The Rise of Product

Essence of business/theory of startup: just because you are annoying asf on the internet or you raise money, does not imply there is a business in the scope. let’s break down the essence of business so we’re on the same page: Solving for distribution: unless you are a celebrity, the children of celebrity, famous like lebron, the princess of king henry the 3rd, daddy’s special child with trust fund money, you actively need to solve for distribution itself. you might need creativity, you might need to be clever, you may lead to nowhere, and dig a fat hole—but for now let’s NOT talk about social media because its trivial and nobody likes reel creators or larpers. Yes, it could work but can we do better. naive: we can and always can try implementing manually loading what you think is ICP for A/B testing… The caveat here is you can’t fail the Mom test by asking irrelevant people(non ideal customer) bcuz they are irrelevant…you shouldn’t be caring about the feedback of anyone but whoever it is you are trying to help. an example: the founder of blind went straight to seattle(home of socially inept software engineers) to host bbq events that required work email forcing users into his app that didn’t exist, which eventually through many drinking nights later formed a community on blind. A heart-warming and real founder story. but you also don’t want to hear “build in public” for the millionth time and not everybody wants to be annoying online. can we do better than naive strategy? is there preprocessing we can do? Are there other people, or companies that can offload distribution for us? Instead of data structures improving runtime, are there channels you can abuse to leverage the running time on distribution? A lot of ideas might just be naive... a flyer in the amazon building hung up is prolly worse than linear—it’s a monte carlos process where it’s not even guaranteed when someone sees it, they’ll care, let alone that they were the right ICP. how can we guarantee ideal customer? how can we maximize conversion? how do we improve our worst case bound on distribution? there are lots of other companies, organizations, etc that may have already preprocessed the burden of acquiring your ICP. the right partnerships would be analogous to knowing or choosing the right data structures for your given problems. however, taking a step back before distrubtion, we need a well sound hypothesis or thesis. Let’s say you’ve acquired some overwhelmingly strong conviction about a problem space, group/type in the market, painpoint within others etc. first of all you are wrong—since you are not god capable of predicting the weather or the future. So let’s say: We now have a business idea, and with that you have to prove it. The tools to prove your idea are through distribution or talking to customers directly or etc etc. but its not rigorous, you are best case guessing why people would care and throw money at you… More likely than not, you should have a series of hypothesis about your ICP or have a written/mentally written thesis about the problem, why is it worth solving(why would other ppl pay for this). how to solve is irrelevant for now, that’s a product issue. We have our hypothesis and we use our distrubtion that you should’ve already cracked, to then seek to justify if some hypothesis about the market were stronger than others. As we lean into a stronger hypothesis it forms our thesis about the market or about the problem with our ICP, which will come into play to shape product. Assuming there was problem worth solving, there was something that could be mapped to a pain point; all goes well some small fraction may pay for the product(only in theory). This phase of 0->1 is by far one of the hardest in startup theory, since it will either validate pain, or continue to confuse you on if previous hypothesis were well-typed. it may be ambiguous, or up for judgment, which is gonna be awkward if the team lacks expertise in the field, shrugging at whether the problem exists or if it’s worth anything. couple months have gone by at this point, and you have either dialed/validated customer painpoint or you should just fucking give up rn. (we cannot get to the point of having ICP sending you money online without confirming problem) Solving for product: At this phase, it should be clear and evident via somehow the market signaling “yes your problem space exists and you can reach ICP”. when it is time to think about product, do not think about how to build it. Its rushing we need to create mapping between what the hell was the problem, where is the friction currrently, and how is the user experience going to be less painful such that they will get to addressing their problem? we don’t need any buttons or authentication or stupid little features. literally, where is the friction currently with how ICP is trying to resolve their issue, and where/why would their problem be addressed through you? it’s a conceptual map of what’s going to happen / expected outcome. Now that in theory, the problem will be addressed and there is less friction, ie your idea has some moat / competitive x,y factor to what customer could be doing instead— we could potentially now think about its minimal viable build or minimal lovable product, whatever buzzword you like. but do you see how much theory or actual thinking of a soln or implementation come before claude code? how much due dilligence is actually required to verify if the problem you are working on is even correct? it could take you years until you realize that the problem itself is faulty…. and to not realize how much thinking should go before building is failing design thinking… the approach towards hypothesis—>thesis—>problem—>soln should be modular/circular in nature, meaning you should be backtracking to previous states if you can’t move forwards. or throw it all away and pick a different field if you realize you are incapable. Modern Day Society:---------------------------------------------------------------------------- “I build viral app over a week” “Watch me vibe code a startup” “insert bs clickbait” Okay, so now we have 100 new useless apps every week on the appstore, or 100 new useless ideas that will never sustain the market because people think building is the same thing as value….. It's not. Building something without understanding the why, is just the modern tech form of masterbation…You’ve pleased yourself through your own thoughts, and completely failed to please or validate other people in the real world…. how selfish …a receiver and not a giver… Evidently, building is the wrong intuition… figuring out how to please others(customers), what, and why behind decisions—are the guard rails abstracting “whats on agent todo list”. Vibe-coding is just attention porn—the tiktok for developers—the gooning for the tinkerers. We don’t need more tinkerers and builders, because who asked? Raise your hand if you care. To not understand product: The great and vast illusion of productivity where millions of tokens get used in parallel, yet nothing is actually accomplished. 