HN user
self-organizing, self-healing factories, anywhere on Earth and beyond.
This is absolutely not my area of expertise, and I can't fully vouch for how well it works, but I was looking for a similar solution for consistency across website graphics and came across this walkthrough for consistent game assets:
https://runware.ai/blog/creating-consistent-gaming-assets-wi...
Maybe helpful?
This might be a sub-strategy of the broader manufacturing technique called “binning” in which products of varying quality, size/weight, performance, or some other variable parameter are sorted for later re-combining into similar groups (often to result in different product “grades”), or dissimilar groups to achieve a more uniform total, by average.
Reminds me of the "Exact Instructions Challenge" videos where kids write instructions for some everyday task - like making a PB&J sandwich - and parents, hilariously, follow the instructions verbatim.
https://www.youtube.com/watch?v=cDA3_5982h8&list=PL2tgThFV6O...
There are quite a few specific examples, all on a spectrum of technical and economic feasibility, of course.
Here’s a great place to start: https://www.factoriesinspace.com/
As a mechanical engineer absolutely obsessed with Lego at a young age, I’ve recently (ie, when I could afford to start buying Lego again!) had to reconcile what I previously enjoyed about building with what I enjoy now.
As a child, the joy of building was in part the satisfaction of following granular directions to see something awesome emerging step by step, and in part the joy of playing with the sets. I was immersed!
Now, I think of them in part as essentially 3D puzzles and, in part, just curios… I like looking at them. Perhaps I’m subconsciously craving that unbound creativity I had as a child?
Thing is, in my work, I get to materialize physical, mechanical things with machining and 3D printing and welding etc. I thought I would, but haven’t found myself designing and building new creations with Lego. Perfectly content following the instructions and putting it up on the shelf to admire!
Seems like an open-source/DIY or kit (requiring some assembly) for educational applications is great in the context of a class or project where building/fixing/upgrading the robot is the point, and where 3D printing is readily available.
If the main purpose is for the robot to do something - the DIY approach is more likely to suffer issues that, without support, may discourage students because they're fussing with/fixing the bot instead of doing the thing.
Out of curiosity, what's the $40k kit you mentioned?
Educational packages are all over the place (some seem to be price-gouging, frankly) but the low-end (in terms of payload + reach) of commercial/industrial cobots is getting pretty affordable.
My company designed and built the first few prototypes for a startup, Throne Labs, who deploy prefab, off-grid, (semi-)publicly-accessible (you have to have their app) bathrooms.
Seems like they solved some core problems: finding suitable sites and partners, using semi-permanent modules that don't require [quite as] complicated/slow/expensive permitting and utility tie-ins, and some clever ways to get them re-stocked and cleaned regularly using gig labor (I always understood this to be the biggest challenge for offering public bathroom - labor cost of cleaning and resupply.)
They moved on to a higher-volume producer so I've been out of touch on their progress, but I thought it was a interesting solution in the face of what I learned to be outrageous expenditure - $300k, $500k(!) - on municipal bathrooms that quickly go defunct or get destroyed, and a lack of public restroom access during the pandemic.
During the project I learned a lot about how rideshare and package delivery companies just assume stops with restrooms are available on the routes (I'd love to hear what the actual company training/guidance says! It's easy to find stories about Amazon drivers peeing in bottles etc...)
I also gained a lot of empathy for people who have more frequent and urgent bathroom needs than my own - having trust in availability of facilities can make a world of difference!
We turn shipping containers into microfactories that can be deployed anywhere on Earth. We also retrofit containers for all kinds of fun modular commercial, industrial, and even residential purposes.
I have no affiliation and no idea what the current status of their effort is, or if they are working with orcas specifically, but this group has crossed my radar a few times:
It may seem cheesy and somewhat simplistic at first, but after you get over the cutesy anecdotes and jokes, the book/method "Slicing Pie" by Mike Moyer has been a very helpful resource for me during several very early-stage bootstrappings.
It really helped contextualize the different types of contributions co-founders and early employees will bring to the table, and is a helpful transition into the world of stocks and equity for those who are less familiar.
I bought copies for all involved and we worked through the calculations together, resulting in equity splits that have so far been low-drama.
MODICA (Techstars '21 with Stanley Black & Decker) | Software Architect (Platform Development) + more roles | Seattle (preferred), Remote (US, intermittent) | Full-time | https://www.madebymodica.com/
MODICA turns shipping containers into robotic microfactories using our next-generation industrial automation platform.
We're the world's first Factory Factory, and we're hiring a Platform Architect!
