A Pragmatic Approach to Custom GPTs: Why I Chose Airtable Over Building a Store

I see a flurry of GPT stores emerging. It’s an interesting, albeit somewhat laughable trend, considering the broader context.

Remember when Sam Altman’s talk sent shockwaves across numerous GPT wrapper app companies? It was a clear demonstration of how quickly the ground can shift in the tech world, particularly in AI. I see a parallel situation unfolding with the current rush to create GPT stores.

With OpenAI planning to launch its own GPT store, it’s not hard to foresee a similar turn of events. The arrival of an official store will likely overshadow these independent efforts, raising questions about their long-term viability. It’s a moment reminiscent of the past, yet it seems lessons haven’t been fully heeded.

So, what’s the point in investing heavily in creating these stores now? A more pragmatic and less risky approach would be to use simple tools like Google Sheets or Airtable Bases. These platforms are sufficient for listing and sharing custom GPTs without the complexity and cost of developing and maintaining standalone apps or websites.

This isn’t to undermine the entrepreneurial spirit but to encourage a more calculated approach in a landscape that’s known for rapid and unpredictable changes. What are your views on this rush to build GPT stores? Is it a strategic move or a short-sighted reaction?

In light of this, I’ve taken a more cautious and practical approach. Instead of building an elaborate, standalone GPT store, I’ve put together a simple, efficient directory using Airtable. It’s a temporary solution that makes sense – a place to list and share Custom GPTs without over-investing in a platform that might soon face tough competition from OpenAI’s official store.

Drop a comment to get the link.

It’s a straightforward, no-frills way to explore and share Custom GPTs for now.

13 Likes

Agreed. I got laughed at last March when I suggested to a bunch of novices that OpenAI will follow the same path as Apple did with their platform.

The thing is you must have your ear to the AI ground as it shifts. But, you also must have your mind focused on Niche content that is 100% unique. Copyright it. Turn it into a system that is Trademarkable. Also, if you achieve that then fill out and send in the Legal forms to OpenAI so that you “own” that unique niche data.

As for using Airtable, would you mind if I took a peak at what you’re doing? My assumption is that you’re using it for distribution of content as well as content management. Correct?

Appreciated your thoughts.

2 Likes

Absolutely, FOMO does drive a lot of the rush in tech. I’ve been there too, learning from past experiences. Your insights are spot-on about the importance of unique, niche content and its legal protections. It’s about creating value that stands out, not just riding the wave of existing tech.

Regarding Airtable, you’re correct. I’m using it as a platform for both distributing and managing content. It’s a flexible, user-friendly way to organize and share Custom GPTs without overcomplicating things. Unfortunately, I can’t share links in the comment. Hit me up in twitter? iamvasee.

Some people like to build sandcastles in front of a massive wave :person_shrugging:

Unfortunately I don’t think most people here are interested in using other people’s GPTs. It’s incredibly easy to create, and prompts are straightforward to “engineer”.

So any sort of third-party store will just be essentially drowned with “entrepreneurial spirits” until the official one releases, and the stars can be begin to shine.

Which IMO will be the ones that can connect all the dots of GPTs functionality and not just focus on “prompt engineering”. Same song, same dance when the plugin store was released.

6 Likes

You make a great point about the impermanence of these early GPT stores. It’s true, the real value lies in integrating GPT functionalities, not just in prompt crafting. My Airtable base is a temporary space for exploration until the official store sets the stage for more groundbreaking innovations.

3 Likes

Yes. During a gold rush, it’s better to be in the business of selling shovels than searching for gold. Being able to gather, and compile datasets for retrieval functions / API function calling I think will be the winner when everyone and their grandmother is creating GPTs.

Great idea. Hopefully you stay along for the ride here so you can actually post it :laughing:

1 Like

With a flood of Custom GPTs (and other content) it will become more about discoverability - how to separate the wheat from the chaff, so to speak.

I’m curious about what discoverability features (search, etc.) will be available in the official Custom GPT store on ChatGPT Plus. Any ideas?

6 Likes

That’s a fun thought!

Knowing OpenAI there will be a GPT to help you find it :laughing: It’s GPTs, all the way down. I’m excited for when I can hold a conference with a selection of GPTs. Have them argue about future prospects of Dream Land Corp.

The only metric we have right now with our GPTs is to see how many times a conversation was initiated, so maybe that will be the sorting method? :person_shrugging:

1 Like

Well, I’m just naming mine…

AAAAAAAAAAAAAAAAAAAAAAAAAaaaaaaaaaaaaaaaaaaaa11111verygoodGPT!

So I’m sure I’ll be top of list! Small smile.

Seriously, though, I do think it will be interesting. You’re likely right on there being a GPT discovery tool maybe.

3 Likes

LOL

Dammit. Took my name.

If they actually make it sorted alphabetically again I may cry and yell at clouds (more than I already do)

I seriously cannot stop laughing at this. LOL

1 Like

Hahaha. Yes hopefully. It’s just something I thought should be helpful for a lot of people. I hadn’t realised people would start building stores.

