YC founder thinks AI Wrappers are doomed to fail

[rachel_44]

@iam815518 Instagram is an s3 wrapper

 

[ams59]

@rachel_44 How can an end user use S3 instead of Instagram?

I see no comparison here.

If Adope adds AI features, I will use Photoshop instead of any other AI tool doing this same feature.

 

[rachel_44]

@ams59 End users can use any of the thousands of b2c cloud storage solutions that exist to save their photos and then share links. But Instagram found network-effects in making that easy.

 

[ams59]

@rachel_44 Instagram isn't about storing photos.

Instagram is a social network, not a Google Drive competitor.

 

[rachel_44]

@ams59 Read up on you startup history, mate

 

[ams59]

@rachel_44 It has been a photo sharing app from the start.

Instagram began development in San Francisco as Burbn, a mobile check-in app created by Kevin Systrom and Mike Krieger. On March 5, 2010, Systrom closed a $500,000 (equivalent to $658,300 in 2022) seed funding round with Baseline Ventures and Andreessen Horowitz while working on Burbn.[13][14] Realizing that it was too similar to Foursquare, they refocused their app on photo-sharing, which had become a popular feature among its users.[15][16] They renamed it Instagram, a portmanteau of instant camera and telegram.[17]

 

[rachel_44]

@ams59 Exactly. And what use case did I describe? And what did you tell me it was?

 

[ams59]

@rachel_44 Sharing is not storing, mate!

 

[ams59]

@rachel_44 Plus, most Instagram users are not tech-savvy enough to store images on S3.

 

[iam815518]

@ams59 So, for the Adobe case, I'd say think like this, Imagine I want an animated picture of a kid singing for my website in a particular way. Previously, I had to find someone who knew Photoshop and ask them to do it, or learn it myself, buy the Adobe subscription, and make it myself. But now, I could go to Midjourney and just type what I want. This is what I actually do now. I know Adobe has Firefly, but it's not even close to what Midjourney is capable of. While the incumbents can innovate, they are generally slow, and the law is tough on them. Startups can move fast and get away with things with less scrutiny, even if they break something.

 

[masako]

@iam815518 Where does long term defensibility come from? You can say Instagram is an S3 wrapper, but the defensibility comes from the user base and brand. How does Tome defend its position when other products get better? Jasper was a success story for a bit, then it turned out it had no defensibility. It's a lot easier to say something works now, a lot harder to say it will continue in 5 years.

 

[iam815518]

@masako I agree, it's hard to build defensibility, but I think that's the case with most startups. If you look at Canva and how it sort of took some market share, probably from Adobe, I believe Tome can do the same. Since these software solutions always lower the barrier to entry, they attract more users and change consumer behaviour. I don’t know how Canva did it, but I know Microsoft and Adobe both have their Canva-like software, and still, they haven’t taken off. I think the ease of use might be how they attract users and if they keep innovating and adding features I don’t see reason why they can’t grow.

 

[jl83]

@ams59 They have any reasons to fail simply because they wrap around openai

 

[alej]

@ams59 I think people who have talent to build an app or web, they are just catching the opportunity. I don't think many of them have any goal to stay around long term. But until then they can earn enough maybe fund their another idea.

Nothing wrong with trying out.

 

[erizza]

@ams59 I disagree.

What's exciting about AI is that it allows us to solve solutions in a fundamentally different way, that frankly, incumbents can't really risk steering their ship into.

For example, I am working on a Slack/Teams/Discord communication app with AI deep integrations. The way we are storing and securing data is so fundamentally different, those other companies are too big to change course to how we are doing it.

 

[masako]

@erizza Agreed with what you're saying, though I don't think I'd label your idea as an LLM Wrapper.

 

[davidgarcia]

@ams59 he's just voicing what we all already know

 

[nehemiah126]

@ams59 TLDR: no moat

 

[gohusu236]

@ams59 Theo is just stirring up the pot. Leaning on AI hype. Playing the AI doomer game.

Yeah, there will be a lot of failed AI projects, for number of reasons, but there will be a lot of successful products. It all comes down to unfair advantages in execution that some people have.

 

[erizza]

@ams59 I love Theo, but even if this is his actual opinion, it is weird that the CEO of a data storage company targeting startups is bashing the most popular startup craze.