> One defining constraint must shape the product... Minecraft is built entirely from blocks. IKEA is flat-pack, self-assembly furniture.
I've been calling these things product primitives. I can't remember where I heard that term, but it refers to things like...
Blocks in Notion. Messages and conversations in Telegram. Frames and layers in Figma. Tweets in Twitter. Cells and sheets in Excel. Tools and layers in Photoshop. Commands in a CLI.
I think what makes for good product design is having a very small number of primitives. A bad product doesn't know what its primitives are. Or it has a very large number of primitives. It feels like everything in the product is some unique thing that works in its own unique way. So users have to learn a ton of different top-level primitives/concepts. It's confusing and intimidating and hard to teach. Ideally you just want one or two or three main primitives.
The complexity/power in an app comes from choosing powerful primitives that have depth, that are composable, etc. You can do a lot with Notion blocks. You can do a lot with Excel cells. You can do a lot with a CLI command. You can do a lot with a Minecraft block. There's depth there.
We used to call this “concept count”. You usually want to minimize the number of core concepts that make up your product. I’ve also heard it as the “nouns and verbs” of a product.
I think this philosophy might be oversimplified. Tana has basically two primitives (bullets and supertags) and manages to be devastatingly complex to use to the point you have to watch hours of tutorials to do very simple things. Conversely Google Maps has a lot of “primitives” but the UX is fairly tight for 90% of use cases.
Doesn't Jira only have one primitive: the ticket.
Everything else just augments it.
You could say that these augmentations are separate primitives, but then the same would apply to all tools in the other cited examples like Photoshop too
I used a similar metric when judging programming languages. The language can get huge but if it's conceptually small, one can learn it and then leave the rest to compounding due to experience. Conceptually large languages had a barrier for me. The case when I felt this was perl.
> Commands in a CLI … I think what makes for good product design is having a very small number of primitives.
Small but not too small. Case in point: shell scripts (POSIX shell, bash) where the scripting part was decided to be modelled as commands thus not introducing another bunch of concepts. We all know what the result is (hot, slow mess).
I know it's in vogue to bash Bash but I feel that criticism is unfair.
Shell scripting is a victim of its own success: it is _so easy_ to get started that most users get value out of knowing the first one percent and never bother to actually learn the rest.
There aren't many who have read the Bash manual, or know what zsh can do that Bash cannot, etc.
"Shell scripting is a hot, slow mess" is the same hot slow mess that you get wherever the barrier to entry is extremely low (e.g. early PHP, early JavaScript/frontend development, game development with a game engine where you can just click around in the editor, etc).
There’s also the fact that shell scripting is for automation of what you may do interactively. It’s not for stuff where you want data structures to manipulate in memory. Trying to use it like python is an exercise in frustration.
> The core tech must be separable from the product
Won't this lead to premature abstraction and application of design patterns everywhere? I mean, sure, of course you should do separation of concerns, keep your business domain layer clean of persistence/network/UI/… concerns etc. But your domain layer will still be very much tied to your product. There is no way around that.
The author really extracted the core tenants of exactly how my former research mentor and I ended up building our business.
We started with the second two points: our core technology was a sampler that enables arbitrary hierarchical Bayesian graph models for sparse data, our constraint was cpu bound tractable compute. The piece that took us the longest to discover was the fact that our end products need to be separate from our underlying technology.
We were given that advice in various words from many people even before we started but some lessons need to be lived to be learned.
One unique core concept per tenant, because the SaaS is nominally multi-tenant but it's really N unrelated websites for N enterprise customers in a trenchcoat
It would have been great if the author could have provided a complete example of the constraints in action, I'm kinda lost on how the third one would look in practice
seems like a made up hook, you come up with something. i like the "everythings a file" idea in linux. you can go a long way with a strong concept like that.
A clear hierarchy may be secured through these constraints. That's the unifier - it'd be hard to achieve these three without it.
A one-pager begs of you to find the foundational value simply - no fooling yourself with a multitude of prospects and complexity.
