ChatGPT Store — Ultimate GPT Update You Should Try!

it finally happened the gbt store officially launches we've been waiting for it since this announcement in November and it promises to give us access to thousands of well-trained models and let us earn money by selling our own models let's look at what's coming to Chad BT and a few more exciting things about the upcoming Google bard update and Apple's secret Siri AI so a few days back open AI started sending emails to people announcing the upcoming launch we want to let you know that we will launch the gbt store next week and the next part is interesting because this update actually requires us to do something review our updated usage policies and gbt brand guidelines to ensure that your gbt is compliant don't worry we will go over it all in a minute verify your Builder profile and publish your gbts public now let's take a look at those usage policies and guidelines we don't want our gbt to not get accepted right so what's not allowed these policies look like stuff that you a user can't do but they surely apply app to the models themselves one illegal activity open AI prohibits the use of our models tools and services for illegal activity this one is pretty basic and obvious no one wants their products to be used for illegal stuff further down the line there will be a couple clarifications regarding this illegal activity that I believe hint us at real possibilities of chat gbt and gbts two you can pause the video and read this for yourself we don't want to get demonetized here three generation of hateful harassing or violent content again there is not much to discuss here except the fact that this part of the usage policies talks about stuff that gets automatically blocked if you try to generate something like this and to me this means that open AI knows that there are ways to write prompts in such a way that the automatic system will not stop AI from generating forbidden stuff of course I'm not trying to make people look for this secret prom but still I find it interesting generation of malware content that attempts to generate code that is designed to disrupt damage or gain unauthorized access to computer system this one I find to be one of the most interesting parts of these usage guidelines if you try and read between the lines it would seem like with enough persuasion it's possible to trick Chad GPT and train gpts to generate malicious code I am not sure that we can just ask to write a code for a worm but the presence of this text means that someone at open AI successfully generated a working code for a malicious software virus or something like that and of course you shouldn't even try to do something like this but I feel like this part of the policy doesn't allow creating gpts that will analyze the code of an application file for malware yes there are antivirus programs already embedded in all operating systems but as an exercise for novice programmers or something like hackaton this could be a nice thing to have activity that has high risk of physical harm including development military management or operation of critical infrastructure energy Transportation water again I feel like this usage policy is basically a description of everything that is possible with Cad gbt and gpts would even try using Chad gbt for developing such dangerous things can we just ask Chad gbt to give us schematics for Rocket or can it help with developing its design this all sounds very interesting to me because it also hints at the information that open AI used to train its model I wonder how many secret and leaked documents were used to train Chad GPT but you dear viewer please don't try to do anything like this don't create gpts for the Pentagon or something like that the next one for me sounds a bit ironic fraudulent or deceptive activity including scams plagiarism academic dishonesty plagiarism and academic dishonesty there should be a joke somewhere but I can't figure out where I don't know what's funnier the fact that open AI doesn't want to use its product for plagiarizing someone's work or the fact that they used a ton of materials themselves to train their algorithms because there have been a lot of scandals regarding the information that was used for training from stolen artwork to full articles that can be accessed by writing your prompt in certain way but if we look at this text from a practical standpoint it basically means that you cannot create a custom gbt that will write your homework for you this is very sad but let's close our eyes and pretend we didn't read that activity that violates people's privacy including tracking or monitoring an individual without their consent facial recognition of private individuals classifying individuals based on protected characteristics using Biometrics for identification or assessment again I see this as proof that gpts can be trained to do facial recognition track people work with Biometrics and so on I didn't even know that there was a possibility to trick chbt into doing something like this but apparently there is a way I wonder what happens if we feed our custom gbt a bunch of photos of me and ask it to identify mean group photos

this sounds like a very fun video to try don't you think leave a comment if you want me to try it in one of our next videos I feel like gbts hold many secret and hidden abilities but let's not dwell on the stuff we cannot do and look at how our gpts must look and act in order to be accepted into the store there is a special section about gpts and brand guidelines titled gpts in Chad GPT gpts should have a short name appropriate for display in chat GPT sidebar choose a name appropriate for an app or service not the title of a document or video having a gpt's name and with gbt is discouraged but not prohibited do I need to clarify any of this is just pure branding nothing more a GPT may not use another organization's trademark and its name or logo unless they're authorized to do so Builders must verify their domain in order to use many common trademarks and open AI May reactively enforce this for others if reported this one is more interesting because this basically cuts out all those gpts that pretend to be official gpts of certain brands this means it won't be allowed to create a GPT that's going to impersonate support agent with the goal of stealing private data this is a big win yes the fact that this rule is hidden in branding seems counterintuitive but still additionally names reference in public figures profanity and harmful topics are prohibited if a GPT makes use of third poty services for example as part of its actions these may be mentioned in the description yep you heard it right no more Elon Musk GPT or impersonating someone's manner of speech congrats open AI you made gpts boring of course I'm exaggerating a bit here and if you name your GPT correctly would be allowed in the store but still this just takes a lot of fun out of the platform as for The Branding guidelines there's not much else to talk about none of these two documents say anything about the gpts or the store itself how will it work how will it be monetized what will the developers cut be will we still be limited to around 25 messages per hour I guess we'll just have to wait and see for ourselves so how about we'll look at more practical things let's verify our Builder profile without a verified profile we will not be able to publish R gbts in the store so do not skip this step did head over to settings find the tab that says Builder profile and click on it right now we have two options to verify our profile one turn on our name this name will be displayed in the store and will Mark you as the creator of a GPT you will decide to sell and the second option add a verified domain for example if you're creating a GPT for your company or you work in a company that creates gpts where you just want to redirect people to your website put a link here and submit it will take some time to do all the necessary checkups but eventually you will get your profile verified and if you are serious about selling gpts in the store I advise you to complete both of these steps people will buy one of your gpts then click on the link and go to your website where they will find more gpts after we have verified our Builder profile everything is pretty straightforward you just create a GPT as usual and click on a down arrow on this green save button to publish your GPT in the store you need to choose everyone and click confirm this way you will submit your GPT and start a process of verifying everything after which people will be able to buy it if you want to know how to create your custom GPT be sure to watch our video about it we made a comprehensive and detailed guide for you I am sure the GPT store will be a huge success before it launched there was basically no incentive for talented people to train Advanced gpts but now when everyone can publish their creations and earn money from it the market will explode with thousands of super Advanced and well-trained gbts this can also mean that some of the free gbts that you're using right now could become unavailable anymore especially the popular ones so you better start creating and training your own GPT but the launch of the GPT store isn't the only exciting thing that has happened in the world of AI recently there's been a big leak about the AI from Apple the leaker uk12 has a good track record for the most part so I think we should take this with a grain of salt so on a Chinese website neighbor shared information that the new version of Siri will use generative Ai and even share the name of the internal project revolve around Ai and generative algorithms Ajax while his claims haven't been officially confirmed they align with the growing consensus that Siri is poised for a dramatic transformation a few months back a well-known Apple leaker Mark German said that AI from Apple is coming tech-wise Ajax looks like any other llm to me and right now it's unclear how advanced it could be according to this leaker Ajax will bring a significant leap in series capabilities conversations will feel more natural with Siri seamlessly

