from Hacker News

After 3 Years, I Failed. Here's All My Startup's Code

by gniting on 12/17/24, 12:38 PM with 268 comments

• by gortok on 12/17/24, 10:54 PM

  > While we gained some traction with some startups, we ultimately weren't able to build a hyper-growth business.

  >"We then decided to pivot into a vertical B2B SaaS AI product because we felt we could use the breakthroughs in Gen AI to solve previously unsolvable problems, but after going through user interviews and the sales cycle for many different ideas, we haven't been able to find enough traction to make us believe that we were on the right track to build a huge business."

  How many wonderful niche products would be around if their owners had tried for a small business instead of a 'huge business' with 'hyper growth'?

• by bwood on 12/18/24, 2:39 AM

  As a customer of Konfig, I am super bummed about this. I think we may even have been their first customer (referenced in the init commit). Dylan and Anh-Tuan solved a number of real problems for us, and they did it spectacularly. They essentially provided a one-stop-shop for generated SDKs and beautiful hosted docs/tutorials that would call our API and immediately allow users to start playing with real data. Prior to Konfig, we were hand-writing a hodgepodge of SDKs to wrap our API and using ReadMe to host the docs (which never quite worked interactively with our API due to request signing that we require on our API).

  We're self-hosting the docs now thanks to Konfig open-sourcing it: https://docs.snaptrade.com/

  Tons of respect for these guys and I wish they'd found a way to make it work.

  Also, for anyone finding themselves in a similar situation as Konfig (a product that (some? most? all?) customers love, but revenue not scaling as needed), please consider charging more. We probably would have paid 2-3x what Konfig was charging us and more as we grow, but they never asked and never built in any usage-based cost scaling (like $X/month/SDK, $Y/month/demo, etc).

• by Hendrikto on 12/18/24, 11:06 AM

  > While we gained some traction with some startups, we ultimately weren't able to build a hyper-growth business. […] We then decided to pivot into a vertical B2B SaaS AI product […, but] we haven't been able to find enough traction to make us believe that we were on the right track to build a huge business.

  American startup culture is so weird. Not everything needs to be a hyper-scaler, not everything needs to be huge. Businesses should focus more on sustainability, and not shut down after 2–3 years after pivoting three times.

• by gkoberger on 12/18/24, 1:44 AM

  Oh, it looks like we were competitors! I'm the founder of ReadMe.com, and we do API docs (as well as SDKs and more).

  I don't really know why, specifically, we ended up being successful. I know I also had a lot of false starts – slightly iterating on the core idea until I hit something people were visibly excited about. I also knew it was a hard market, so I put a lot of my time and effort into non-engineering things. Hacker News was always good to us, especially as I was more vulnerable.

  I'd love to hear more! I'm going to reach out privately :)

  Overall, though, congrats on all the work you put in. It's incredibly hard to start a company and believe in yourself, and you should be proud!

• by asdev on 12/17/24, 9:32 PM

  Thank you for sharing.

  >"We then decided to pivot into a vertical B2B SaaS AI product because we felt we could use the breakthroughs in Gen AI to solve previously unsolvable problems, but after going through user interviews and the sales cycle for many different ideas, we haven't been able to find enough traction to make us believe that we were on the right track to build a huge business."

  I think we are unfortunately going to see this outcome appear a lot in the near future as the AI bubble pops

• by codingdave on 12/17/24, 12:57 PM

  This is not the first story I've heard of a product that isn't quite right, but when deciding where to pivot, fell into AI as the answer and then failed. I've said it before, but it bears repeating - when we figure out what AI is good at and what it is not, we will start treating it as a tool instead of a goal, and our product ideation will stabilize.

• by mcnichol on 12/18/24, 12:51 PM

  I'm not sure how I feel about this. On one hand kudos to them for the self-reflection, facing hard truths, and building a really wide set of tools.

  But when you look under the hood it's tooling that wraps tooling. The API categorization tool arguably hands off a large portion of the heavy lifting to OpenAI.

  "You are a world class categorizer. Fit these APIs into one of these groups."

  The rest of the file is just wiring and a little blurring of the lines of model, view, and controller. I saw some testing and was like, okay this is going to be important if you are wrapping a lot of tooling because "change outside of your control" but then it's just a the default contextLoads() functional test Intellij gives that makes sure dependencies exist and nothing fails at compile.

  I think the vision is there and it is definitely aligned to the Pareto principle but it feels like the idea was tested that markets aren't interested in maintaining their stuff while internally they haven't even addressed maintaining their own stuff.

  Feels like a Catch 22 where if they could address that reason for themselves first then they could probably solve that for other people. But addressing it means having a product that is being used in order to feel the pain and empathize with the end user.

• by ltbarcly3 on 12/17/24, 12:55 PM

  A failed startup is nothing to be embarrassed about, most of us don't have the guts to try it!

  You shouldn't expect much interest in the code, the value of a company's code without the business and people is quite negative when there isn't some technical secret sauce. Even if there was some technical wizardry people would just care about understanding how that one part worked and probably not use any of the code itself.

