so I took some time and I just went through what are people building with cursor right get a sense tangibly I put a bunch of these on YouTube with links but fundamentally what you see is a pattern is that people are building simple web apps they’re building weather apps they’re building a chat bot app they’re building a Trollo copycat and you might think I’m about to say well what’s the value in cursor then but I actually think that this shows that this is what you get when the cost of software goes to nothing people will build exactly what they wanna build for themselves it doesn’t really have to distribute to be valuable to them it just costs a little to make it’s an afternoon project they could just do it I do think it emphasizes the value of polish and distribution in what you build in a world where software is free so for example when um Amar who’s the head of design 11 Labs put out his weekend project which was also a cursor app but it was super well designed he’s the head of design of course it’s well designed it looks great it looks like something that you would be able to actually pay for it’s a little app that lets you take your voice or voice track and pull it into a video it’s relevant for him and his team on their creator toolset at 11 Labs that’s fine and it’s something that like I could see someone paying for like I don’t know if they intend to do that but it’s so well designed so well thought through it’s professional and I think that’s the piece I’m kind of waiting for as I’m waiting for more professional software that is built very transparently using an LLM assist and there is some out there now I’m curious to see as we get farther along does that look like lolm assisted software for more complex code bases I don’t know the context window for chat is 20,000 tokens right now that doesn’t mean that your entire code has to fit into 20,000 tokens or a couple of dozen pages but if it’s bigger than 20,000 tokens you can’t refer to the whole code base when you’re talking to cursor and that generates the possibility that it won’t remember or be aware of previous elements of the code when it’s making edits of that could cause downstream issues so one of the things I’m waiting and watching for is a good tutorial on someone working using cursor on a very large code base and explaining how they work with it probably there’s one out there I didn’t find it if you found it let me know regardless I think that’s one of the bigger questions how do we handle editing and working in very large code bases with something that has an inherently limited context window because guess what even if they raise the context window to 40K or 60K tokens next week it’s still a limited context window there’s still large code bases it will not be able to handle in one chat window and we probably shouldn’t expect it to cause that’s not how we expect our developers to work so I think my question is are we ready for a world where software that’s cheap and single use or like a few use is something that costs almost nothing to create and professional software is easier to create but still takes the kind of design and product thinking it is always distinguished good software and what is that design and product thinking look like in a world where you can prototype this quickly is the expectation that you prototype it up and you show it and then there’s a software engineering step that comes after we’re still figuring out that workflow I’m curious what you think in the comments let me know what cursor use cases I miss