What should you build?
exponentially.substack.com
exponentially.substack.com
Paul Graham: "Almost all founders learn brutal lessons during the first year, but some learn them much more quickly. Obviously those founders are more likely to succeed. So it could be a useful heuristic to ask, say 6 to 12 months in, "Have we learned our brutal lesson yet?"" "The most common lesson is that customers don't want what you're making. The next most common is that it isn't possible to make it, or at least to make it profitably."
Can you make it? Will people want that? Will it make money?
Related PG tweet: "When young founders build something that they don't want themselves but that they believe some group of other people want, 90% of the time they're building something no one wants."
And also related: https://mitchellh.com/writing/building-large-technical-proje...
1. If my company were to fail, why would it have happened? 2. What would have to be true in order for the company to succeed?
(From Rob Fitzpatrick's "The Mom Test")
It might be tempting to think this will always be obvious, which is dangerous because other scenarios can appear to be very similar on the surface. e.g the classic, your customers are asking for "a faster horse".
You don't have to be making something as world changing as a car, but if you are trying to get people to switch from a niche horse to a niche car, it simply takes time for people to adjust to your solutions different perspective, to realise that many of the things they are asking for are not fundamental to the problem but artificial, ancillary parts of the old solution.
You can end up second guessing yourself a lot while waiting for the turning point that validates your idea - speaking from experience. To make things more complicated I think there are also going to be scenarios where you are making a car which is only marginally better than a horse, and the fundamental change to approach is just not worth it to people - differentiating all these subtly different things from customers simply not wanting it is not easy, I think people just need to think very deeply about their products and customers to reach the right answers.
> Make something that people want
I still find this kinda complicated to follow in practice. Make something that you want. What you need, day to day at best. You longed it for long, searched all the possibilities and it just doesn't exist yet. It helps with motivation, and you know the basics to start from. Other people/the market can then help steer it in the right direction.
Also, although “make something people want” is pretty simple, still people forget it all the time (Including me btw). So it’s as much a self reminder as anything!
Or something that significantly reduces people’s living costs?
Or something to increase equality and equity?
There is far too much priority placed on making profits in tech and not nearly enough doing what’s actually good for people.
Years back I took a (big) pay cut to work with a non-profit. We brought affordable internet to low income and refugee status apartment buildings, developed tools to help those that are homeless or at risk of becoming, wrote at-cost software for the non-profit and community health sector. I was there for 7 years as their head of platform engineering and automation. A very rewarding experience.
Others, with the little time they have left, want to work towards not working for someone. That should be what’s encouraged. Because that is more likely to produce jobs, give freedom to those that profit, and has a higher chance of something good happening (versus working for a corporation until retirement - which is perfectly fine, but a path none the less)
I also have a gripe with individual citizens being responsible and taking the burden of climate change. Why should the working class, whose taxes have gone up, wages stagnated, be continually punished with higher costs due to proposed carbon taxes, or initiatives that drive the price of staples up purely due to climate change. That just foments resentment and contributes to the opposite of such activists goals.
Policies are too one sided. We should be pushing for reasonable policies, that acknowledge the enormous progress oil and gas has given the world, and understand changing it rapidly will cause more harm to people than good, and in all likelihood will delay projects. We should look at proven solutions, like nuclear technology. Invest in renewables, in research fusion research and the likes.
But the vast majority of social platforms represents a very very small percentage of people, and those that actually tweet/post/etc represent an even smaller percentage.
The hard truth is when things like gas and food increases, Americans notice that pain far more than what may happen 30 years from now. They care about how their family will do in the next week, month, and few years.
Punishing them to try and understand the problems via monetary taxes, or guilting them to participate in activists activities, or pushing education material that puts shame to what real education should stand for, won’t work. It will backfire.
What's this supposed to mean? Gas has been useful for the world, and gas companies have already gotten tremendous profit as a result.
Who are you expecting to pay for climate change anyways, if it's neither the working class in the form of taxes with rebates, nor the owning class in the form of oil and gas companies. Who's left?
You can build something that is good for the planet, but if nobody wants it, can it make an impact?
But there are pathologies for these abstract ideals you mention too. Namely that they are hard to measure and therefore attract bullshit artists. They make it easy to fool yourself and others.
The best thing is to solve a real problem others have, thereby doing good, and to do so in a self-sustaining way, which is also called profit.
Was thinking about a small 2d scanner/ccd camera in the profile of a SpO2 sensor. Connects with Bluetooth. Would take pictures of garments, fabric, yarn, etc for purposes of color matching in a store.
The app would have color calibration with white LEDs. 2 main functions: "store" and "match". Store would allow storing color samples. Match would show the samples you have and %fit to the stored sample/samples.
Could easily integrate with online stores for matching colors as well.
But yeah, I don't have the capital to do this. Maybe someone else can run with it.
It’ll work similarly to those heart rate measurement apps.
Except it's pretty pricy. I was also thinking a clip-based device, but I can see how an eyeglass would also be good.
(I was thinking of a clip so you could have white balanced LEDs on the same and opposite sides of the ccd chip.)
But yeah, thats definitely the thing. Thank you!
Better spot colorimeter devices will have two different calibrated white LEDs and use both of them in sequence, to get a more accurate measurement of the true color of the object you're sampling.
The big trick with these devices is the color library that they're matching against. Sometimes you really do want to match against the Pantone library, and I don't believe any devices other than the official Pantone CapSure can actually measure against that library. But most paint manufacturers, etc... have their own color library, and there are a lot of other color libraries out there. And many of those libraries are exclusively licensed to only one particular spot colorimeter device.
So, sometimes you need to have multiple spot colorimeters/spectrophotometers in order to be able to use all the libraries you want.
Yes, I have several, including the Pantone CapSure and the ColorMuse. No, none of them are perfect. They're all missing one or more libraries that I might want to use.
It takes time, tears and sweat and a lot of failures to actually build something successful, it won't be a one night sucess from idea taken from someone.