3 New Free Auto-GPT Alternatives (No Code Install Required)
Now, if you’ve been hearing a lot of hype about Auto GPT just recently, which is basically a Chat GPT on steroids, it has the ability to kind of run on your own machine through a lengthy install process. On my end, you’ve got to install Python and things like that in your local machine, and then you give it tasks which it then goes out and tries to achieve. That includes searching the web and doing all kinds of things. In fact, it creates lots of autonomous agents, which are other GPT instances that try to achieve each goal. So, it’s a pretty cool system, with mixed results online. I’m seeing some people saying it’s the best thing they’ve ever seen in their life; it is amazing. But you know, is it easier to use that than it is for you to be able to do it yourself? That’s often the question that people are asking. Nevertheless, in this video, I’m going to show you three easy alternatives to using Auto GPT, which are all browser-based. So that just makes life a lot easier. You don’t have to install all this stuff, which some people will find a struggle, others won’t, but it’s much easier to just go to a website and then sign up and have all of that capability right there.”
“This is the first one, I don’t know how you pronounce it, but I’m going to call it ‘Omni.’ Okay, omni.com. And I’m going to put links to all of these underneath the videos so you can always go and check them out. This one gives you five free queries a day. Off the top of my head, I can’t remember whether I had to put in my own API key when I set this up, I’m not sure. But what it will do is, you give it a task, you know, an objective as it says here. So I’m going to say, ‘brainstorm 20 trending ideas for a Twitter thread on recent breaking AI news and write to a text file.’ I press ‘research,’ and then it has a process that it goes through where it thinks about how best to achieve that goal. And you can see the progress down here, and it also spells out its thinking, too, which Auto GPT does as well. And it’s quite handy because you can see it working and thinking how to deal with this problem. You can watch its thought process live in action.”
“While that’s happening, we can go over to another one that I’ve been trying out, and you can too. Here it is. This one is called ‘God Mode.’ Godmode.space is the URL for you to go and try this out. This one, I did enter my own API key, I do remember that, although I think you are able to use it at a slower pace with less reliability if you just use theirs. But what this will do is the same as the other one, Omni. You will put in a task, there’s mine, ‘brainstorm 20 training ideas for Twitter,’ and so on. And then it will break down what it has to do to achieve that, and you can see this long list here of things that it’s done and attempted to do. Each time it’s done the thing that it says it has to do to achieve your goal, it will have you approve its plan of action just down here. And that is very similar to Auto GPT because also GPT will say yes or no to each of its steps that it says it’s going to take. The first one here says, ‘Create a Twitter thread from a list of five trending AI topics using the specified API keys and account.’ And then it goes, ‘Search for recent AI news on Google, retrieve the 20 latest AI-related articles from Science Daily.’ And then it will go on and on, now it’s saying, ‘Search for the article, ‘Smart Microbots Walk Autonomously with Electronic Brains’ on Google,’ and it keeps going through, and then it’s basically writing the contents to a text file. It’s just going through all of this stuff. And every time I just click ‘approve plan,’ and now it moves on.”
“But what you can also see here is that it’s got the option for you to give feedback. So if it’s going off down some rabbit hole, and it’s just all getting a bit too wrong for you here, and using up your credits within your playground API key there, then you can just stop it anyway. So, there’s no problem there, and you can guide it down a different path. All of the files that it writes are just down here on the left. So we’ve got ‘Trending AI Topics.txt.’ I wonder actually what would happen if I download that and take a look at what it’s done so far. Let’s have a little look.”
“Okay, look at that, there we go. So it has already kind of achieved its goals. So what it’s doing now is thinking about how it’s going to tweet it all. I suppose potential trending AI topics for a Twitter thread: new method for comparing neural networks exposes how artificial intelligence works, an AI tool for reducing common drug side effects, new programmable materials that can sense their own movements, and some other stuff like even the smartest AI models don’t match human visual processing. So there are some topics there, but those are the kind of top-level topics that it’s come up with. But now I’m guessing it’s going to be turning those into tweets effectively. And I’m thinking at some point, surely it’s going to have to ask me for my Twitter API details so that it can do that. But it doesn’t seem to have done that at the moment, and it keeps trying to figure out how it’s going to post using an API. So it could go on and on. I don’t think, to be honest with you, it’s been that useful for this task. I mean, you know, I think there are probably easier ways to do it than this at the moment. But it’s still fascinating to see how this autonomous GPT system is working out and dealing with problems.”
“Another one that I really like is this one, which is called Cognosis.AI. Again, a link underneath, and it’s a similar layout, similar deal. Here it is. You do have to put in your Open API key if you want. What does it say actually? It tells you why, increased limits. Okay, otherwise you have to use their system, and then you can pay out for that. But it’s free to just try this out, but I’ll just go ahead and create a new secret key, which I’ll just put there and copy that. Don’t worry; I am going to delete the key afterwards, so my account is not at risk from someone copying it and so on. If I go and paste that in and save, there we go, now I’m in a position to use this. I type in the name of the research agent, then the objective. We’ll use the same one: ‘Come up with some ideas for Twitter,’ we won’t use automatic tasks because I think that basically lets it run all on its own and just carries on and on and on without your approval. And that’s it; we can just click ‘start,’ basically. So, it’s having a think about it as well, and over on the right-hand side here, you can see the completed tasks, and you can download them. Now we can see what it’s up to here. So the current task, let’s just move that over here a little bit. Okay, it’s done the first thing then. To analyze recent AI news articles and extract key topics and themes, I will utilize NLP algorithms.”
