One of the key points that I think is missed whenever people say do things that don't scale in software is that today's software is not comparable to 10 20 years ago software. The users' standards are much higher now. I wish they talked about this too.
No one in SV is talking about this, which is so odd. The unit economics in regards to building a software company are insane but people are still talking about doing things that don't scale. Since everyone has access to these same cloud tools, the quality of products has drastically increased. No one is deleting their notion to sign up to someones shitty productivity MVP. This also means you have to build something thats 10X better than whats currently out there, which defeats their thesis of doing things that don't scale.
I mean you might be taking it too literally as SOFTWARE that doesn’t scale. I think what these folks were trying to describe is the mentality of not creating problems before you meet them.
I think the idea could be to have some killer feature that Notion doesn't have, that people will put up with a worse overall experience to get that feature. Then if it really is killer maybe notion will copy it, but maybe by the time they copy it you can copy the normal features and get an up to par product.
Thank you for publishing this, I’ve been living your things that don’t scale series. My business is at TheFacebook stage where we’re turning the water on and fixing the pipes as we go along with onboarding customers in a legacy B2B industry. I’ve been learning a lot and using your wisdom to try to improve things as we plug along. Thank you and I hope you keep putting this content out there.
Great story about Facebook servers guys! Never heard those details before. Doing the same, using a DO droplet for each campus. Figure out real scaling later.
this + Ansible, and keeping the configuration inside Gitlab and then issuing the server changes over a gitlab pipeline... you can orchestrate the campuses rather than worry about central scaling. Automating updating a distributed network is just as favorable as a centralized solution especially if each location has its own consistent load. Also, such can be done with an engineer or two.
Most of these really aren't problems any more, especially with pay as you go infrastructure that can be scaled up at any time in minutes via a control panel.
I love you guys. Can you do an episode on non-tech founders and tech founders and the relationship between them? I know Michael Seibel talked about this in some separate videos, but think it could be an interesting topic to make another episode on.
Packed with fantastic insights and great war-trench stories from really smart founders with great instincts. This should be part of every single "Product thinking 101" course. Thanks guys for posting!! 😀🙏👏👍
My only thought is regarding doing stuff that does not scale is I think it matters where you are in the tech timeframe. Back when facebook started in the early 2000's, serverless was not a paradigm or I don't think had existed or sold commercially yet. That makes sense for to build this way because it was based on access patterns per college and was probably faster. However, with the advent of serverless technologies and pay as you go dbs, if you tried to build an app like that today, it would be really insane to manage. Today the easability of building it scalable initially is not that much more difficult that building in an unscalable way. The largest barrier is just finding the hire that knows that already. In other words, if you build stuff today because tech has advanced soo much in the past 10-15 years, you can't apply that style of thinking. However where you can apply this thinking is understanding who your target market is for the product. You don't necessarily have to build it fully out initially but over time.
This advice came at a critical point where my team is building an MVP and I am worried about technical debts when we don't even have a single user yet to break anything. Watching them laugh at FB and Google early days, makes me laugh at myself harder 😂. Great episode! 🤯❤🔥
Yeah totally agree that for specific software there is really little risk to failure and fix on the fly. However, there can be software where life’s can be at risk or the cost of failure is so high that it may terminate the product. So yeah good lessons but important to keep this in mind. MCAS is a good lesson.
These are so good and full of learning! I am already applying my learnings to my startup idea. You guys are creating a lot of aspiring entrepreneurs!! Keep these coming!!
Awesome as always! I would love to see a whole series specifically on stories doing things that don't scale. It would be so cool if you invited founders on here directly to tell their stories, even if it was a simple remote conversations like you've been doing. It would also be cool to hear from Paul directly with more examples of the 90/10 rule. Great stuff!
I have watched countless videos and FINALLY I got real insight about scaling! A thousand thank yous for helping take my products out of launching limbo!🎉
Awesome Series! You both just saved our startup from the perfection Syndrome! We thought we had to not do things that scale but that's exactly what we need to be doing at these beginning stages ! All your resources have being a life saver ! Thanks
The Gmail story is fantastic, but the FB one is a good reminder for me to stop worrying about technical debt and poor architecture when building MVPs. Thanks for yet another great episode, Dalton and Michael!
"A recurring theme is a lot of the best products decisions are the ones made kind of fast and kind of under duress... its like when its 8pm in the office and the sites down, you tend to come up with good decisions on this stuff" - Dalton Caldwell 🔥🔥🔥💯💯💯
Really like your serie guys. There is a real complicity between Dalton and Michael, the mix of you two is just perfect. Pure rationality, pure understanding of people, pure wise advice.
As a founder, I cannot agree more with you. don't make the perfect the enemy of the good. Though it sounds simple it's the one of the hardest things to do as there are a lot of temptation to lose your focus and start over engineering the product or service you are working on. Love you videos and love you guys, keep on sharing such information. Cheers
YC: "How would you feel if Michael and Dalton stopped posting these videos?" Me: "Very very upset" YC: "Ok, next we're doing a talk on how to measure PMF"
If that's what you mean by doing things that don't scale, then it makes sense. What I thought you guys meant is things like Dennis Ritchie offering to personally compile his colleague's C code into assembler without first writing a C compiller, Tim Berners Lee offering to personally format documents for his friends without writing an HTML parser, or Mark Zuckerberg offering to personally courier messages from friend-to-friend at Harvard without building the first version of Facebook.
Problem i have is that I happily do the 90/10 thing on stuff I dont like doing but things I do like doing or am passionate about, I become a perfectionist
The hack Seibel mentions at 20:09 is what Facebook famously did with its own translations as well - helping the platform breakout to 1B+ users (at a time its own forecasts were suggesting the website won't grow beyond 500M-750M users).
I own the most innovate ISP with hardware in Data Centres so I still have to drive to the Data Centre at 2AM when things go wrong :( I remember about 10 months ago having to drive to the Data Centre at 12AM due to one of my Routers locking up. I ripped that Router out of my Data Centre colocation as fast as possible. Internet Exchanges (IXs) are where BGP peering with other networks is done :) I loved hearing your hostage tactic to force the Swedish ISP to peer with you lol.
Hello! Have you done a hardware edition :) i've seen most of the other videos, didn't see anything that's hardware specific (pebble comes to mind) but thats a little bit in the past. Anyone else got a lead/video i can follow?
*☝️☝️I am pretty new to financial planning. I have just started my journey.but the way gave the plan and their approach gave me a lot of confidence and peace🥰of mind.their comprehensive financial planning is very clear and simple:their interaction is very professional and caring😘*
this advice doesn't work as well in the enterprise space imo. there's still tolerance for things not being perfect but unless you already have an entrenched position with the client you have less room for fuckups or janky initial products.