> The moat isn't money for out-marketing your idea that 750 other people are building, it's having a good idea that solves a problem that nobody else is solving well.
This is a very naive take. Any good idea you have can now be cloned trivial in no time at all.
The clones just need to out spend you on marketing even though it is your idea that the LLM cloned.
Which is my point: it seems easy to clone something that seems conceptually simple, but is the result of all kinds of UI, UX, performance, etc optimizations. The reason someone might choose Obsidian over these so called clones is not just marketing, I assure you. The reason people attempt Obsidian clones is that they think creating todo-list management tools, etc is all it takes to implement a PKM, and that that is easy. It is not.
Right now there are kinds of tools I wished existed, that I would pay for, but AI does not automatically provide the insight, good taste, technical excellence, and grit needed to create these products. I could do them, with AI assisting, but do not have the time. It is not a simple matter of saying: Claude, create or clone X, Y, Z.
There are so many of them, aren't there? There's Roam Research (might be the OG one), Logseq (FOSS Obsidian basically), Notion, Emacs' Org-Roam, Anytype, etc. Neovim has like 5 extensions implementing the same idea (such as Neorg), Bram's Vim probably has its own plugins in Vim9script.
> Just on HN alone, throughout 2023 - 2025 we were seing like one new TODO app show up on HN weekly!
This response shows you missed my point entirely. I am saying a todo app does not a PKM make! I'm not interested in a vibe-coded todo app, it is useless to me.
> Because money is the moat, and they have it!
You're sidestepping issue and contradicting yourself. I asked: if cloning is so easy, why has no one cloned the JetBrains IDEs, for example?
Remember, I'm not talking about what happens after the cloning. Are you saying no one has cloned Jetbrains because it takes a lot of money to do so? That would contradict your claim that AI makes it easy.
What's the point you are trying to make here? That it is not possible to trivially clone good ideas?
Let's start with my actual claim - "It is now trivially possible for someone to clone your good idea".
I want to clarify your position: Do you think that the bar for cloning someone else's good idea is now:
1. Harder to do with AI,
2. Easier to do with AI,
3. Exactly the same level of difficulty it always was.
Because if you are arguing that #2 is an incorrect answer, there's no real point in continuing the argument, is there? I'm taking #2 as a given, and you appear to be arguing that it is a baseless assumption.
My point is that "easier" is not the same as "trivial".
Yes, for a skilled and determined hacker, judicious use of AI can enable them do more. That is a far cry from "AI makes cloning complex flagship apps trivial".
Furthermore, the value of software is not primarily in the arrangement of bits. It is about the technical, domain, and contextual knowledge you gain as you develop the software, the understanding you gain about your customers, etc. AI cannot give you that on a whim.
> That is a far cry from "AI makes cloning complex flagship apps trivial".
That is not what I claimed, though.
I said "Cloning your good idea", and context in this thread and this story is not, nor was it ever, about producing Windows 11 or a similarly large and non-trivial product.
It was, IIRC, about small teams (the actual story is about a solo founder) executing a good idea, and then seeing someone with a $20 CC account cloning that product in a week.
This is a very naive take. Any good idea you have can now be cloned trivial in no time at all.
The clones just need to out spend you on marketing even though it is your idea that the LLM cloned.