It is way easier to fix a specific problem than to find one and sell solution for it.
For example, Hi I’m John and my accounting software can increase your profits by 10% by reducing time spent doing billing.
Vs
Hi I’m John, we do custom software and we can do an app for you in accounting for example, but also a crypto wallet or a booking website.
If you are in the second route, there is no mass outreach strategy because you are not offering anything specific your customer is anyone and your solution is anything. The outreach works for the first category, because spam or not, if someone is offering to fix a problem I have I’m ready to listen. Hence the relationship advice, if you provide services, people around you need to know that you “do apps and stuff”, and if you do products you can throw 1,000 emails fixing one thing with some level of certainty that someone within the ICP will give you a chance.
Just to share a counterpoint: I think it’s impossible to get a product market fit without active outbound verification. If it’s a brand new project, balancing development commitment and GTM hours is crucial. If I understand your comment correctly, the brand marketing approach you are talking about works for products that verified their need, but if this is a brand new idea/demos committing to GTM instead of active outbound/ emails/ calls/ in person sales may be too costly.
Did you have a chance working on small / startup projects; how you balance development commitment?
I'd seperate PMF hunting and go to market just in case that's not what you're doing. I've done tons of small/startup GTM projects, most notabily I was on the deviantart zero to 1 and the digitalocean zero to 1. I wasn't clear enough. I'm not advocating that you should skip validation or assume a PMF, rather, you shouldn't even be thinking about a full, formal GTM until you've proven repeatable product market fit.
You're right that active, direct customer outreach in early stages, talking to real people at events, user interviews, forcing them to talk to you etc is the only way to validate and refine PMF. But to me, this isn't really "outbound" in the spammy, growth hacky sense that I critiqued above. It's thoughtful customer development: qualitative, deliberate, personal, strategic, and hopefully founder lead.
The real outbound I'm against is that lazy "spray-and-pray" approach, or the obsession with hyper personalized yet fundamentally shallow cold outreach that's mistaken for scalable growth. Real GTM, once you truly have PMF, is actually pretty mechanical and clinical. That said, personally, I've always treated early GTM as inseparable from product refinement. Every interaction early on is about learning and adjusting, (sorry Arch, you just had to go!) not aggressively scaling. "Rapid growth" is alluring, sure, but if it isn't anchored in real market validation, as you're pointing to: it ends up as expensive noise.
Am I correct in that much of the 0 to 1 is just like, getting some interactible artifact in front of potential customers, getting them to kick the shit out of it, identifying which of those people liked the artifact best, refining the artifact for THOSE types of people, then repeating the cycle? Like, there's no "market" yet, only your buddy Joe and some guys he knows.
Not really, well kinda, although lots of people skip the first step (well, most people try to skip a lot of steps sadly). You're describing something closer to early stage customer validation (ux/ui or otherwise) imo, (although experienced folks will tell you the playbooks and motions matter a lot, I agree, but that would be too much to get into), and while that iterative process is crucial, there's a key difference: real zero to 1 requires that you've also done a broader market analysis to understand where you're headed in a strategic sense, your "vision". That is not just about refining an artifact based on Joe and a few friends, it’s about deeply understanding the context your artifact fits into, including the broader market dynamics, existing customer pain points, competitive landscape, and whether this thing has legs to actually scale beyond those first friendly testers. (In business the answer is the answer, not what you want the answer to be: 5 customers, 500 customers, 5 million customers, there is an answer.)
Early feedback loops from Joe & friends can be misleading without also stepping back to ask strategic questions like: “How big is this pain?”, “Is this pain widespread or niche?”, and “What are the economic dynamics around solving it?” You’re not truly going from zero to 1 until you’ve mapped those answers and built a clear hypothesis around repeatable product market fit, which involves deliberate strategic judgment calls beyond just refining based on individual user reactions. When we launched DigitalOcean, several major tailwinds aligned perfectly, fueling rapid adoption. Chief among these was the dramatic, yet largely unnoticed, decline in SSD pricing, which Ben and Moisey Uretsky recognized early. Once they explain that change was happening in the world, I was able to understand and explain the whole business: SSD prices dropping allowing us to offer faster infrastructure at price parity (and eventually even lower than traditional providers). The Comp Sci degree was outpacing the MBA handily as an area of study, coupled with the explosive growth in developer focused startups (and startups generally), dissatisfaction with AWS complexity, and a rising wave of DevOps and containerization practices, we positioned ourselves perfectly for the incoming generation of developers. We came to a market with a genuine commitment to simplicity, love, community driven marketing, and API-first infrastructure resonated deeply, creating genuine trust and loyalty that became the foundation for rapid, sustainable growth. I was literally online 24/7365 making sure everyone was happy with everything (except that one time: sorry about that!!!!), just ask anyone here, I was on it.
tl;dr yes, iterate closely with early users, but pair that tight iteration with rigorous market understanding and strategy, if you're unaware of the broder market dynamics, you're at risk of solving only for a handful of early adopters and missing the real market entirely. (I don't like to mention my blog on HN, but I have a most you might enjoy: https://b.h4x.zip/founders/)
By definition, the Digital Ocean Zero to 1 can't have been a thing.