We are looking for an experienced full-stack engineer to design, prototype and ultimately lead the development of a software platform that addresses the needs and opportunities found in Contract Manufacturing and industrial systems integration. The platform spans the internet (cloud-hosted environments and Web Portals) down to client systems (apps and applications) and further down into robots, CNC machines, and embedded controllers and sensors. This is a foundational role that sets the stage for our entire software team.
We're looking for someone with sufficient knowledge of hardware (specifically: industrial automation and robotics) to comfortably work with microcontrollers and PLCs, debug industrial comms (like EtherCAT, CANBUS, MQTT), and quickly build working prototypes in software to link it all together. Ability to build is more important than formal education. We don't have a strong opinion about what frameworks you want to use, languages you prefer, or databases you admire - the slate is blank!
Reach out directly to Founder and CEO Will Gibbs to talk more: will (at) Mmicroindustries.com
Other roles we're actively hiring for (Onsite in Seattle):
- CNC Waterjet Operator (experienced - no entry-level, please) - Program Manager - Modular/Prefabricated Structures - Certified Welders (several roles) - Mechanical Engineers (several roles)
Granted, no, I don't want my $2,000 exercise entertainment center to show up visibly rusting to pieces, but as a mechanical engineer who works a lot with weathering steels (like Corten), and steels of all kinds that rust more or less, all the scare quotes in this article around "rust converter" and "cosmetic oxidation" are so over the top it's comical!
Rust conversion/"reformation" is a common chemical process that does exactly what it says - it turns iron oxide into a more stable composition to protect the underlying steel. Cosmetic oxidation is also a known and normal engineering thing - it's EVERYWHERE in architectural cladding, planter boxes, etc.
I understand Peloton's being pummeled in the press recently for the old (and new) CEO, and this is a continuing hit piece, but I can only imagine the furor if Peloton had decided to scrap/trash all those frames instead of applying a completely normal and minor metal finishing touchup.
Tip for anyone who's disturbed by rust: don't look too closely anywhere under your car, or within your appliances. You might witness some cosmetic oxidation! The horror!
Disclaimer: my company is a Vention "Certified System Integrator" in the Pacific Northwest USA.
I can say that Vention, along with their comparables in the modular framing/light automation world (see: Bosch Rexroth, Item North America) will put some great tools in front of you to piece together solutions within their ecosystem, and you'll be able to check your requirements against their various specifications with regard to maximum loading, linear axis speeds etc, but they will of course not inherently perform a full engineering analysis of your application, or really make any guarantees on your design's suitability for any particular purpose.
For that level of services, they will loop in an integrator, who will work within the hardware ecosystem, but add value in validation, assembly, service/support and non-ecosystem component integration.
Not entirely accurate. Certainly for metals, stone, and other hard materials the entrained abrasive is doing the cutting, but water-only waterjet cutting is used for a lot of things: foams, rubbers, plastics, fibrous materials, food, sensitive items like diapers and medical materials (because the pressurized water is inherently sterilized, and there's no blade to gum up and hold bacteria).
And what's really going to bake your noodle is there is an additive called...wait for it...SUPER-WATER that helps maintain stream coherence out of a high-pressure waterjet nozzle, allowing even more effective water-only cutting.
Sure thing! Looking forward to hearing from you.
Sure - we're a startup (MODICA Microindustries) working on a related problem (modular manufacturing equipment), and we spun up a subsidiary (Harbor Island Waterjet) offering waterjet cutting services as a job shop since we had excess capacity on our three machines.
There are pros and cons to waterjet cutting but for what it's worth, it can cut virtually anything. There are a lot of techniques for reducing splashback and other wetting/soiling from the tank. We cut wood fairly often. Some people like how laser singes the edges better, but to each their own!
For floppy materials we use a few approaches: first, instead of putting the material on the standard "metal slat" bed, we use a product called Rhino Board, which is essentially a whole bunch of plastic straws glued together into a rigid mat. It provides more granular support. For materials still needing more support, we put a sacrificial board (either OSB or plastic sheeting) underneath. And for materials that we can't glue down, or are even more problematic, we also put a sacrificial sheet on top.
Site is down, so I'm winging this comment until I can check again later or find an archive, but I wanted to note that my company has cut all kinds of fabric (and also food!) using a "water-only" (ie, no entrained abrasive) method with an ultra-high-pressure CNC waterjet.
Depending on the material and tolerances, you can cut through significant stacks simultaneously to save time, and with the correct settings and setup you get very little splashback or other wetting.