The separable aspect makes explicit the need to build the foundation to stand on its own. You can't lean on the branches prematurely as if features are solid ground.
The single-defining constraint forces one to conceive and recognize the single-most fundamental functionality - and its shape, and its abilities; its character.
The constraint I'd add for solo SaaS is "can I find one beta user this week." Time, scope, tech all stayed reasonable on my one pager but a project can pass those filters and still die because no one walks in the door. Adding a distribution constraint upfront forced me to validate the audience before I went deep on features.
I love the one page idea. If you can't spend the time to articulate a page worth of constraints, you're going to flail around discovering them as you go. And these aren't "bugs", they're "oh shit we're building the wrong thing" flaws.
I have no hard data to back it up, but in my experience, projects that take the time to put everyone on the same page conceptually (even if it's a 1 pager, high level, here's what we are and are not doing) end up succeeding far more often than projects that wing it. The wing it projects always end up disappointing everyone who had opinions but never bothered to articulate them.
We are trying to design our kitchen for a renovation and I can see how these 3 constraints would be useful for us to do for something more about design than software.
people space, storage space, operations space... everythings a space? too basic? what about space and flow, flow of people, flow of light, flow of food, transitions etc... heh, this is fun
I think one of the biggest constraints is not directly visible to users, but its effects are: HN is built in Arc, a small LISP dialect that for a long time didn't have great performance. Some info here: https://lisp-journey.gitlab.io/blog/hacker-news-now-runs-on-...
stay thoughtful and focused on the tech industry or other nerdy subjects and dont get emotional and make one line snarky comments that immediately attract an avalanche of downvotes and risk getting your account shadow banned.
Well, yes, but no. It didn't matter to the user what it was built around. What mattered was that it was much faster and much more relevant search than what was available at the time.
> The biggest product of the century thus, LLMs, are the core tech.
I would not say LLMs are products. It's still early adopters stage and it's going to be skewed on HN -- a large portion of people here evangelize the virtues of digging through an electronics store's parts bin to finish off a pcb they made in their garage then run an obscure version of linux on it for work. It's a lot of tech kludged together, not a product, not in the context of this article anyway. Same for the current state of LLMs. It's a tech waiting for a product to make it useful for the general population. For developers, Github's copilot is probably the closest, it bundles LLM tech to leverage their tech (github) creating a product you don't have to piecemeal together if you don't want too.
The internet was a tech that was first played with like we play with LLMs now. It was the web browser -- a product that leveraged a core tech -- that made it widely usable. Large parts of the population have no idea the internet is not their web browser (or now apps that access that web through a different interface).
I read a quote from the new Apple CEO on AI, that I think highlights the tech vs product separation and why Apple is where it is: 'We never think about shipping technology. We always think about 'how can we leverage technology to ship amazing products'
I don't get that quote. Does that apply to teams working on SDKs, GPU design, internal tools, etc? Are those all reframed as "products"? Seems like there would be a lot of groups at Apple concerned entirely with shipping technology? "Leverage" makes it sound like they just use other people's tech but we know that's not true. Why throw shipping technology under the bus? I don't get it.
I would guess they label things differently or use different language to talk with their internal product teams/employees. But the end goal being to ship an end-user Apple Product orchestrating tech/products in ways that the individual tech/product could not do on its own.
> One defining constraint must shape the product... Minecraft is built entirely from blocks. IKEA is flat-pack, self-assembly furniture.
I've been calling these things product primitives. I can't remember where I heard that term, but it refers to things like...
Blocks in Notion. Messages and conversations in Telegram. Frames and layers in Figma. Tweets in Twitter. Cells and sheets in Excel. Tools and layers in Photoshop. Commands in a CLI.
I think what makes for good product design is having a very small number of primitives. A bad product doesn't know what its primitives are. Or it has a very large number of primitives. It feels like everything in the product is some unique thing that works in its own unique way. So users have to learn a ton of different top-level primitives/concepts. It's confusing and intimidating and hard to teach. Ideally you just want one or two or three main primitives.