“I’ll start by collecting a set of recent AI news articles from reputable sources. Once I have the articles, I will use techniques such as tokenization, part of speech tagging, and named entity recognition to identify and extract key topics. Jeez, this thing is smart! Then, it’s saying, ‘uh, use machine learning models to predict which of the extracted topics.’ I don’t know how it’s going to do that.”
“And it’s off! Look at this, generate a list of 20 potential Twitter threads based on the predicted trending topics and write them into a text file. Bam, bam, bam! Done, task complete, and task is shutting down.”
“So there it is, the potential Twitter topics. And then I’m thinking, what you can do here is just click on this, which is going to then download that as a text file. I’d be interested to know what happens when we open that. What have we got here? Let’s have a look at the goodies and see what’s collected.”
“If you’re finding this useful and interesting, by the way, and you want more videos where I just go and test all of these crazy new AI software platforms, then please do like and subscribe to the channel. Here we are! So, understood, and then it’s gone through and explained how it’s done it. Now it’s got all of these topics, which it thinks are going to be really good. Now, I suppose there would be nothing to stop me from taking these or adding another task to this initial prompt effectively, that says create actual tweets now for all of those different topics, but there’s some interesting ones here.”
“Look, the use of AI in agriculture and its potential to increase crop yields and reduce food waste. Something I would imagine, food and then it’s got the use of AI in the fashion industry and its potential to create personalized fashion products. So, some interesting topics, and it did it very quickly and without the need for me to go in through the lengthy install process using Auto GPT. So I think so far, this one’s pretty good. Now, let’s go back over to God mode. This is still thinking about how to go and tweak this using APIs and stuff, so I think this one has definitely just gone off on one.”
“How about Ammoni, Ayomni, a Omni? Let’s call it a Omni again. Um, brainstorm 20 training ideas for a Twitter thread on recent breaking news. Blah blah says the system is making progress towards generating 20 trending ideas for a Twitter thread. The second task of extracting information from the news articles is currently underway. Once completed, the system will move on to rewriting the extracted information into Twitter thread ideas, creating a list of 20 trending Twitter thread ideas and writing it to a text file.”
“So it is doing the job; it’s just that it’s taking forever. I guess that’s just the way that the code’s written or what, but it’s a lot slower than Cognosis, which has done the task quickly and efficiently, so you might want to check this one out. Maybe your results will vary, you know. One thing I do quite like about this God mode one is that it explains its reasoning as well as its thoughts, you know. So you can understand why it’s doing things a certain way and then put it back on the right track by using that additional feedback.”
“Then the other one, which I actually spoke about in a previous video that’s sort of taken off quite well, is this one which was Agent GPT, where you enter your own API key or you go and use theirs. If you want to really go for it, use your own. But what this one does not go online yet, but I do understand that it is on the cards very soon, that they’ll be able to go online and do more as well. But these guys are already going online and doing a good job.”
“This one here says as well, this Cognosis.AI Beta is limited to GPT-3.5, but you just know all of these tools are going to be using GPT-4 as soon as they possibly can because it’s just obviously so much more powerful. So, to its credit, it’s done a very good job if it’s just using GPT-3.5.”
“The other thing I’d like to say is actually about this specific task is that you know with Chat GPT and the browser plug-in and everything, you could probably do this just straight up from there as well. So I see these as really, at the moment, I feel that they’re kind of filling the void for everybody who can’t use Chat GPT fully yet with all of the different plugins that they’re bringing out. But once all of those plugins are out, I think Chat GPT by itself will still be able to do mostly what you would like it to do because it’s super clever. We’ve all seen the demo videos; it is unbelievable with those plugins, and they’re only going to get more and more powerful and better as well.”
“If you’re into Chat GPT and autonomous GPT and all of those different things to do with AI, you’re going to love this group. This is my group; we’ve got over 4,000 members now, so you can go and interact with fellow GPT nuts who just love talking about AI, prompting, prompt engineering, the latest and greatest software. Come and get involved in here; we’ve just had so many people join recently, and there are some really, really good people in here, some very knowledgeable people on all aspects of Chat GPT. You can also access some free training in the classroom there as well. It’s like a collection of all of my YouTube videos and tutorials to really get you up to speed with Chat GPT and all of the cool things that it can do.”
“And then look at all these different videos; there’s absolutely tons: Chat GPT for website building, for YouTube channel growth, all sorts of different stuff. Hopefully, you’ll get a lot out of it. So, that is it for this video. I think we have done a pretty good test of these three on that one simple task. My favorite out of all of these has got to be Cognosis.AI so far. I think it’s just quick, efficient, explains everything really nicely; the user interface is simple, works really well. But there are always going to be improvements with the other ones as well. This one’s still working on that task; it’s just a bit too slow for it to be of any practical use, really, in my opinion. It does say it takes anywhere between 5 and 15 minutes to be fair, but yeah, I think that’s a little too long in this day and age with AI, unfortunately.”
“So that’s it. I hope you enjoyed it. If you did, like and subscribe, and there’s some more videos coming your way any second.”