Since the 60,000+ PSI stream is coming out of the nozzle at something like mach 2, the water is travelling too fast to make the edge wet!
Within our lifetimes (say, the next 40-60 years), no, personally I don't think we'll see completely autonomous end-to-end manufacturing widely implemented (as much as I'd like to, considering it's a problem space I focus on!).
Some pockets of industry are much further ahead than others, but it will take A LOT of work to reach parity across the board. If not for technical reasons (which I'm more optimistic about), then for political and social reasons, as these systems and understandings adapt. That's a whole 'nother discussion...
AI/ML will play a huge role. Not only in machine resilience once commissioned and operating, but upstream and downstream as well. Better (AI/ML-assisted) tools for designing products and the factories/equipment that make them will preempt some of the challenges caused by the currently disjointed process.
I disagree with the comment that AI/ML techniques are only useful once you've physically built a plant - there are of course emergent behaviors that only crop up when dynamics of the whole unique factory are at play, but any given problem that arises is almost always traceable to one or a small number of subcomponent failures, for which better, more granular datasets are becoming available to train AI upon.
And, as I mentioned in my comment about throwing virtual wrenches in virtual works - simulations can begin to generate training data sets as well!
Speaking as someone who has been responsible for "turning the lights back on" to fix problems with "fully-automated", "lights-out" factory lines, much of this paper still rings true forty years on - if nothing else as a check against our engineering hubris. It remains tremendously difficult to quash entirely the long tail of things that can go wrong in a factory.
That said, many contentions raised here really have been resolved substantially with increased computing efficiency and ubiquitous connectivity. The touted expert human operator's ability to see and understand processes from a high-level, informed by years of observing (and hearing, and "feeling") machine behavior has truly been eclipsed by an advanced machine's capacity to collect increasingly granular snapshots of its complete operating state - the temperatures, vibrations, positions, and other sensations of its various organs and elements - every few milliseconds, hold on to that data indefinitely, and correlate and interpret that data in ever-expanding radii of causation.
The best human operators (of any technology) not only respond to problems, they anticipate and prevent or plan around them. Massive data, advanced physics-based simulations, and "digital twinning" capabilities of manufacturing equipment afford pre-emptive testing of virtually infinite scenarios.
Not only can you simulate throwing a wrench in the works - you can simulate the effect of the wrench entering the works at every possible angle!
It's not infallible, and will for a long time still require a human-in-the-loop at some level, but as the author rightly put it themselves near the end of the paper:
"It would be rash to claim it as an irony that the aim of aiding human limited capacity has pushed computing to the limit of its capacity, as technology has a way of catching up with such remarks."
Can't speak to living spaces since we don't do that, but you have the basic process in mind. You might rip the floor planks/panels out entirely, eliminating the smelly, pesticide- laden lumber and giving access to more metal surfaces during blasting. Or, much less ideal but possible: you can seal over the nasty flooring with epoxy then leave it or put another flooring layer on top.
Haven't heard of powder coating interiors - that would be a lot of wattage! - but maybe doable with radiant methods and insulation blankets, or putting the whole thing in a giant curing oven? Most often it's just spray-on epoxy over rust-converting primer. There are low(er)-toxicity formulations out there. For undemanding applications you can paint with household enamel over primer - wouldn't hold up on the exterior though.
Interestingly, the Cor-Ten steel many (most?) containers are made of is a weathering steel that forms a protective rust layer when left to its own devices. So you could conceivably just leave it there to develop a fine rusty patina!
All told, there isn't a single "right way," to clean up a container, but there are some best practices and trends. YouTube seems to have a wealth of folks documenting their container home builds.
Structural modular with offsite prefab, for sure, but I don't know any specifics on the Wuhan hospital builds. From some other coverage, the units appear to have ISO corner castings, but the proportions don't look like 'normal' containers. There are other proprietary connectors for volumetric prefab - like Z Modular's "VectorBloc" - but these don't seem to be using anything fancy.
What stands out to me, watching the time-lapses: SO. MANY. WORKERS! They are everywhere on site, moving constantly, and the equipment never sits still either. That workforce, along with the certainty that zoning and environmental impacts weren't even a question mark show what the "command" side of a hybrid economy can pull off.
Certainly - the companies (mostly in China) that manufacture containers will do one-offs. They can be specified with less toxic epoxies and composite/alternative flooring, and any cutouts or other features you want to end up with. This avenue is expensive though. You don't have the economies of scale in comparison to "standard models" pumped out by the tens of thousands, and you don't have the advantage of shipping the container over full of goods instead of empty. Caveats abound, of course.
