![](https://images.ecency.com/DQmZuoQ79q1YWEQHr1cTumo8VPBfZk23ptJMKH4a66rrc8X/ai_gen.jpg)



I've been getting a little heavy into Ai again as for the last two years from the front facing things nothing has really changed about it. We have images, video, LLM (large language models) etc. but they still all have issues and seem to be advancing rather slowly. 

In fact so slow the quantum computers got a big headline and became the focal point once again. However AI is now starting to see some changes itself which many are missing. Let's review what's new in the Ai sector and what we might be seeing soon.

# Ai Agents

The main thing that makes Ai agents so powerful is their way of "thinking" I mean if you want to call it that. Basically it's inputs from various places and then it acts based on those inputs. In a way you could think of that as thinking aka a concisions. I mean really people don't understand how AI works in robots and how it figures things out in these new models. It's pretty wild and not fully understood yet.

This ability for the Ai agent to "think" allows it to perform tasks on it's own based around a certain command structure or goal for the AI. 

These Ai agents have core functions in which communicate to each other to figure things out. Those being, sensors to pull data and make note of change happening such as price changes. Control centre which takes that data and compiles it into something useful that the effector uses to perform an action based on that information it was given.

Now the biggest win for all of this is that we can see it's using a vast amount of systems interconnected to come up with a result. But the biggest win here for Ai agents is it's ability of memory. It pulled information from years of data already on the web and new data and compounds ALL of it quickly to help make decisions itself.

This is the big win as the human mind can't do this. The human mind either fills in gaps with incorrect data which the person feels is true or simply can't hold enough of that data to figure something out. 

# AGI

AGI gets thrown around a lot but as we can see above the current AI is simply a few different programs with goals compiled to create something that feels human like however is not.

AGI on the other hand would be fully thinking like a human of which is only still just theory. 

# Other Applications

What we will most likely see AI for in the future will be things like traffic lights for example. In which sensors can calculate optimal times to keep traffic flowing as quickly as possible. 

Also automated cars (self driving) Think of a world where AI talks with lights and everything is integrated into a single system. Now that all sounds great sure. But if something were to go haywire it could cause devastation in no time or completely shut down a grid, traffic etc if say a enemy Ai was able to hack it and manipulate it.

So while there are great things about Ai the more you automate things the more you run the risk of something going out of control and creating more damage. On the flip side it can also help and create less damage so it's a trade off to the extremes of both ends.

# Back To Agents

An important factor to take into account here is that Agent Ai is still rather new and has only been put into place for a few months. While the first versions where rather basic we are seeing growth.

At first Ai agents simply did small tasks for you based on your input commands such as booking you a flight or compiling analytics. But today they already doing much more such as the most recent cases of an Ai not only creating a ton of content and creating a community but also launching and trading tokens. 

This is a rather rapid step forward in a short amount of time and now begs the question of what's next.

I think what we see next are Ai agents that are mainly given goals and then set free and might come back to human input for structure or moves to make next or it could be totally autonomous depending on the person launching the AI. 

What we have to realize though is these Ai's are not really "smart" they are just compiling vast amounts of data and creating predictions or acting in what they see should move them closer to their goal. 

The only issue with that comes the Ai has no real rules and could do other things at some point it's not designed to do if given too much power, information or leeway to grow and expand taking over systems etc.

Posted Using [INLEO](https://inleo.io/@bitcoinflood/ai-generative-whats-coming-next-3ux)