OpenAI is out for all wrappers - I said it, now it's happened

[birch]

OpenAI has now made it possible for ordinary people to create wrappers.

I said it here

https://www.reddit.com/r/ycombinator/comments/17n3i52/ai_wrappers_lessons_from_twitter_api_post_elon

- that it will happen. I thought openAI will charge more, they didn't, instead they opened the market to the general public.

With one blow, openAI has killed all wrappers (such wisdom). And now, YC's $500k for both 2023 and most of W24 is truly down the drain.

You know, I was going to integrate Medisearch API into my business, now I'll just do it myself.

I was on ion's early access list, now I'll just turn figma to code myself.

I loved Artisan AI, was going to pay for a worker, now I'll just train my worker myself.

And lots more like that....

Meanwhile, we should realise that openAI has not fully killed the market, cos we have over 8billion people on this plannet and only about 1% is technical. What openAI really did is to open their AI to general competition. Which means these startups don't deserve $5k not to talk of $500k investment.

So I think openAI really targeted investors like YC not the startups themselves. If they really wanted to kill the startups, they could have charged crazy like Twitter for their API, instead they made the startups worthless by exposing their secret to everyone, dragging investments into the mud.

The startups will survive, but simply won't make as much money as they could. Any company that would have subscribed to these wrapper startup services will just implement it once and own it forever.

Meanwhile, openAI will still make as such money as they could, even more, now that ordinary people can create their own wrappers. All while making sure that no wrapper startups is milking their hardwork.

Now I guess YC will focus more on nonAIs.

 

[heisalive1]

@birch I’ll say it. I’m glad this happened. There are only so many PDF readers or marketing AI or you name it that developers can make that are differentiated. This will force developers to be more creative and that’s never a bad thing in an oversaturated and trending market.

 

[tim53]

@heisalive1 When you say Pdf readers, would you consider a company like ChatNode to be a pdf reader?

 

[heisalive1]

@tim53 Never heard of it but based on the headline on the website it looks like a PDF chatbot.

 

[timmdogg]

@heisalive1 “Look I made a whole company in 2 weeks and we have a product launching tomorrow”

Yea I’m glad too.

 

[jesuspleasesaveme]

@birch meh, you act like it’s zero sum and that startups don’t evolve. that’s why you invest in founders who can build and launch, fast. on to the next.

https://preview.redd.it/bgxjfjcs3yy...bp&s=cc10fb0a4e9a7b5f0c513bf1c6a0a80b2249b439

 

[grubberg]

@birch oh openai doesn't target anyone. They just follow their power users and they listen to their feature requests. Most of the time there are already some startups addressing a particular missing feature. If it's your company - well, too bad for you, because you are just an ant on openai's way. They will steamroll over you, not because they hate you, but only because they follow the needs of their users. Ps - I'm one of the startups that was steamrolled too

https://www.reddit.com/r/ChatGPT/comments/17ptif4/post_mortem_for_glome_social_a_wrapper

 

[divanir]

@grubberg If you try to capitalize on the trajectory of a single company, don't get mad when you get steamrolled by it

 

[marissavchon61]

@birch IMO, not a lot of small to medium size companies will write their own wrappers. Doesn't make sense to have employees work on building and maintaining something which they can subscribe to for $1k/mo.

 

[mmntc]

@marissavchon61 i think you're misunderstanding that it's not about them being willing to write or not write their own wrappers causing these startups to get hit, it's more so that other devs and possibly non-tech people will be building these gpts and putting them on the openai store to make money.

sooner or later, there will be so many gpts on the store that it'll be hard to not find something that'll suit your needs, removing the need of going outside of the ecosystem to find a Saas solution. most of the companies will probably be on an openai subscription already, so why not just stay there instead of incurring another monthly subscription. I mean you literally have one subscription and access to millions of gpts. kind of like having a netflix for your saas needs

i actually see people on the internet offering to pay someone to create a gpt to suit their need becoming a big thing, causing a whole new 'job market' or 'occupation' if u wanna call it that

 

[pmsh]

@mmntc I 100% agree with your logic

 

[writerhomework]

@mmntc Anyone can always come up with a reason it won’t work right now, but they always dismiss the trajectory. OpenAI did this in a matter of months, I genuinely don’t even know what 2 years in looks like, forget 5 years.

 

[birch]

@marissavchon61 You think? Or they can pay $10k for it once (even less soon enough) and own it forever.

The future of AI is 'open'.

 

[marissavchon61]

@birch With that logic, most B2B SaaS shouldn't exist.

It's not always about the money spent. Building something which is not your core business doesn't always make sense.

 

[birch]

@marissavchon61 So I agree that any wrapper that really provides true value other than simply wrap AI in a specific service will survive.

But majority of YC's AI startups are simply wrappers. Meddisearch simply tried to make chatGPT answer only health questions, still if you ask it nonhealth question, it will answer.

Many more like that don't deserve $5k not to talk of $500k.

 

[arustadb]

@birch Why not? You are looking at the technology as the core of the investment. But that’s not how it works. The investment is into the business not the code.

 

[dominic12]

@birch This is just unbelievably wrong with any business. Of course they subscribe vs. internal development or consulting.

 

[jsj123]

@birch I doubt they’re “out for” all wrappers. It seems like they’re doing a logical progression, if they just wanted to be an API then they wouldn’t have launched chatgpt.

 

[aloraz]

@birch I don’t see how this kills all wrappers. For the lazy ones that only provide a nicer UI sure, but a lot of value is with integrating AI into specific ecosystems that can’t be done without coding. Maybe I’m leaning too much on convenience as a value add, but web builders didn’t kill web dev and no-code tools can only take you so far.

 

[pbentoncron]

@aloraz Yeah-- this shouldn't really impact those that have adapted to the actual workflow of the customer/niche and made this work in the workflow vs just being another chatbot.