While AI agents or agent AIS is launched as the next leap in human efficiency, advertising -free reality is as good as people who design and use them, just like other AI technologies.
They still need a person to know what the problem is and to give the right command. What if you would have combined an agent’s horsepower with a AI that could think like a human? How long can it be achieved?
This question is that many people in the AI industry are already looking at past agents and only focus on net, artificial general intelligence. There are some leaders in the technology industry who believe that AGI is about two years. Google co -founder Sergey Brin, sending a note to employees to work 60 hours a week in the near future by sending a note, “the competition is extremely accelerated and the last race for AGI is standing,” he said.
But let’s be real, are we really so close to AGI? Can it really increase the adaptability and complexity of the human mind? And even if so, should we look at the agencies now and wait for the promise of AGI?
Dhaval Jadav is the global CEO and founding partner of Alliant.
Chris Stephenson is the General Manager of Smart Automation, AI and Digital Services in Alliant.
True AGI is still a long road away
Most of the speech around AGI started with AI reasoning models. Reasoning models, such as Openai’s O1 and X’s Gok 3, are designed to “think” problems before responding. But the human mind is much more sophisticated than a reasoning machine. Reasoning models do not solve the three biggest problems related to obtaining AGI.
The first problem is that AI models, even reasoning models should be constantly updated and trained. This requires constant human supervision to strengthen AI’s logic. He will constantly remain in a embryonic state without telling a person that he thinks of a AI correctly or wrongly.
The second problem is that existing AI models cannot assimilate information or adapt to new situations. If an AI model can consider a mathematical problem, it does not mean that he can think of it with a legal problem.
Finally, the biggest problem is that AIs cannot create unique ideas. Artificial intelligence is limited to the data in which “thought” is trained. In other words, AI can only re -express the already existing concepts. We are struggling to find a way for AIS to find something original.
NASA’s software former chef Robert Ambrose, Robotics and Simulation section, a possible solution is to encourage AIs to encourage AIs. If you can make it possible for an AI to think abstractly and spontaneously by suspending normal reasoning for a long time, and then you can create artificial original thought by enabling AI to include these abstract thoughts in educational data.
But scientists don’t even know why people dream or how they dream, they provide the possibility that AI’s dream of electric sheep is far farther.
Don’t ignore the intermediaries – now they can help you
Whether you believe or believe, you don’t have to wait for AGI to check your representatives, the general intelligence is still working. If you have been muscles with how you set up your AI agent, you can use it to significantly reproduce your productivity. Instead of performing separate AIs and automations to perform tasks manually or do business, you can rely on an AI representative who will project management for you.
Do not make an error as AI agents as a kind of magic “easy button .. Whichever technology companies promise you, there are things that need to be done at the front end to make AI agents really effective.
Imagine this way, if you probably rent a person with general intelligence and give them a working guide, would you expect that they can do your job on the first day?
Like a new employee, your representative needs to be trained in your processes to learn how to do the job. Vehicles should be given and trained in how to use them to perform every task in the process chain. This means that you need to create basic automations for your AI to run it according to its goal. In addition, it should be trained on how to handle different exceptions, barricades and variations that may interrupt the nominal process flow.
The difference with an agent is no longer under the risk of losing an employee after being built once. Hiring and educating people is not only expensive, but no matter what the role is permanent. Time and deposit to a representative is often a single cost, you do not need to employ or train a new one.
This may seem scary, but today we are doing it and guess, even the net could not work without leaving this floor.
When the first work is done, you have an agent that will not only work at a speed that one person can never do, but at the same time he will be able to perform much more tasks at the same time.
Do not fix your hopes to SCI-FI Timeline
According to the Return to the Future, we would have a travel of flying cars and time until 2015. Blade Runner said that we will have completely sensitive Androids until 2017. Skynet’s artificial super intelligence comes online in 1997. AGI is still in the field of science fiction, if we wait for it, we are now missing the current progress.
It is not about choosing between agents and agents, but to benefit from the concrete benefits of existing AI technology while maintaining a pragmatic perspective on future developments. Today, by focusing on the thoughtful implementation of AI agents – creating net processes, creating the necessary automations and providing appropriate training – organizations can realize significant efficiency gains without waiting for the uncertain timeline of AGI.
Instead of determining when the machines can think like people, we should focus on how they can help us think better and work better at the moment.
We evaluate the best task management application.
This article was produced as part of Techradarpro’s best and brightest minds in the technology industry today. The views expressed here are the author’s views and are not necessarily the views of Techradarpro or Future PLC. If you are interested in contributing, learn more here: https://www.techradar.com/news/submit-to-techradar-Pro