1 Like

I’ve been thinking about it ever since the announcement. It is so simple to create a GPT, what’s stopping people from flooding the store with useless apps. At that point, you might aswell just use the ChatGPT and not a CustomGPT. And I have no idea how its go

2 Likes

This is something I’m trying. it’s called “ExploreGPTs” Hahah. I’ll find a way to connect my Airtable Base to this GPT.

2 Likes

Nah, I think they learned their lesson with the beta Plugin store… Let people do what they will do and watch and note. Hopefully this next iteration will be more robust and use the technology itself.

I’ve been thinking a lot about book discoverability for a few years now. There was once a project known as Book Lamp (?) - the Book Genome Project. Apple bought them out and the tech disappeared. It was basically a system that would allow you to easily recommend similar books based on the books “genome” pieced out by machine learning… I haven’t heard anything from Apple since they gobbled up the company, but I’m sure Amazon and others would be interested…

2 Likes

Interesting. I wonder if they will also introduce some sort of recommendation engine. Only time will tell!

Meta, I love it

1 Like

I haven’t heard anything from Apple since they gobbled up the company, but I’m sure Amazon and others would be interested…

sometimes i am in situation in a game where i have a power weapon but the scenario is very unfavorable. the enemy team closes in. if it’s an option i’ll jump off the map. could have been nice to create some chaos with that weapon but i’m sure not going to let it be unleashed on my team :slight_smile:

1 Like

Heh. Yeah, what I figured - hide it. If they would use the technology, though, they could give Amazon Kindle a run for their money, I bet. Like I said, with generation tools here now, the markets are being flooded, so discoverability is important - not based on sales or reviews (which both can be faked), but an LLM looking at the text of the book and reviews and finding books with similar “genomes”… we’re close… I think Book Lamp was 2013 maybe? Way early…

I’m thinking the OAI “official store” has more gravity because of all the ChatGPT subscribers.

But it would be relatively straightforward to create an alternative platform with different GPTs, built on the already more powerful RAG systems, that would absolutely BLOW AWAY anything in the OAI store.

The problem? Really nothing, just put in the work to show value. :rofl:

But what value is the OAI store offering to the GPTs developers? Well, I think revenue sharing for one.

So really all it takes is to create another store, with more powerful and useful bots. The GPTs are cute, but easily blown away from what developers can create now.

At first, I was wondering what market the GPTs were going after. At first, I was thinking: “Aww cute, you can easily create your own personal bot”. So this means direct to consumer … OK, I get that, it’s what ChatGPT does. :+1:

But then I see this “store” and “revenue sharing”, and it looks like your bot is currently constrained to limited resources, and now I’m thinking WTF??? This will easily get destroyed in a the marketplace when the Goliath behemoth bots, ones you can create as a developer, get “discovered” by the public. :interrobang:

So, to wrap this up, I don’t think the store will currently compete with already mature bots, but it’s more of a showcase of “look what I did” with GPTs. So it’s a nice little shelf of accomplishment within that ecosystem, which would (maybe) drive others to create bots in that space too.

But like I’ve said on other threads, the bigger threat is consumers seeing that GPTs are “good enough” and not see the value in your bot(s). So how to counter this? Marketing basically. :rofl:

3 Likes

I (mostly) disagree, sir.

There’s very rare (if any) instances where people go elsewhere than the official landing page for an app store. Plugins don’t count (lol). Convenience is king. It can be two clicks and I have loaded a GPT.

OpenAI’s GPT ecosystem is becoming very complex and nuanced. We’ve seen how fast they are branching and moving. Any third-party will be playing catch up & reacting to changes while OpenAI will release their updates (hopefully) already aligned and tested.

The way I see it, is the beginning of building blocks. Their RAG system is a start. Very, lacking. Yet they are going to be continuously reviewing the data and adjusting. This goes for CI, browsing, Whisper, TTS, GPT-4V takes breath and finally Dall-E. Not only are they going to be improving them individually, but also working to harmonize them all together.

Anyone and everyone can be a developer now. Which is incredible. But I think it’s more important to think of the consumers.

I can sell a daily ingredients-to-your-door, attach a GPT and then feed it the daily cooking options for a homeowner to talk to and help them cook their food while they also speak to other GPTs. They can also goto my website and watch a video while talking about it (add dimensions)

Right. Like they can literally scan an NFC or QR code and bam, they are talking to the GPT that already knows who they are, and their ingredients and a conversation can begin

I see GPTs being the future of personal assistants. They are going to tie in with Home Automation. So I can wake up, ask GPT what my schedule is today, what was delivered and what are the options for cooking.

I truthfully don’t care about the revenue sharing. I think it’s just OpenAI’s way to prevent nasty third-party providers offering money by injecting ads and most likely selling information

3 Likes

True. There is a moat. But the store is in its infancy, and the moat is a little trickle, and can be easily crossed right now. Mostly talking now, not in the future if that helps add more context.

Agree, the moat will get deeper and wider over time.

Yes, I can see the potential, PA’s are going to be huge. But competition will drive this. I’m now looking 5 years out :rofl:

But all good points. :+1:

1 Like