Do you know someone whose GPTS already has many users? Which categories of GPT do you think will have more commercial potential in the future?
I don’t see any commercial potential in custom GPT’s that are based on just a few prompts or some pdf files.
I think a lot of those that have commercial potential are the ones made by corporations who already have some kind of service they can offer as an “action”
I think your analysis is very reasonable. By the way, what do you think the significance of our ordinary GPT is?
It’s hard to say, but I guess they won’t be that significant on average. I think most of the custom GPTs will only be relevant to their creator and maybe a few friends.
Can i help you? I summarized the basic behavior of the store yesterday.
There are many websites selling AI-related products and services, from sales prompts to hiring.
At this point what we do is to improve the MOE architecture of OpenAI, specializing the gtps, but at a commercial level this does not make any sense, we are just playing with this new technology. Another thing is if you manage the API directly and you have the capital to promote the chat bots through social networks, SEO; advertising, is where you can try to conceive something interesting as a chat bot that replaces a human being in customer service of a company.
The rest goes the other way, chat bots with explicit sexual content are pupulating in the network, but they are created with the API directly, the gpt and the store is currently focusing on improving the architecture and expert modules, that is to say when someone creates an insignificant bot the most benefited is OpenAI and chat gpt.
In the long term and depending on competition, I see an environment where in an airport you see a qr code and scan it and find out the departure and arrival of flights
Remember the golden rule: A successful product is one that people want.
Any kind of commercial success would happen because people are offering a product or service (in this case, a custom GPT) that people want.
It is both incredibly simple and extremely complex. Many people are making GPTs that they want, but not what many people want. Figuring out what people want can be easy, but figuring out how to make what people want is difficult.
I agree with what you mention, but it is not only creating the gpt, it is marketing, social networks, paid advertising to publicize the chat bot, this is not only creating it with gpt builider, because otherwise the caht bot will be lost in the store in millions of gpt.
Here is something that as a gpts creator, I am the one who uses my own custom chatbots, usually I do not look for another, except for dalle, which I still use under the paid subscription.
In short, we are in a revolution and as long as the API is closed, the business will be in the API, buying tokens to sell tokens.
That is true to a certain degree, but at the same time, if one needs to dump vast amounts of resources on marketing and publicizing with little to no return, then the problem might be the product. I have learned this lesson before, and this is not GPT specific.
Of course, getting a product known is always the hardest part. It also can take time. However, I’m a firm believer that if something is good and useful, people will tell their friends, and those friends will tell their friends, so on and so forth.
Let’s also not forget that the usefulness in GPTs come from Actions. You can build your own API endpoint for the custom GPT to use. Is it easy? No. Is it the best way to stand out from the fluff right now? Absolutely.
Ones that are fully-integrated or partnered with existing successful brands. Or, ones that solve a social or medical problem.
I’m a fire safety engineer and I’m working on implementing the fire safety codes and regulations using the APY.
And I find the custom GTPs useful because I can test batches of information before I use the APY.
And sometimes I send links to my clients with a custom GTP and they are very happy to use it. The GTPs are not public, but for some of the services the custom GTP can resolve a 200-300$ problem.
I think that the business model and the niche is very important. I can ask my clients for 100$ instead of 300$ for the private link. And probably I can make more money than leaving the GTPs public.
In that context it does not work with chat gpt builder, it would be working directly on the API, here you put some health in the builder and they will remove you from the store.
I think we are all trying to answer this question.
I used GPT heavily in my own business but it does not translate well in a public product. Customizing GPT requires a lot of data (and fine-tuning in my case) and it’s not evident that there are “public” type data that can become useful products.
I don’t know about everyone else, but I’m asking myself: what good could a GPT be if we add only a handful of RAG items to it compared to the million of things it’s already trained on. So far I don’t find any.
The background to all this is that the creators of gpt with the builder are participants without realizing it in the training stage of fine-tuning the gpt model.
Because the entire API business is focusing on sexual chat bots and that does not improve the model at a general level.
Your example is great and I think it can be a useful tool in many small scenarios. Perhaps more people in the industry need to try and share to make our ideas more active.
I’ve been thinking about this a lot recently.
I feel like there needs to be some unique knowledge that people care about (advice from celebs / influencers maybe?) or expertise that is proven (from verified health care professionals or legal professionals perhaps).
I’m curious to see if the GPT Store becomes a source of customer acquisition for GPT builders (reviews, ratings, recommendations, effective search) or not…
If it doesn’t, then surely most brands / businesses will drive traffic to their own sites where there is an embedded API version that you can easily monetise.
Like if you have to market the GPT, doesn’t seem to make any sense to point users to the version in the GPT store that a) sits behind the ChatGPT Plus paywall b) do rev share with OpenAI?
Yes Jeevika, I’m glad to see your reply again. You are a very thoughtful person.