100 new features, 10,000 new lines get pushed, yet no additional change in who is actually affected. A technician who builds for the pleasure of building, rather than an entrepreneur who really knows their audience, building a business with some meaning or purpose. Or rather—the technician who doesn’t understand people or products… So what the hell is “product”? Another word that gets thrown around in corporate speech, yet nobody actually knows what it is….like when foreigners get chinese phrases tattooed—it sure looks like it means something, but we don’t actually know what it means. [insert and think of explanation here] The skill is not about speed, or output—It is judgement. A deeper level understanding of the world, one that may come from adjacent fields irrelevant to engineering. What problems actually matter in society. What ideas are worth pursuing. Who’s paying for this and why the hell. At its surface, startups have nothing to do with engineering. Vision has to do with understanding people… and I am not in the slightest convinced that anybody can just obtain judgement. “We gamified coding and made it addictive with strong instant gratification. They are the new casinos, tokens are the new chips we buying to play, to have a sit at the table. But then again, many gamblers play for the thrill of playing not to win. The same is true with coding now, is the thrill and confirmation you get when Claude tells you "your instinct is right, that's a killer feature” Says an anonymous internet thread commenter. [] The Rise of Product Thinking------------------------------------------- You walk into his/her room, 90s rnb and rock is blasting in the faint background, 15 tabs of various agents are running–each attempting to verify if the other agent is correct. Millions of tokens are flying around being consumed and you watch dis mfer doomscroll while the agent completes all the work… is this the flow state you dreamt of? … lame hook huliu, needs work. I recently read some “decades in tech” articles (random engineering managers who yap on substack that I tried reading) on their takes for staying grounded in a market that wishes to reduce its entry-fish roles. One take being that if AI democratizes technical or building throughput(reduced barrier for task completion), employee ability to articulate / communicate(the soft skills of engineering) rises along with personality/culture fit. Not too sure honestly, seems a bit black and white(feels a bit clickbaity imo), but I do see the fundamentals being so abstracted in the next decade or two, it becomes 1-to-1 of PM’s who articulate requirements and army fleet of agents completing (w/ human interventions just to verify for correctness). I don’t think any engineer will agree ai replaces need for swe entirely in the next two decades, but I do want to comment on how I think it could shape the market as a whole. I’m going to use fashion as an analogy–where all the money spent in the world can never give someone taste. And yet other pieces are so timeless within the fashion economy they can never actually die with the trends. It almost feels the same with computer science or math—it doesn’t matter what is in trend/style in a given era, what’s timeless are its fundamentals. Similarly, you also can't really notice items that are “timeless” in fashion, if you are someone who is drowned by the noise of “is it in style”... a.k.a you need to know the fundamentals to know what is timeless. That’s why lots of people think “timeless pieces” are ugly, they just have zero taste. My point is it kinda feels the same in technical fields. How will you make logic, problem solving, or design thinking, etc, go extinct in the next thousands of years? they are just a concepts, or a way of thought. All the ingredients given to someone who can’t cook won’t produce a good dish. All the tools and most advanced building assistance in the world won’t dictate if it's “actually correct” or that “people want this”, “this is as desired”, etc. You can fundamentally never replace correctness itself(its a concept). You don’t need to be making ramsey beef wellington to ruin a procedure, you can fuck up even lobster congee with hundreds of dollars spent on ingredients. While master chef junior with cheap walmart ingredients has dishes turning out better. That would be “expertise”, or “knowledge”, I digress. Even in the next thousand of years, people still want to know if you can thrive under ambiguity, if you can make well-sound logical decisions(increasingly more important the more leadership is involved), if you can discuss trade-offs(exists in both tech and business fields), can you do the analysis(whatever that means in respected field). None of this is replaceable, if anything before ai, if it was 20/80 split between thinking and execution, with the intro to broken tools, it's making the actual roles more demanding if execution is taken off your plate–lets say 80/20 thinking to execution. You can totally see why hiring is so difficult today, because literally moving your fingers or opening your mouth was never the challenging part 20 years ago either–it's being thoughtful, articulate, or having a brain that contributes to organizations across any position. [unfinished]