Up until my current job, my entire career involved physical things. Either doing electronic design, or writing firmware that controlled something or both. The current day job's not that interesting, so let's talk about the side gig instead.

My most interesting product converts position signals from old machine tools (e.g., large planers and mills built in the 70's, 80's and before) into signals that modern motor drives can read. It makes it much easier to retrofit old machinery with modern controls.

It was built at the request of a single customer and I did it mainly for fun, but over time it's opened my eyes to the market for tools to retrofit and modernize old equipment. They're still actively buying these from me, but I know there has to be a larger market.

Product #2 is still being developed: a generic "leveling module." Basically it's a motor controller designed to keep the load level in one axis.

For the last 15 years or so I've also been making modules take a pulse input at one end, and send an accumulated count out the other end over an RS232 serial interface. I don't sell a lot of them, but the surprising thing is that demand hasn't dropped off over all this time. Could probably sell more, but it's really not interesting enough to put a lot of marketing effort into. The fun part is all the customizations I'm asked to do.

It's not even just retrofitting and modernization. Even just repair of larger CNC machines is unbelievably expensive and very cool.

Are you removing old motor drives from machines and installing new drives to extend life? There is certainly a larger market. I'm in Norfolk, VA and there are machine shops all over the place. I'd love to hear more from you.

I personally am not; my customer's core business is repairing and retrofitting the machinery. They kept running into the issue doing some retrofits where the new drives weren't compatible with the signals from some of the old machines and it was causing them to have to modify the drives (breaking the warranty) to make it work.

They realized that if they had hardware to read the original position indication, do some DSP and convert to something the new drives could handle, it would simplify everything. That's where I came in.

It's been working out pretty well for everyone concerned so far and I'm now working on version 3 of the product!

I am a SWE at stenon [0], we make a portable nutrient analysis probe for agriculture. It is basically an electronic shovel that you stick in the ground, and you then can view a real time breakdown of bioavailable and mineral nutrient content of that soil on your phone or tablet.

I often say we're actually building three products, ML, hardware and software. It is pretty challenging but a lot of fun, and very rewarding.

[0] - https://stenon.io/en/

Great if you can make this work. My questions would be how long does a probe last before you have to re-jig it? Does it work with sandy soils? Clays? To what extent do readings swing massively based on soil substructure and poor contact/other random chance? Does it internally capture material and create liquid solutions to get around this problem? What range of data can you sample?

Ideas: Don't just target industrial agriculture, consider the residential and commercial garden and surveyor markets. For inspiration or allies, look at the mining industry's established ecosystem of prospecting and sampling related service and software companies. They could also be customers: with open cut mines, they often have to do site remediation, eventually, and it's well funded.

This is really cool.

The job that pays the bills is pushing code. On a personal level, I build airplanes. I've done a couple different techniques end to end - composite and classic cloth and tubing. If things don't totally fall apart, hope to start my first sheet metal construction this summer. Were one to compare what I spent to what I got back... well, some of these dot bombs don't seem so silly. :)

I have a side hustle where I produce niche desk ornament sculptures and busts made from concrete. Etsy link: https://www.etsy.com/shop/jurgenstudio

This brings me around 500 euros per month. Recently the sales of Marcus Aurelius took off nicely, so I'm seriously thinking about expanding this business by making busts of more famous philosophers and leaders. I'm in the process of adding mini concrete busts of Alan Turing, Steve Jobs, Slavoj Žižek, Nikola Tesla, Plato, Aristotle, Jordan Peterson and more.

If anyone is interested in how I these are made (3d printing, molding, concrete pouring), please check here: https://twitter.com/matixmatix/status/1243116129289146368

It's hilarious how there's so many sellers for that same damn Venus bust on Etsy.

How did you do the digital sculpting? Any tutorials?

I make and sell (used to sell) art toys on Instagram, and I make my digital sculptures in Blender since it's Free and Open Source. CGCookie.com just came out with an up-to-date course in sculpting with the newest version of Blender (2.8x). I highly recommend it. It's mostly up to par with zBrush, and you get the benefits of the FLOSS community. https://cgcookie.com/course/fundamentals-of-digital-sculptin...

