The development of the Great Action Models begins its career

The world of artificial intelligence continues its unstoppable path, opening more and more paths on which to develop, and it seems that now it is the turn of the LAM —Large Action Models— or Large Action Models in Spanish, the main attraction of the Rabbit R1.

Although it is true that it is a door that has just opened, the technology giants are beginning to see the paradigm shift to move AI from “saying things to doing things”.

And as we have learned, OpenAI, the company behind ChatGPT or SORA among others, I would be developing a model that can execute orders at the user's request, which would cause a real leap in terms of Artificial Intelligence.

Currently the models we have available through OpenAI would not allow you to really do anything with the text they generate, since if you wanted to send an email to a colleague or book flights and hotel rooms on your vacation, you would have to do it manually.

AI Agents: OpenAI is trying to change that.

Rabbit Inc. is already working on Large Action Models and as we have learned, the company OpenAI is developing similar technology based on "AI agents" to execute tasks on phones or computers. so that we cede control of execution to the agents.

OpenAI would be training its AI agents by showing them examples of people using computers so that they can understand the way in which we interact with the different interfaces so that they can imitate it.

One of the agents will, as far as we know, be able to handle web-based tasks, such as booking tickets or collecting data from websites. The other could work with apps on the device; as an example, transferring data from a document to a spreadsheet.

All this is very similar to what we already saw in the presentation by Jesse Lyu regarding the upcoming launch of the Rabbit R1. From ordering an Uber to finding the ideal hotel by simply asking it to do it for us.

It appears that OpenAI has been developing these AI agents for over a year and is training them by showing them examples of people using computers, This is the same approach that tech startup Rabbit used to train the AI agent on its R1 device..

At the moment we do not know the launch date of the OpenAI AI agents and it has yet to be determined, although seeing at the pace we are advancing in recent weeks or months, perhaps we will have news before summer.

Challenges ahead: ChatGPT is based on a Great Language Model, which bases its output on complex models that predict what is most likely to come next. But those AI models sometimes “hallucinate” or present false information as accurate.

OpenAI will need to convince people that its AI agents can protect their private information.

That's not a big problem if your only interactions with ChatGPT occur on the OpenAI website, where you can verify the veracity of their answers, but the appeal of an AI agent is that it is autonomous: you just give it a command and it does it.

It's unclear how OpenAI plans to ensure its AI agents don't make similar, but much more costly, mistakes. if customers give them access to their email, credit card details and other personal information. Even if the agents generally work, the company will need to convince people that the system can protect their private information.

Still, if OpenAI manages to overcome these issues, its agents could signal a new era in the way we interact with AI, just as ChatGPT did in 2022.

Will it be competition for Rabbit and his LAM?