adapting to users preferences and context this enhanced understanding will allow for more Nuance interactions making Siri more versatile and helpful if you don't care about Siri and apple you should still care about Apple's AI because it's going to be something totally different from the competition and there are a few reasons for it one apple is super focused on privacy so it will be safer to work with any sensitive data two it will be available only on Apple devices free it's going to be embedded in the system itself and at some point become a core feature of the devices UK 112 also hints at a new subscription service that will unlock even more advanced AI features for Siri this could include personalized recommendations more proactive assistance and even the ability to generate creative content like poems or scripts despite all the uncertainties there are now the underlying message from the leaker and other leakers is clear apple is betting big on AI it may not be as multimodal like Bard or Chad gbt but it's definitely going to have its perks unavailable to other companies maybe this will be the time of Jarvis like assistance with good voice recognition and access to apps because no matter how good chbt gets it will always lack the Deep integration that Apple could do but other competitors aren't resting too with Google ready to launch another huge upgrade to Bard Dylan rousell and non leaker recently posted an X about some new features Google might be added to its Chad gbt competitor he found these details in the code of Google Bar's website one of the headliners in his reveal is an advanced version of Bard dubbed Bard Advanced which could potentially be a premium offering rell's findings showcase through screenshots describe Bard Advanced as a more capable large language model with advanced math and reasoning skills this version is believed to Leverage The Power of Gemini Ultra this is a significant Step Up from the Gemini Pro which pars the current Bard version however it appears that accessing Bard Advanced might require a Google one subscription the cost of this Advanced service within Google one's various tiers remains a mystery rousel also found a feature code n Mato it seems to allow users to create Bots the exact nature and functionality of these Bots are still a bit hazy Rell screenshots show a sidebar labeled your MOS and an option to create threads though he admits I can say what they mean by Bots exactly if they will be sharable or anything but yes Mato is the code name for Bots I wonder if these Bots are in any way similar to my gpts another feature Rell showed is Gallery this feature appears to help users unfamiliar with AI Chad Bots allowing them to explore different topics to see what you can do with Bard it seems to show sample prompts across various topics explaining how Bart works good stuff the more people know of AI the better tasks is another feature rouselle mentioned but its purpose remains unclear it speculates it could be a button to check on Long running processes like generating a photo some kind of a progress bar additionally he noticed a sharing background foreground option which seems a purely customization focused thing I don't think it's worthy of our attention the last feature in rell's Twitter thread is PowerUp this could be a tool for users to improve their prompts kind of like the prompt perfect Chad gbt plugin that we featured in one of our recent videos but don't get your hopes up just yet it's important to remember that all this information is based on speculation at this stage there is no guarantee that these features will see the light of day but the possibilities they hint at are certainly intriguing Google bard more and more starts to look like a better version of Chad gbt even now it can do certain things better than gbt 4 for example sometimes Bard can write much better and more natural text than GPT 4 it has no cap on messages and it's totally free we have actually done a comprehensive comparison of Bard and Chad gbt so be sure to check it out Bard still has a long way to go but at this rate will outsmart Chad gbt in half a year or so and let's not forget that Bard is multimodal which means it was trained in different types of data such as audio text and images this allows it to have a wider range of possible features and capabilities chbt on the other hand was trained on text first and all the additional voice and image related features were added on top of it 2024 really looks like the year of AI with almost every major company investing huge amounts of money into the development of their own models even Apple cannot resist am I the only one excited about all this just think about it by the end of this year we could get a really smart virtual assistant with a deep integration our phones and everything on them we could get GPT 5 B advance and who knows how many more AI

tools and by the way be sure to check out our video about gbt 5's super intelligence thanks for watching and see you in the next one