OpenAI is out for all wrappers - I said it, now it's happened
- Thread starter birch
- Start date
-
- Tags
- startup y combinator
@birch You have some logic, and I agree that wrappers are not the best thing to build however, startups are about providing value and then scaling it, and wrappers can help do this faster
99% of the time what a startup does in the early stages is not even what the founders want to do, it's just a step to get into a strategic position where they can make validated actions, and the only way to get there is to make something small that people want and design your startup in a way that can grow to your big idea.
YC will buy your big idea, not your first step.
We don't know what the real plan for their startups is, we only know what they have publicly chosen to say, the real thing is in the founders' minds and probably in their YC applications
like you I can't see why this startups can be billion dollars companies, but I also don't expect them to see this in my small idea (the one that I tell to everyone who is not an investor)
99% of the time what a startup does in the early stages is not even what the founders want to do, it's just a step to get into a strategic position where they can make validated actions, and the only way to get there is to make something small that people want and design your startup in a way that can grow to your big idea.
YC will buy your big idea, not your first step.
We don't know what the real plan for their startups is, we only know what they have publicly chosen to say, the real thing is in the founders' minds and probably in their YC applications
like you I can't see why this startups can be billion dollars companies, but I also don't expect them to see this in my small idea (the one that I tell to everyone who is not an investor)
helarikoski
New member
@fsheppard But what will you even build? I only see value in going open source, contributing, and then building solutions that the GPT-4 model is bad at.
Open AI has the most advanced model out there and their priorities for who gets access to it is: Microsoft, Internal Dev teams of Enterprises and then traditional devs who want to build on top of the API (treating it like AWS essentially).
Open AI has the most advanced model out there and their priorities for who gets access to it is: Microsoft, Internal Dev teams of Enterprises and then traditional devs who want to build on top of the API (treating it like AWS essentially).
@helarikoski Why are you assuming that they will build everything using Open AI APIs
Doing this is wrong, I agree, but that doesn't mean that they shouldn't do it specifically at this stage, if this is the fastest way to provide value then this is the to go, if the founders are as good as they look like they will figure out how to deal with this challenge later.
Doing this is wrong, I agree, but that doesn't mean that they shouldn't do it specifically at this stage, if this is the fastest way to provide value then this is the to go, if the founders are as good as they look like they will figure out how to deal with this challenge later.
From openAI -
We’re rolling out custom versions of ChatGPT that you can create for a specific purpose—called GPTs. GPTs are a new way for anyone to create a tailored version of ChatGPT to be more helpful in their daily life, at specific tasks, at work, or at home—and then share that creation with others. For example, GPTs can help you learn the rules to any board game, help teach your kids math, or design stickers.
Anyone can easily build their own GPT—no coding is required. You can make them for yourself, just for your company’s internal use, or for everyone. Creating one is as easy as starting a conversation, giving it instructions and extra knowledge, and picking what it can do, like searching the web, making images or analyzing data. Try it out at chat.openai.com/create.
We’re rolling out custom versions of ChatGPT that you can create for a specific purpose—called GPTs. GPTs are a new way for anyone to create a tailored version of ChatGPT to be more helpful in their daily life, at specific tasks, at work, or at home—and then share that creation with others. For example, GPTs can help you learn the rules to any board game, help teach your kids math, or design stickers.
Anyone can easily build their own GPT—no coding is required. You can make them for yourself, just for your company’s internal use, or for everyone. Creating one is as easy as starting a conversation, giving it instructions and extra knowledge, and picking what it can do, like searching the web, making images or analyzing data. Try it out at chat.openai.com/create.
kobidobidog
New member
writerhomework
New member
@kobidobidog What the hell do you think a wrapper is, not fine tuning? 
kobidobidog
New member
@writerhomework 
. Your output of anywrapper depends on the data you train it on - fine tuning is not just giving it exmPlea in the prompt. It’s loading 2K+ examples of quality training data
I doubt more than 1% of people building wrappers will do this
. Your output of anywrapper depends on the data you train it on - fine tuning is not just giving it exmPlea in the prompt. It’s loading 2K+ examples of quality training data I doubt more than 1% of people building wrappers will do this
writerhomework
New member
@kobidobidog Maybe, even if so then only for now. Training will become 10x easier sometime soon. I can already load up 1000 examples in a google drive folder and “train” gpt locally, may not be exact same as training the way you’re thinking but point is we’re not far off as we stand.
Oh and if you’re building a wrapper and not training what the hell are you building.
