized that modern AI processors typically generate heat far exceeding one kilowatt, underscoring the necessity of liquid-cooling methods. While immersion cooling—fully submerging servers in dielectric fluid—is under consideration industry-wide, Lim cautioned that its 'high infrastructure cost' and 'management complexity' make it impractical for many datacenter operators. In contrast, MACS utilizes narrow (tens to hundreds of micrometers) microchannels, optimized flow paths, and capillary-driven fluid dynamics to achieve effective heat removal even at modest flow rates. Lim pointed to NVIDIA's AI H100 processor—which dissipates approximately 700W—and noted that future generations ('Blackwell,' 'Rubin,' and 'Panther') could generate heat loads up to 1 kW per chip, intensifying the demand for highly effective and reliable cooling solutions. In May, despite a global slowdown in venture capital investment, MHS secured Series A funding from WeVentures and Gyeongnam Venture Investment to accelerate the commercialization of MACS and expand R&D capacity. With MACS poised to meet the rising demand for datacenter cooling, MHS plans to expand its international presence and explore partnerships with global AI infrastructure leaders. Media Contact Company Name: MHS Co., Ltd. Contact Person: Chris Oh(Head of Sales) Email: Send Email Phone: +82-10-2731-5730 Country: South Korea Website: Source: MHS Co., Ltd.

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into a world of global content with local flavor? Download Daily8 app today from your preferred app store and start exploring.
app-storeplay-store