The complexity/power in an app comes from choosing powerful primitives that have depth, that are composable, etc. You can do a lot with Notion blocks. You can do a lot with Excel cells. You can do a lot with a CLI command. You can do a lot with a Minecraft block. There's depth there.
We used to call this “concept count”. You usually want to minimize the number of core concepts that make up your product. I’ve also heard it as the “nouns and verbs” of a product.
I think this philosophy might be oversimplified. Tana has basically two primitives (bullets and supertags) and manages to be devastatingly complex to use to the point you have to watch hours of tutorials to do very simple things. Conversely Google Maps has a lot of “primitives” but the UX is fairly tight for 90% of use cases.
Doesn't Jira only have one primitive: the ticket. Everything else just augments it. You could say that these augmentations are separate primitives, but then the same would apply to all tools in the other cited examples like Photoshop too
Vaguely feels like "Atomic Design" but applied to engineering.
what is Tana?
This I think: https://tana.inc/
Seems like there's quite a bit more to it: https://outliner.tana.inc/learn
I used a similar metric when judging programming languages. The language can get huge but if it's conceptually small, one can learn it and then leave the rest to compounding due to experience. Conceptually large languages had a barrier for me. The case when I felt this was perl.
> Commands in a CLI … I think what makes for good product design is having a very small number of primitives.
Small but not too small. Case in point: shell scripts (POSIX shell, bash) where the scripting part was decided to be modelled as commands thus not introducing another bunch of concepts. We all know what the result is (hot, slow mess).
I know it's in vogue to bash Bash but I feel that criticism is unfair.
Shell scripting is a victim of its own success: it is _so easy_ to get started that most users get value out of knowing the first one percent and never bother to actually learn the rest.
There aren't many who have read the Bash manual, or know what zsh can do that Bash cannot, etc.
"Shell scripting is a hot, slow mess" is the same hot slow mess that you get wherever the barrier to entry is extremely low (e.g. early PHP, early JavaScript/frontend development, game development with a game engine where you can just click around in the editor, etc).
There’s also the fact that shell scripting is for automation of what you may do interactively. It’s not for stuff where you want data structures to manipulate in memory. Trying to use it like python is an exercise in frustration.
> The core tech must be separable from the product
Won't this lead to premature abstraction and application of design patterns everywhere? I mean, sure, of course you should do separation of concerns, keep your business domain layer clean of persistence/network/UI/… concerns etc. But your domain layer will still be very much tied to your product. There is no way around that.
The author really extracted the core tenants of exactly how my former research mentor and I ended up building our business.
We started with the second two points: our core technology was a sampler that enables arbitrary hierarchical Bayesian graph models for sparse data, our constraint was cpu bound tractable compute. The piece that took us the longest to discover was the fact that our end products need to be separate from our underlying technology.
We were given that advice in various words from many people even before we started but some lessons need to be lived to be learned.
core tenets
For some SaaS products it's both. :p
One unique core concept per tenant, because the SaaS is nominally multi-tenant but it's really N unrelated websites for N enterprise customers in a trenchcoat
> Google has Kubernetes
This is more to disable its competitors than anything.
It would have been great if the author could have provided a complete example of the constraints in action, I'm kinda lost on how the third one would look in practice
seems like a made up hook, you come up with something. i like the "everythings a file" idea in linux. you can go a long way with a strong concept like that.
your example also reminded me about the "file over app" concept in Obsidian. this constrain actually limits and shapes the product decisions
A clear hierarchy may be secured through these constraints. That's the unifier - it'd be hard to achieve these three without it.
A one-pager begs of you to find the foundational value simply - no fooling yourself with a multitude of prospects and complexity.
The separable aspect makes explicit the need to build the foundation to stand on its own. You can't lean on the branches prematurely as if features are solid ground.
The single-defining constraint forces one to conceive and recognize the single-most fundamental functionality - and its shape, and its abilities; its character.
Constraints are underrated.
The most elegant solutions typically arise not out of unbounded degrees of freedom, but building specifically with a constraint in mind.
I think that this goes with point 1: composing the one pager helps define those constraints.