If one wants to avoid stripping paint and floors out of secondary market boxes, it's also possible to buy the constituent pieces of a shipping container individually - the corrugated sidewalls, doors + fittings, floor channels, corner castings etc - unfinished or with primer only. In this way, you can ship one or more deconstructed shipping containers INSIDE a shipping container! But then you have to fixture it and weld it etc, so you're back to losing economy of scale...
For our purposes, we usually use 1-trip/low-trip for client projects avoiding the nasty ones by careful selection, and we have a custom machine that media blasts the interior and exterior down to bare metal when we work with older WWT and partial/damaged containers of unknown heritage.
My company works with shipping containers daily - we modify them for industrial and manufacturing purposes. We reached out to the CURA team several weeks ago, offering our capacity if they scale/deploy their concept. They are currently building their very first pilot unit(s), so we'll see!
As a basic structural shell, shipping containers are great - plentiful, compatible with global logistics, cheap, dense/stackable. But, to bring them up to habitable standards, let alone medical standards, takes A LOT of work. Hard manual labor and also precision assembly work. We have some custom equipment to help speed that retrofit process for our own purposes, but most container mod providers don't - so while yes, you can deploy finished units anywhere, stack them densely and connect them quickly, there isn't a reserve of these units standing by, and you have no scaling advantage up-front in manufacturing them right now.
I'll add that by the time you retrofit insulation, ventilation, utilities and paneling inside (and you have to put this inside if you want to maintain side-by-side stackability and weather impermeability) what was a "decent" small room size becomes a bit claustrophobic.
For modular, dense, deployable, and durable emergency hospital facilities, it's best to look at one of the many architectural prefab approaches - volumetric, panelized, or otherwise. See: BLOX, Blokable, FullStack Modular, Katerra, etc. This field is growing rapidly, and many of these companies are already tooled to produce room and structure modules very efficiently.
Some pose the question of re-tasking hotels as temporary alternative medical spaces. I could see it for housing medical staff in a more dedicated and perhaps centralized manner, if nearby a hospital. There are several military slide decks circulating around that describe exactly what's necessary to create a field hospital out of a hotel/office - including ripping up the carpet in the entire facility, heavily modifying HVAC (central or standalone units) etc - and that's all doable, but I've wondered about the implications afterwards. You'd have to basically rebuild the interiors entirely, battle future customer perception ("Oh, the hotel that 100 coronavirus patients died in?") and I'm certain the insurance situation will not be straightforward...
Does anyone here have any personal, specific, actionable knowledge of the melt-blown plastics process? Are there alternatives for achieving the N95 (or sufficient) filtration levels in a different sheet material?
Automating many steps of assembly are [more] straightforward, but this precursor material supply really is a problematic bottleneck.
I balked at first glance because "free energy," but here we have a unique method for turning thermal into kinetic energy. Slightly outside the bounding box of this energy system, of course, lies the expending of more energy than what's produced to keep the hot water hot and the cold water cold.
But using natural low-grade temperature differentials as the author suggests, what might this arrangement achieve that a Stirling engine doesn't? More torque? Advantage of fewer moving parts? Novelty only?
In 2016 Seattle hosted a recreation (or maybe it was a memorial?) of 9 Evenings called 9e2. Organized by John Boylan, whose "Conversations" series is worth checking out, 9e2 was fun and intermittently thought-provoking but from what I understand not much more than an echo of the original NYC event which, as outlined in the McCray piece here, was itself an experimental shout into the dark that received many negative critical reverberations, but also some resonance that lasted.
There's a fine line where technical art - "new media" and beyond - and artworks made purposefully to reflect upon or expose specific topics, issues, or tools in science and engineering stop holding their own artistically and becomes either "gadget show-and-tell" (symptom: when it takes more time to put on the VR headset and wait for the creator to reboot the computer and scene than to view and interpret the 5-minutes-of-Unity-tutorials art within), the kitschy sculptural, musical, or whatever media equivalent of Popular Mechanics cover art, or interactive science museum displays.
That line is fine, and I wonder where each of the works and performances in the 9 Evenings program sat on the spectrum, but there are many artists currently and in the past who have adapted or interpreted science and engineering material successfully and to great effect (Refik Anadol comes to mind at the moment).
It's nice to see more options becoming available in this space. Industrial PCs (and, in the same vein, PLCs) from the old-guard industrial automation providers are generally lackluster for the cost, in my experience. So many acts of industrial automation and robotic system integration can be achieved with a Raspberry Pi and/or an Arduino, with a little hardening.