What exactly does Dots do when it's offline?

When offline, Dots continues to execute predefined tasks using its independent operating system. For example, it can open a calendar every morning, review the schedule, and prepare a to-do list. Even if the internet connection is lost, Dots can still access local files and folders to complete its tasks.

How does Dots differ from other chatbot tools?

Dots stands out because it doesn't rely on user prompts or internet connectivity to function. It operates autonomously in the background, handling routine tasks without needing to be supervised or connected to the web. This makes it unique among other chatbot tools that typically require user interaction and an internet connection.

What kind of tasks is Dots designed to handle?

Dots is designed to handle small, regular tasks that users often put off, such as checking a calendar or preparing an invoice. These tasks are typically simple but necessary and often get overlooked. Dots can be instructed to perform these tasks without user supervision, aiming to simplify daily routines.

Can Dots handle more complex tasks, and if so, how do I know when it's ready?

Dots is currently designed for small tasks, but users can gradually give it more complex tasks as they become comfortable with its capabilities. However, the article doesn't provide clear guidelines on when Dots is ready for more complex tasks, leaving some uncertainty about its full potential.

How does Dots compare to a human assistant?

Dots excels at handling small, routine tasks without needing emotional feedback or appreciation. Unlike human assistants, Dots doesn't require emotional relation and performs its tasks regardless. However, it currently lacks the ability to cater to user preferences or handle more complex, non-routine tasks.

What happens if Dots encounters a problem while offline?

The article doesn't specify what happens if Dots encounters a problem while offline. Since it operates independently, it may not be able to troubleshoot issues without an internet connection. Users should consider this when relying on Dots for critical tasks while offline.

OpenAI Dots: Offline Functionality

OpenAI's latest feature, Dots, launches a personal assistant that runs even when users go offline. The function keeps working offline, making it unique among chatbot tools. The feature was tested by a company that reported a forgotten invoice was noticed, prepared and sent after approval.

The AWOL assistant

Dots is an automated personal task manager; a personal assistant with its own to-do list, working actively in the background to tackle routine activities. It is cobbled together from a bundle of OpenAI technologies to create a service that takes action. As a standalone product, Dots doesn't respond to user prompts. Instead, it executes self-contained tasks without user supervision, aiming to simplify daily routines. Dots' ability to function offline makes it distinct. It can continue to execute tasks by an independent operating system even when the internet connection is lost. Users can instruct Dots to access local files and folders to complete its job.

Small jobs waves

The tasks Dots handles reflect an unstated assumption that users have a set of small but regular tasks to perform. These tasks generally don't even seem to be capable of being automated — they are those little jobs one leaves for 'when I get a chance,' and we never do, says OpenAI. In a visual display, OpenAI demonstrated the routine of opening the calendar every day. Just one task: "Every morning, review my calendar. Tell me what needs preparing. Don't change anything." Then, to check for mistakes, the user can see what Dots' output looks like. The idea being that if Dots passes the muster of this small task, it can be tried on larger tasks. Once users get comfortable with the feature, they are instructed to give it always the smallest job before they hand it something more complex. At this point, it is still quite unclear from the visual presentation what other tasks Dots can handle. Depending on which ChatGPT plan the user has purchased, Dots is not available on all of them and may not be available to users in all regions. For example, Pro users outside the EEA, UK, and Switzerland can access Dots. Business Premium users in supported regions can also access it. The Enterprise beta users have to enable the admin to access the service and the rollout is gradual.

Start small

The visual data shows a robust demo of Dots. While the display is successful at advertising, there are still several unanswered questions that remain. For instance, at what point will the user be confident to hand over Dots a more complex task? The visual data leaves this crucial piece of information out, raising questions about how Dots capacities have been tested. Will Dots will tackle the job correctly if the user is operating offline without being able to troubleshoot? It remains unclear what other system capabilities Dots has. As of now, it seems like a small, simple assistant.

Where Dots excels**

Small tasks that Dots are capable of currently are limited — compared to human capabilities — and currently do not cater to user preferences. Dots is a machine. It does its job regardless of if it is appreciated or not, unlike human assistants that require emotional relationships and trust. Additionally, Dots is currently showcased as an assistant for adults, but the company does not explain how it differentiates tasks according to individuals' needs. Most tasks are repetitive and routine, like opening an Excel file every day for a job. So far there are no tasks shown to be tailored to individual needs or diversity. The idea of having an automated personal assistant is that it takes tasks off a human's to-do list, but it is unclear if this assistant can actually reduce the heavy mental burden of daily tasks.

To do lists

Dots is not able to create a to-do list for the user. This is somewhat expected as it is difficult to automate the creation of a task list that is so variable from person to person. The best Dots can do is to execute a series of tasks it has been instructed to complete by the user. In the visual sample, Dots is shown to take a routine task, like opening an Excel file, and you can see the completed task on the desktop. There, in the background, Dots is supposed to do the routine work of maintaining a database like an invoice file.

Fundamental questions

OpenAI still hasn't answered the essential questions on how Dots will handle its more complicated tasks. For complicated tasks like sorting out a client's files and sending invoices, will there need to be an active user engaging with the process? This is a daunting task because the user needs to reach out to clients to get the necessary files. If the user does contact the client, then what tasks will Dots be able to complete of this task on its own?

Personal towel management

In the visual presentation, Dots was only tested in one capacity. It was asked to spot and send an invoice. The company publicly reported the test by using an actual invoice in a real-world situation. The test showed that Dots took the invoice, prepared, and sent it after approval. While it seems impressive, it does not show how reliable Dots can be for its functions. The user still has to supervise the assistant because it still does not have a standardized method of communication. The brainstorming of the tasks is shown in the form of messages that users can read in the chatbox like this: OpenAI says one tester's dot: -Noticed the missing invoice.- Prepared it. -Sent it after approval. Company-reported. Not our own test. It doesn't show what Dots actually looks like doing the task, making the visual less than impressive. Additionally, there is no insight into how Dots takes tasks from the user. If the user has to communicate with Dots through an interface, how much time does it take to learn how to communicate with it?

Where Dots makes the difference

For those who are interested in buying Dots, currently, it is helpful to know that it can do small tasks like managing files. Depending on the user's needs, it might be useful in the future to have Dots manage emails, send messages, and schedule meetings. Additionally, Dots can be used to automate responses to phone calls. While this can be useful for some users, it is unclear if this will be useful for all users. Finally, for those who are skilled at changing settings and using computer systems, having an automated assistant is a good idea. However, for those who lack the time and knowledge, it may not be as helpful.

When Dots can work

The requirements to make the Dots function are relatively straightforward if you already have a computer. The user needs to have a functioning computer with access to the internet. But Dots is set up on a desktop, which means that the machine needs to have enough storage space and the capability to run at a high speed. If the user wants to try a real-world test, like the invoice system, they also need to have a well-stocked file system so that Dots can identify what needs to be done. Always remember to check Dots work because errors usually happen. **When all is said and done, the future of Dots remains to be seen. Will Dots be able to construct a to-do list from scratch? Or will it always need to be supervise? A machine can perform tasks according to a set of instructions, but what will happen when a mistake is made? Will Dots be able to make the necessary reparations, or will it need to be supervised the whole time? Finally, how does the user know if the Dots is a better assistant than a human?