I share your skepticism of B2B companies making claims of being AI driven! Most stuff I see is a thin wrapper around an existing workflow where the AI isn't adding any meaningful functionality (also see: crypto).
We've been very deliberate about integrating AI functionality in a way that's additive: recommendations, proactive insights, and summarization are all areas that are a leap forward in the data space. What the CommandAI team has done is legit- it's their fastest growing functionality in terms of usage. We'll have more here soon.
literally 90% of AI startups right now are some combination of the above and their "edge" is just a long system prompt that tells them to "be this niche and output response that mimics an expert from that niche"
This was never going to work out because OpenAI is literally watching everybody and how they are using it including prompts. AI wrapper startups serve these functions:
- wrappers prop up NVIDIA which are owned by the people giving VCs the money to invest in wrappers who are long Nvidia too.
- wrappers give OpenAI RSU holders a fuzzy warm feeling and they are able to see what the market is doing with their product so they can pull "Amazon Basics"
- IRS, attorneys, accountants, politicians love them all because they all get to squeeze milk from the fume
While the space is hot and most will fail, I disagree with the general negative sentiment. The same could be said for most/all startups. They all can be distilled into some simple implementation.
> The same could be said for most/all startups. They all can be distilled into some simple implementation.
Hard disagree. There are lots of niches and businesses are complex. If everything was CRUD the people who are using no-code app builders would win against the ones who code, which hasn’t happened despite continuous attempts for 20-or-so years.
Now, I don’t know if GP is right, but assuming that it’s true (many prompt wrapping startups), then yes absolutely they are completely owned if they sit on top of an existing “platform” which can see everything they’re doing. That’s much worse even than selling IaaS on top of aws.
It was the same during the blockchain hype peak – you’d see lots of startups that forked some existing project and replaced the name of a token, and created a website to drive hype.
Exactly. "All SAAS for b2b is just an SQL wrapper". That's until they do something distinguishing, make implementation easier/faster/cheaper or come up with a new technology. Yes the playing field is level now, but what can they do in 1,3,10 years? How will new models play into this?
I have a feeling people are overthinking the whole "we need to have a moat" and really the best moat you can have until you're a multi-million/billion dollar business is your craftsmanship (as silly as it sounds).
Good UI/UX is still the edge as technology becomes more of a commodity.
I share your skepticism of B2B companies making claims of being AI driven! Most stuff I see is a thin wrapper around an existing workflow where the AI isn't adding any meaningful functionality (also see: crypto).
We've been very deliberate about integrating AI functionality in a way that's additive: recommendations, proactive insights, and summarization are all areas that are a leap forward in the data space. What the CommandAI team has done is legit- it's their fastest growing functionality in terms of usage. We'll have more here soon.
Relevant TikTok: https://www.tiktok.com/t/ZTFHap9QW/
literally 90% of AI startups right now are some combination of the above and their "edge" is just a long system prompt that tells them to "be this niche and output response that mimics an expert from that niche"
This was never going to work out because OpenAI is literally watching everybody and how they are using it including prompts. AI wrapper startups serve these functions:
- wrappers prop up NVIDIA which are owned by the people giving VCs the money to invest in wrappers who are long Nvidia too.
- wrappers give OpenAI RSU holders a fuzzy warm feeling and they are able to see what the market is doing with their product so they can pull "Amazon Basics"
- IRS, attorneys, accountants, politicians love them all because they all get to squeeze milk from the fume
While the space is hot and most will fail, I disagree with the general negative sentiment. The same could be said for most/all startups. They all can be distilled into some simple implementation.
> The same could be said for most/all startups. They all can be distilled into some simple implementation.
Hard disagree. There are lots of niches and businesses are complex. If everything was CRUD the people who are using no-code app builders would win against the ones who code, which hasn’t happened despite continuous attempts for 20-or-so years.
Now, I don’t know if GP is right, but assuming that it’s true (many prompt wrapping startups), then yes absolutely they are completely owned if they sit on top of an existing “platform” which can see everything they’re doing. That’s much worse even than selling IaaS on top of aws.
It was the same during the blockchain hype peak – you’d see lots of startups that forked some existing project and replaced the name of a token, and created a website to drive hype.
1 reply →
Exactly. "All SAAS for b2b is just an SQL wrapper". That's until they do something distinguishing, make implementation easier/faster/cheaper or come up with a new technology. Yes the playing field is level now, but what can they do in 1,3,10 years? How will new models play into this?
Way more than just a tarpit.
1 reply →
In terms of unit economics, I think if the price per token goes down (less than $0.00002/1k) some extremely niche companies can survive.
The main problem with those companies is that none of them will have any moat and those businesses aren't defensible.
I have a feeling people are overthinking the whole "we need to have a moat" and really the best moat you can have until you're a multi-million/billion dollar business is your craftsmanship (as silly as it sounds).
Good UI/UX is still the edge as technology becomes more of a commodity.