The constraint I'd add for solo SaaS is "can I find one beta user this week." Time, scope, tech all stayed reasonable on my one pager but a project can pass those filters and still die because no one walks in the door. Adding a distribution constraint upfront forced me to validate the audience before I went deep on features.
I love the one page idea. If you can't spend the time to articulate a page worth of constraints, you're going to flail around discovering them as you go. And these aren't "bugs", they're "oh shit we're building the wrong thing" flaws.
I have no hard data to back it up, but in my experience, projects that take the time to put everyone on the same page conceptually (even if it's a 1 pager, high level, here's what we are and are not doing) end up succeeding far more often than projects that wing it. The wing it projects always end up disappointing everyone who had opinions but never bothered to articulate them.
We are trying to design our kitchen for a renovation and I can see how these 3 constraints would be useful for us to do for something more about design than software.
I’m gonna go do these…
people space, storage space, operations space... everythings a space? too basic? what about space and flow, flow of people, flow of light, flow of food, transitions etc... heh, this is fun
How big is this page and what's the font size?
What are hacker news constraints?
I think one of the biggest constraints is not directly visible to users, but its effects are: HN is built in Arc, a small LISP dialect that for a long time didn't have great performance. Some info here: https://lisp-journey.gitlab.io/blog/hacker-news-now-runs-on-...
- Just submitted links and threaded comments
- An advanced ranking algorithm
- Moderated contribution and discussion
HTML 4.0
stay thoughtful and focused on the tech industry or other nerdy subjects and dont get emotional and make one line snarky comments that immediately attract an avalanche of downvotes and risk getting your account shadow banned.
> The core tech must be separable from the product
I don't know... none of the examples makes sense for me. Especially:
> Google has Kubernetes
I mean, yeah, and? Google was originally a product built around PageRank, the core tech, wasn't it?
Well, yes, but no. It didn't matter to the user what it was built around. What mattered was that it was much faster and much more relevant search than what was available at the time.
I like these. I have never thought about it that way, but I guess that I generally have the same constraints.
> The core tech must be separable from the product
The biggest product of the century thus, LLMs, are the core tech.
I don't doubt these rules have helped the author, but readers should be mindful when heeding them.
> The biggest product of the century thus, LLMs, are the core tech.
I would not say LLMs are products. It's still early adopters stage and it's going to be skewed on HN -- a large portion of people here evangelize the virtues of digging through an electronics store's parts bin to finish off a pcb they made in their garage then run an obscure version of linux on it for work. It's a lot of tech kludged together, not a product, not in the context of this article anyway. Same for the current state of LLMs. It's a tech waiting for a product to make it useful for the general population. For developers, Github's copilot is probably the closest, it bundles LLM tech to leverage their tech (github) creating a product you don't have to piecemeal together if you don't want too.
The internet was a tech that was first played with like we play with LLMs now. It was the web browser -- a product that leveraged a core tech -- that made it widely usable. Large parts of the population have no idea the internet is not their web browser (or now apps that access that web through a different interface).
I read a quote from the new Apple CEO on AI, that I think highlights the tech vs product separation and why Apple is where it is: 'We never think about shipping technology. We always think about 'how can we leverage technology to ship amazing products'
https://www.tomsguide.com/computing/macbooks/i-interviewed-j...
I don't get that quote. Does that apply to teams working on SDKs, GPU design, internal tools, etc? Are those all reframed as "products"? Seems like there would be a lot of groups at Apple concerned entirely with shipping technology? "Leverage" makes it sound like they just use other people's tech but we know that's not true. Why throw shipping technology under the bus? I don't get it.
I would guess they label things differently or use different language to talk with their internal product teams/employees. But the end goal being to ship an end-user Apple Product orchestrating tech/products in ways that the individual tech/product could not do on its own.
Agreed, but he is talking about limited-scope projects, that he probably does himself. That’s how I work, myself (these days).
In the past, I worked in teams, building much more ambitious projects, and these rules would likely not apply.
The product is automation; the tech is llms