I outsource that. I find 3D design sculptors on Upwork and CGTrader and tell them what I need. When I'm satisfied with the result, I send it to my 3D print guy who makes a print for me.

Cool to see your process making the busts!

How do you put the two halves together when they come out of the molds?

The two halves of the mold are put back together and the concrete is poured into the bottom. Once it's set, the mold is removed and the result is once piece. Here's a good video from Tested on the process (not with concrete) https://www.youtube.com/watch?v=DEVi0mEaJJQ

Any advice for running a business on Etsy?

I'm also running a website where I ask people exactly that! :) https://myhandmadestory.com/ (success stories from independent makers who sell handmade stuff online)

Bread: https://pearlbakery.com

We are moving the operation out of downtown Portland OR, but this virus will slow down our opening.

There’s a surprising number of ways to apply technology to bread baking. Inventory management, formula calculation, delivery planning, etc. There’s also the problem of trying to optimize the production schedule , i.e. when to mix, how long to rise, when to shape, and when to bake, across your entire menu with orders changing day to day.

> There’s a surprising number of ways to apply technology to bread baking. Inventory management, formula calculation, delivery planning, etc. There’s also the problem of trying to optimize the production schedule

Yup! For example, here's an old press release from a vendor of combinatorial optimisation software (LocalSolver) -- the type of software that you might use to optimise Operations Research type problems: inventory management, supply chain, planning, routing, scheduling -- about a successful integration and deployment of their product to cost-optimise a large Japanese bakery's supply chain decisions:

> [Pasco Shikishima Corporation]'s supply chain involves 15 factories in Japan, each one with several production lines, and more than 100 distribution centers. Pasco's catalog contains more than 1,000 products. 900,000 orders have to be executed each day in Pasco's factories. For each order, Pasco has to decide where and when to produce it. Moreover, Pasco has to decide where to source raw materials and which routes to deliver distribution centers. The goal is to minimize production and distribution costs over several days of horizon, while respecting production and distribution capacities.

> Here is the scale of a LocalSolver instance solved by Pasco to plan the next 3 days: 32,670,717 expressions (that is, intermediate variables) including 8,307,431 binary decisions, 991,251 constraints, 16 lexicographic-ordered objectives. This model was solved in 3 minutes of running time on a modern but standard server, including input and output processing times.

-- https://www.localsolver.com/news.html?id=70

(& apologies for dragging the conversation back towards software, but there's some challenging mathematical & computational problems involved in figuring out how to manufacture & distribute physical products efficiently)

Lighting systems - most usually illuminated props of various varieties for ceremonies of big events; things broadly like the Olympics. It's fun but generally NDA'd to hell, unfortunately.

Vehicle Driving Simulators This consists of three major parts: 1) The game engine that simulates the physics of different driving scenarios and rendering them to the drivers display. 2) A six DOF motion platform that is used to recreate the motion of the vehicle. 3) The vehicle body mounted on-top of the motion platform that has the steering wheel and other controls exactly like a real vehicle.

The driver inputs are read and fed into the game engine which outputs the rendered video to the vehicle LCD's. The output from the game engine is also used to control the 6-DOF motion platform to recreate the physical motion one experiences while driving.

Is there a link to view if you dont mind sharing. This is just purely out of curiosity.

Are you competitive in any non-military markets?

I'm a mechanical engineer working in a university research lab. We're a pretty broad lab with researchers in diverse domains but focused around one idea: innovative design. Making stuff that works.

As a research engineer, I get pulled to work more on "development" projects close to or with industry. Grad students get the more early phase research stuff.

I'm currently working on a contract for a large aerospace company, which involves a new kind of thrust-generating machine. I'm solely responsible for designing and building a prototype and test bench as well as doing the experimental work when that is done.

It's diverse and rewarding work, but it gets a bit lonely sometimes compared to working in teams with 10s of engineers and techs in a large multinational corp. It's still the most enjoyable job I've held so far.

As a small sideline we've developed a new type of house that is CO2 neutral in materials over lifetime, low energy to passivhaus, very quick to build, and healthier to live in.

Designed for Nordic climates but insulates just as well if it's warm outside :)

The initial target was a modern starter house pattern, that could be expanded as the family grew, then split for renting out parts when the owners got old. It had to be cheaper, environmentally friendly, and healthier to live in.