Providing services like virtual machines, managed databases, and Kubernetes clusters wasn't a groundbreaking innovation by the time DO came out in 2011. Even Heroku had been around for four years by then.
I was tired of coming home after networking events and shift through pile of business cards, so I made an app to just scan cards and export them to csv. Pretty much just for fun app for myself, friends, and friends of friends, but other people started using it too.
Here’s a challenge: Get that thing to record license plates at night.
All the closed source ones I’ve seen set the aperture/exposure based on the average exposure of the frame, so the plate comes out as a pure white rectangle.
Have it sweep through bright and dark when recording at night.
About a month ago, I bought a toaster for Christmas, and, while walking in the alley and checking prices offline and online, I was mesmerized by the price strategies. With quantities that toasters operate in, their production cost is around a few dollars, and I bought one for $50 without any digital features.
If you are planning to build one in the USA, you will probably need to tell yourself what "build in the USA" means. Is it PCB printing, PCB assembly, plastic production, mold production, cool design features, coil design, final assembly, or some combination of the above? You will be surprised to find some of the requirements to put "made in the X country."
I do product development professionally, so take the comment as my bias, but it is a light project to start with as there are minimum components, you can build the entire thing in the garage, and it is something that everyone around you can give feedback on. Almost like an art project.
Anyway, I would love to share any info you may find useful to start. Email in the profile, if you are interested.
You are right, it is faster,cheaper and absolutely possible to take the existing enclosure and modify it. OEM enclosure suppliers will gladly add or remove a few openings on their design for a few thousand. They will modify or build a new mould just for your new product, and it works just fine as long as PCB can be designed around their enclosure.
Not the topic poster, but speaking from the similar experience, the cost can be as low as $4,000 in China for the size of these parts. In USA/Canada it would start with a $60k minimum, and with a 6 month timeline instead of 1.
And 9 times out of ten, the local fab is making their moulds in China anyway.
This is similar to my experiences, the really good ones still in the US have heavily specialized on high-end boutique orders. They can be very competent but also very non competitive for any normal work.
Same with with US based PCB shops. They all want to do high dollar ITAR restricted boards for govt contractors.
There is a middle ground. US sub shops that oversee captive Chinese operations with their own crew and keep you in the loop. This worked well for my personal projects and wasn't too horrendous. About 16k for a simple design that I could've had done with our work team for about 8k. But I never had to fly over, play WeChat tag in Mandarin, and I got PowerPoint DFM updates regularly.
We make edge AI cameras. After focusing on services for a while, it has been a different type of journey to switch to the product. We did an open source pilot product on esp32, and got surprisingly more interest than we thought, so now we are working on a high performance (4k, 60fps, AI chip) device.
Luxonis is way more powerful and has an Intel VPU (Movidius). It's not really meant to be a standalone platform, so it needs a host board (Linux SBC like the Pi). It takes a lot more power, but it can do 60fps on small Yolo models. Its resolution is a lot higher as well.
ESP32-S3 has a pretty small memory capacity, and doesn't have H264-H265 hardware encoding, so you'll be on low-res, low-fps. It only does MJPEG (AFAIK) streaming, so you'll also have to deal with high latency if you want to send that data somewhere. The big bonus is that it's low-cost and low-power, and you're running it directly on the core without an OS.
This means you can do stuff like sleep the cores until something wakes it up (like a PIR sensor that detects people), and it will start streaming in a second or two.
TL;DR: Luxonis stronk, but needs big batteries or plugged in. ESP32-S3 can run on small batteries or solar.
By "AI" here I mean cameras that support running machine vision models directly on the device. We provide the device and firmware, and then on the application and model level it is up to the user.
Hi everyone, we made an open source camera for developers. There are some models we already tested (face, parking spot recognition, animal detection,) but this is mostly a hardware project, so there are no custom AI models. Here is a GitHub links for the designs:
For example, Hi I’m John and my accounting software can increase your profits by 10% by reducing time spent doing billing.
Vs
Hi I’m John, we do custom software and we can do an app for you in accounting for example, but also a crypto wallet or a booking website.
If you are in the second route, there is no mass outreach strategy because you are not offering anything specific your customer is anyone and your solution is anything. The outreach works for the first category, because spam or not, if someone is offering to fix a problem I have I’m ready to listen. Hence the relationship advice, if you provide services, people around you need to know that you “do apps and stuff”, and if you do products you can throw 1,000 emails fixing one thing with some level of certainty that someone within the ICP will give you a chance.