Order ice cream delivery and build schedule spreadsheets
A user tried OpenAI's Operator on two tasks: ice-cream delivery succeeded only after the human supplied payment and approval, while a schedule spreadsheet came out error-prone and needed heavy manual correction, with continuous supervision throughout. No time-saved or error-rate data is provided. Task: Baby Brain OpenAI has finally debuted "Operator," its very own AI agent, a type of autonomous model designed to do digital tasks on your behalf like shopping for groceries online. But calling it an AI "toddler" might be more fitting. As a Bloomberg reporter describes her experience using OpenAI's new toy, the experimental tech needs a "lot of adult supervision," frequently screwing up and asking for help when it gets stuck. It's also pretty sluggish, as echoed by other users, and slow on the uptake — as a still-developing brain might be. "For several agonizing moments, I watched as OpenAI's artificially intelligent agent slowly navigated the internet like someone who's had the web described to them in great detail but never actually used it," wrote Bloomberg's Rachel Metz. "I had to monitor it the entire time." ParentGPT Experiences like Metz' suggest there's still a long way to realizing OpenAI's vision — and the industry's at large — of releasing agentic AI models that can serve as virtual employees or personal assistants, supercharging your productivity by doing the work for you. Typical large language models are limited to words. But AI agents are capable of interacting with their environment, like a user's desktop computer. That potentially enables them to do anything from browse the web — itself a fount of infinite possibilities — to using installed software. In its announcement, OpenAI highlighted Operator's usefulness in making reservations, booking flights, and creating shopping lists. The AI model is only available to subscribers of the ChatGPT Pro plan, which costs $200 per month. If it could do any of those things without help quickly and reliably, it might be a huge time saver. The tech is in a very early stage, however, and isn't as hands-off as one might hope. As OpenAI warned, Operator has to ask you for confirmation before actually completing any important tasks — betraying that the tech isn't yet trustworthy enough to be left alone. Helicopter Parent Bloomberg's Metz says that Operator successfully handled tasks like getting ice cream delivered, although it required some "guidance and permission," like providing payment info and approving the purchase. With more serious applications like creating a spreadsheet for a schedule, it frequently messed up the details, she said. OpenAI admitted that Operator still struggles with "complex interfaces" like creating slideshows and managing calendars. If it can Instacart you some food and do some shopping, cool. But is it worth shelling out $200 a month for? "Given it was just a test, I was ready and willing to keep a close eye on the product," Metz concluded. "But if OpenAI and its peers want agents to take off, they'll need to convince people that they can trust these services to act autonomously on their behalf. Otherwise, we may decide if we want a job done right, we should just do it ourselves." More on OpenAI: OpenAI Hit With Wave of Mockery for Crying That Someone Stole Its Work Without Permission to Build a Competing Product About our ads About our ads; tools: Operator. Human role: The original post does not say how a person stayed involved.; limitations: Operator is currently limited to ChatGPT Pro subscribers ($200/month). It struggles with complex interfaces like slideshows and calendars. The agent requires human confirmation for important tasks and often needs guidance when stuck. It is described as sluggish and prone to errors, making it unsuitable for fully autonomous work. Human supervision is mandatory to catch mistakes and provide necessary inputs like payment details..
Tools
Published on Agent Work Atlas:
Overview
A user tried OpenAI's Operator on two tasks: ice-cream delivery succeeded only after the human supplied payment and approval, while a schedule spreadsheet came out error-prone and needed heavy manual correction, with continuous supervision throughout. No time-saved or error-rate data is provided.
Workflow
Baby Brain OpenAI has finally debuted "Operator," its very own AI agent, a type of autonomous model designed to do digital tasks on your behalf like shopping for groceries online. But calling it an AI "toddler" might be more fitting. As a Bloomberg reporter describes her experience using OpenAI's new toy, the experimental tech needs a "lot of adult supervision," frequently screwing up and asking for help when it gets stuck. It's also pretty sluggish, as echoed by other users, and slow on the uptake — as a still-developing brain might be. "For several agonizing moments, I watched as OpenAI's artificially intelligent agent slowly navigated the internet like someone who's had the web described to them in great detail but never actually used it," wrote Bloomberg's Rachel Metz. "I had to monitor it the entire time." ParentGPT Experiences like Metz' suggest there's still a long way to realizing OpenAI's vision — and the industry's at large — of releasing agentic AI models that can serve as virtual employees or personal assistants, supercharging your productivity by doing the work for you. Typical large language models are limited to words. But AI agents are capable of interacting with their environment, like a user's desktop computer. That potentially enables them to do anything from browse the web — itself a fount of infinite possibilities — to using installed software. In its announcement, OpenAI highlighted Operator's usefulness in making reservations, booking flights, and creating shopping lists. The AI model is only available to subscribers of the ChatGPT Pro plan, which costs $200 per month. If it could do any of those things without help quickly and reliably, it might be a huge time saver. The tech is in a very early stage, however, and isn't as hands-off as one might hope. As OpenAI warned, Operator has to ask you for confirmation before actually completing any important tasks — betraying that the tech isn't yet trustworthy enough to be left alone. Helicopter Parent Bloomberg's Metz says that Operator successfully handled tasks like getting ice cream delivered, although it required some "guidance and permission," like providing payment info and approving the purchase. With more serious applications like creating a spreadsheet for a schedule, it frequently messed up the details, she said. OpenAI admitted that Operator still struggles with "complex interfaces" like creating slideshows and managing calendars. If it can Instacart you some food and do some shopping, cool. But is it worth shelling out $200 a month for? "Given it was just a test, I was ready and willing to keep a close eye on the product," Metz concluded. "But if OpenAI and its peers want agents to take off, they'll need to convince people that they can trust these services to act autonomously on their behalf. Otherwise, we may decide if we want a job done right, we should just do it ourselves." More on OpenAI: OpenAI Hit With Wave of Mockery for Crying That Someone Stole Its Work Without Permission to Build a Competing Product About our ads About our ads
Reported result
The ice cream delivery task was completed successfully, but only after human intervention for payment and approval. The spreadsheet creation task failed to produce accurate results, with the agent frequently messing up details. There are no measured metrics for time saved or error rates, but the experience suggests that the agent is not yet reliable for autonomous execution of complex tasks without constant supervision.
Human role
The original post does not say how a person stayed involved.
Limitations
Operator is currently limited to ChatGPT Pro subscribers ($200/month). It struggles with complex interfaces like slideshows and calendars. The agent requires human confirmation for important tasks and often needs guidance when stuck. It is described as sluggish and prone to errors, making it unsuitable for fully autonomous work. Human supervision is mandatory to catch mistakes and provide necessary inputs like payment details.
Try it
I want to test OpenAI's Operator agent for a low-risk task. My goal is to [replaceable input: e.g., find a specific product on a retailer's website]. I have a ChatGPT Pro subscription. Please guide me through the process. First, I will prompt Operator to [replaceable input: specific task instruction]. I will monitor the screen closely. If Operator asks for confirmation or gets stuck, I will provide the necessary input or correction. Do not let Operator complete any purchase or send any message without my explicit approval. After the task, I will document: 1) Did Operator complete the task? 2) How many times did I need to intervene? 3) Was the final output accurate? If the task fails or requires more than 3 interventions, I will stop and note the failure. I will not use this for critical business decisions until I have verified reliability over 10 successful low-risk attempts.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.