Oh and if you’re building a wrapper and not training what the hell are you building.
kobidobidog
New member
@writerhomework You’d be surprised what people are willing to pay for
@birch I'm just a random person who is so far away from YC and the rest of this "scene" they might as well be on the moon. But I have been a software engineer for a long time and have started a successful software business, so I feel that's "good enough" for a random comment here ...
As unfortunate as it is that these startups failed, that is supposed to be a major element of what venture capital firms do. Risk management, risk analysis.
Someone pitches you a product idea tied directly to a major company's API, with the intentions of that company unknown and motives unclear? So with one swift move they can destroy the whole concept and just do it themselves?
I, the unsophisticated investor that I am, would still see that and put that well into the"very high risk" "extremely high risk" "absurdly high risk" category, next to the "don't even think about it" category.
But, I'm not a VC so maybe I just never got the memo with the details.
Instead, I see hundreds of thousands of dollars being invested on one side, and then the other side comes along and does the thing that made the investments risky.
It wasn't shocking, quite the opposite ... it seemed perfectly rational. Not something the startups wanted to happen, but must have been at least on the radar.
OpenAI being a business, that's usually "money is good" and you want to get more of it. So they ran some numbers, they turned out good, and so they did it.
I simply don't see why this is surprising, even shocking ... and the resulting drama.
As unfortunate as it is that these startups failed, that is supposed to be a major element of what venture capital firms do. Risk management, risk analysis.
Someone pitches you a product idea tied directly to a major company's API, with the intentions of that company unknown and motives unclear? So with one swift move they can destroy the whole concept and just do it themselves?
I, the unsophisticated investor that I am, would still see that and put that well into the
But, I'm not a VC so maybe I just never got the memo with the details.
Instead, I see hundreds of thousands of dollars being invested on one side, and then the other side comes along and does the thing that made the investments risky.
It wasn't shocking, quite the opposite ... it seemed perfectly rational. Not something the startups wanted to happen, but must have been at least on the radar.
OpenAI being a business, that's usually "money is good" and you want to get more of it. So they ran some numbers, they turned out good, and so they did it.
I simply don't see why this is surprising, even shocking ... and the resulting drama.
@betsy789 It's basically a glorified affiliate marketing funnel. Somehow some stakeholder thinks there that they can can find a place in the market as opposed to "Another-generic-Chat-GPT-enhanced-UI.html
", they market it as:
"Copywriting helper chatbot for single moms and dads in Florida
"
"Medix - Answers to medical questions using ADVANCED ARTIFICIAL INTELLIGENCE
"
"Better Help - A psychologist at your fingertips"
...
But also some things can only be marketed if you have a good UX for it. Writing code within your IDE that recognizes the context, as opposed to having to copy and paste and explaining over and over again to ChatGPT what you want.
ChatGPT doesn't help you structure your process and the UX is just a chat window. Look how excited people got when it was able to click around the web for you, or put a code sandbox to execute python it generates. As opposed for you having to copy and paste that somewhere and having to install the packages, run in an environment, etc.
", they market it as:"Copywriting helper chatbot for single moms and dads in Florida
""Medix - Answers to medical questions using ADVANCED ARTIFICIAL INTELLIGENCE
""Better Help - A psychologist at your fingertips"
...
But also some things can only be marketed if you have a good UX for it. Writing code within your IDE that recognizes the context, as opposed to having to copy and paste and explaining over and over again to ChatGPT what you want.
ChatGPT doesn't help you structure your process and the UX is just a chat window. Look how excited people got when it was able to click around the web for you, or put a code sandbox to execute python it generates. As opposed for you having to copy and paste that somewhere and having to install the packages, run in an environment, etc.
clajsmith100
New member
writerhomework
New member
@writerhomework I thought so too. But have you seen their employee count growth?
Only last month they increased their headcount by ≈ 20%
(excuse the Portuguese labels, but it's from linkedin)
https://preview.redd.it/jgxok3m2i1z...bp&s=a751038ea85713033cc199f30453d01865a643c2
Only last month they increased their headcount by ≈ 20%
(excuse the Portuguese labels, but it's from linkedin)
https://preview.redd.it/jgxok3m2i1z...bp&s=a751038ea85713033cc199f30453d01865a643c2
writerhomework
New member
@martyrs44 Well yeah they have to spend the $125m they raised. I could be wrong too overall, will take a bit to see their downfall if I’m right though
janellhanery
New member
@birch https://preview.redd.it/4szb6e7jkwy...bp&s=e92b8c48872a9d6e54bed0bac1215b691deaef4d
Called it already 3 days ago on Twitter.
Called it already 3 days ago on Twitter.