• by yjftsjthsd-h on 12/17/24, 3:53 PM

  Although the code dump could be interesting, this appears to have no license attached. Making it properly open source would make it a lot more useful.

• by ianpurton on 12/18/24, 9:41 AM

  The developer tools market is REALLY difficult.

  Nearly every architecture decision is "which open source tool can I use to solve [problem]".

  No one wants to pay for anything and that's ok.

• by CodeWriter23 on 12/17/24, 9:13 PM

  Admiring your courage in dumping your repo raw and uncut.

• by jaza on 12/18/24, 5:59 AM

  Good on you for open sourcing the code. Its destiny may yet be as a hidden gem that makes some peoples' lives better. In which case, not what I'd call a failure.

• by williamstein on 12/17/24, 1:48 PM

  > “Our main product was an SDK Generator that could take any OpenAPI specification and generate high-quality client libraries in multiple programming languages.”

  This is something I could really use! Is there a product that does this and doesn’t cost an arm and a leg?

• by tugu77 on 12/18/24, 7:12 AM

  Not every little idea, as useful as it might be, needs to turn into a hyper-growth business. Just make it an open source project on the side. If it gets popular then you get famous and hopefully some big company finds it so useful that they hire you and let you work on it full time. Konfig seems like a perfect example for this.

• by vitro on 12/18/24, 3:54 PM

  Well, with Latudio [1] a language learning app, we're 6 years in since the first line of code and we still haven't thrown in the towel even if we have little traction. But we still do those small little steps and hope they add up so we have a snowball big enough that it starts rolling by itself, perhaps, one day.

  In my mind, I also had a thought that we would rise and be huge, but I let go of that one. Instead, we just put one foot in front of the other and what warms my heart are those stories like when a person in Uganda got so confident thanks to our app that he opened his own language school. For this it makes sense to continue.

  What I want now is just to have a nice work/family balance and the app paying my bills and something extra. That would be a great start.

  [1] https://www.latudio.com

• by deadbabe on 12/18/24, 1:35 AM

  Has there ever been a scenario where someone gave up on a startup, dumped the code, then someone picked it up and turned it into a successful business?

• by andrewstuart on 12/18/24, 5:45 AM

  "I Failed".

  YOU didn't fail. Your business failed. It's an important distinction.

  You are not your business - for your mental health.

• by stunt on 12/18/24, 8:38 AM

  He is young. At that age, I tried multiple times and failed each time. I did manage to sell one and it earned me good money.

  At that age you have plenty of time to pursue many things but you often lack the experience needed to succeed in the B2B market.

  You definitely need support from someone with sufficient B2B experience, or you might have better luck focusing on something in the B2C space.

  My only advice if he really wants to continue pursuing B2B is to find a partner who can sell the product. Selling is a full-time job on its own.

• by popalchemist on 12/17/24, 8:49 PM

  I remember Konfig. Sorry you couldn't find fit, but great work nonetheless.

  About the code - would you be willing to add an MIT License? I don't see the code being used without a permissive license.

• by vkweb on 12/18/24, 3:05 PM

  I would prefer to build a sustainable business and live a calm life. :D

• by block_dagger on 12/18/24, 10:11 AM

  It takes a strong will to admit so publicly that you failed and to release unpolished code. I’m sure many will benefit from your efforts. Bravo.

• by Loxicon on 12/19/24, 5:59 AM

  What is absolutely wild to me is how many developers build "developer focused tools". That is the red'est sea to swim in as a small dev because it's easy to build things you can use every day.

  I have a lot of friends who build tools in book shops, martech, payments and other niches.

  There are so many markets people. Most of them are unknown, and therefore, low comp.

• by thecleaner on 12/18/24, 7:38 AM

  Remarkable show of persistence. Honestly, I really like that the Repo is as-is. Congratulations on the journey, hope you learnt a ton. PS - you built some of your static content on Docusaurus, a Meta open source project. They're hiring like crazy. Maybe you get to work on that one.

• by newusertoday on 12/18/24, 6:44 AM

  from the bottom of my heart thank you @dylanhuang . Releasing it without modification gives us rare opportunity to inspect the codebase of actual startup that had paid customers. Its more valuable than polished/redacted versions.

• by simmo9000 on 12/18/24, 7:39 AM

  Well done having a go. Most don't.

  Publishing your results makes you part of history. Respect.

• by farceSpherule on 12/18/24, 1:07 AM

  A product in search of a problem…

• by sbochins on 12/18/24, 4:45 PM

  “It’s refreshing to see a founder’s honest postmortem that doesn’t just blame ‘bad timing’ or ‘market conditions.’ Instead, he’s ripping off the Band-Aid and laying bare the pain points—no hype, no quiet pivot into mediocrity. It’s the kind of retrospective that should be required reading before your next ‘growth hack’ session: get the fundamentals right or prepare to pen your own ‘closing up shop’ blog post.”

• by mettamage on 12/18/24, 11:41 AM

  Awesome, will be learning a lot from this. I feel this is much of an operational playbook as one could get.

• by nikolayasdf123 on 12/18/24, 1:24 PM

  yeah, "software developers" is notoriously one of the toughest markets