As a bonus it has excellent fire resistance, and is very quiet.

It's fun to work with this kind of problem. Quite far from my normal projects. But holy hell, I didn't realise how truly broken the bureaucracy around construction is in many regions.

We make water meters, radios for water meters, mobile radio receivers that go in utility trucks, fixed radio receivers that go on towers. I test the new and old products, using a lot of Python to automate and speed up testing.

I love building physical products. My first ever physical product was https://www.instagram.com/humblepen/ but after couple years was shut down.

Now, I don't treat these projects as startups, but solely to share the joy of building, learning and teaching.

Recently, I built a project called Wise Charlie which is a compact list of mental models in a deck of cards form. https://www.wisecharlie.com/

When not running my software biz, I make exercise machines which use flywheels for resistance. The difference compared to traditional weights is you get constant resistance at all angles throughout the entire range of movement. Plus it allows many exercises in a compact package, and the resistance can be variable based on how hard you work. It also includes an app + sensor (Electronic hardware is a complete PIA, but we got there in the end) https://exerflysport.com

We're a small team of doctors that redesign everyday household products to be more hygienic.

Our first product (launched in December 2019) is an antibacterial/antifungal 2-sided bath towel designed to separate natural bacteria from exposure to your face.

Several medical research studies show we spread pathogenic bacteria and fecal matter to our face when we reuse a bath towel.

Our first product helps reduce natural pathogenic bacteria exposure: https://halfff.com

Network Engineer / SysAdmin / Helpdesk by day in Singapore.

I’m part of https://www.dwarvin.com that provide lighting solutions to model railways. Its a small family business and I help out with online side of things. Recently was ask to help out with to build a programmable lighting system and am having tons of fun, brings me back to my embedded systems class during my CS degree time.

Small, tailored, temporary demonstration and applications with sensors plus very simple software running locally are becoming quite popular for home-made domotics over here. Many times, it is non-technical people willing to see how this could be going to work in case it becomes mandatory. They lose interest soon, though, smart is not smart if there is no visible gain, the like you can follow with smart meters and bills or with fitbits and health.

We build the computer accessory Luna Display[0] which turns your iPad or extra Mac into a second display - particularly useful for working from home!

We were originally pure software people, so it's been fun and challenging to explore the hardware space. I've found it to be a lot like software development but with much much longer "compile times" aka manufacturing.

[0] - https://lunadisplay.com

Hey, that's pretty neat. I'm on a MacBook Pro that doesn't support Sidecar with our iPad Pro. Going to seriously consider this.

I'm the founder of Huan [0] - we make smart pet tags. We're the only ones making tiny cat-sized ones too (only 17mm in diameter)!

Once I'm done raising this round (great timing, I know) we'll start deploying our mobile sensors that let us create ad-hoc BLE networks that can be deployed dynamically and find missing pets very quickly without requiring a GPS.

[0] - https://gethuan.com/

Those animated buttons are horrible.

That's what I said, but my web designer insisted. Gotta pick your fights!

Systems that automatically prepare, package and retail personalized meals direct from fresh ingredients in a 2m² footprint, and can self-clean.

Actually, a network of those plus: a factory to build them; operations centers to supply them; and a logistics network to resupply, maintain, and repair them.

Now 4 years in, we have a factory, we're currently finalizing for mass production, we own our own production equipment, we are way below budget per unit and we're funded.

We make smarter occupancy sensors for your home that know how many people are in each room while protecting your privacy. You can finally automate your home to instantly react to you with zero commands, like a sci-fi movie. https://www.hiome.com

My background is actually entirely software, so this is my first physical product. It’s been both challenging and a ton of fun!

I built a digital chip tester from scratch, including schematic/layout, firmware and software. It's still a work in progress, but I thought I'd share.

I wrote a blog post about the process here:

https://www.geminicomplex.com/about

My latest company makes custom furniture, architectural millwork and woodwork. www.valkyriewoodwork.com

3d Digital design, CNC manufacturing with laser and router machines, and we still depend on a lot of skilled and talented people on the bench.

So much of my time in software project management has been an asset here.

I make ice cream sandwiches that I sell wholesale as a side business. It's hard right now, good thing I have my full time.

3D printers. Although I just do the software bits.