Yeah, I was referring to spicy autocorrect in the more general sense of something that uses a faster statistical model to replace a slower theoretically derived exhaustive calculation.

The regular old FEM based models can be quite misleading and when I had the chance to dig into them some years ago, it made me vaguely anxious. Except that nobody trusts the existing CAE solvers, there’s always a process to verify that actually the structure does what you think it does.

Aerodynamics, at least the coefficient of drag, is actually really good for this because you can’t cheat the air and it’s mostly obvious when you screw it up. Which isn’t true for flutter or the more structural details.

So, yeah, there is that risk, that they’ll get high on their own supply. But thankfully the management already thinks that the current crop of CAE solvers are magical and so the credentialed professional engineers already know how to fight that battle for a lot of the structural details. (The long-suffering assembly line folk who are trying to assemble the airplane properly are, of course, a different matter and have had a lot less leverage)

Although, I’d also propose that there’s a second risk, which is that the current validation process is oriented towards the ways with which the existing FEM models screw you up and it’s likely that when the large physics model screws you up, it won’t be the way FEM models do.

Yah, I have some vague experience in the space and, without getting into things covered by NDAs, I guess I can say…

First, The popular media talks about the classic style of physics solvers as these magical black boxes but my experience is that they are sufficiently unreliable that I would never trust my life solely to the answers of a solver. They do provide very valuable feedback for refining a design without an endless hardware-rich cycle of destructive testing. Thus, I think that a large physics model is probably going to be the same sort of useful tool.

Second, while the CAE engineers can be very very protective over the time they spend on the two week cycle the article talks about, it’s fucking drudge work and a waste of a good mind. At the same time, the article does not really talk about some of the nitty gritty details. Aerodynamics is a great place to start because there’s less setup but the coefficient of drag is only one problem that needs to be considered.

Third, the good engineers can “see” things intuitively because things do operate with a pattern. Vorticies from protruding features… stress fractures from square holes in a beam… etc. This does feel like an area where spicy autocorrect can spicy autocorrect you to a useful answer.

Finally, cycle time for real world engineers is just like the cycle time for software engineers. Nobody wants to go back to the world where programmers submitted a deck of cards and got the printout back a week later.

The only real risk here is that somebody gets high on their own supply and decides that a large physics model is actually predictive and we don’t need the same set of actual physical tests that validate the models.

So the conundrum with the Voron world is that the Trident and 2.4 are basically as good as you are going to get for the constraints. You can add the Monolith gantry but that’s an involved mod. You can add a toolchanger or a filament changer, that’s another involved mod. And there’s a bunch of really great mods but each of them adds complexity to the build, makes the BOM larger, etc.

The one thing with the Voron is that if you want the highest accels, you probably need to ditch the extrusions. But then you can’t make it as an open source printer because you’d need to do a lot to get a rigid metal frame and there would be minimum orders, etc.

And, overall, if you look at Qidi, they’ve been making Klipper-based cube printers with active chamber heaters for quite a few years now, so the more recent Qidis are really just mods atop the ur-Qidi, kinda. So a lot of the new hotness, outside of a few Bambu things, exists as mods for the Voron. We’ll ignore that Prusa had problems delivering new printer designs for a while.

Allegedly the INDX that looks actually pretty neat that’s going to be on the Prusa is also going to be available as a kit for the Voron.

Neither of my Vorons are stock. The Trident came from Formbot so it already was a CAN-bus design with some Formbot tweaks. You definitely want a filament motion sensor, there’s a bunch of options there. I swapped to the DragonBurner toolhead, I’d probably try the A4T instead if both of them were full-sized printers. My Trident has the inverted electronics mod, that felt pretty handy. My 0.2 has the electronics compartment rearranged.

Well, think about it this way…

You could hit AGI by fastidiously simulating the biological wetware.

Except that each atom in the wetware is going to require n atoms worth of silicon to simulate. Simulating 10^26 atoms or so seems like a very very large computer, maybe planet-sized? It’s beyond the amount of memory you can address with 64 bit pointers.

General computer research (e.g. smaller feature size) reduces n, but eventually we reach the physical limits of computing. We might be getting uncomfortably close right now, barring fundamental developments in physics or electronics.

The goal if AGI research is to give you a better improvement of n than mere hardware improvements. My personal concern is that that LLM’s are actually getting us much of an improvement on the AGI value of n. Likewise, LLM’s are still many order of magnitude less parameters than the human brain simulation so many of the advantages that let us train a singular LLM model might not hold for an AGI model.

Coming up with an AGI system that uses most of the energy and data center space of a continent that manages to be about as smart as a very dumb human or maybe even just a smart monkey is an achievement in AGI but doesn’t really get you anywhere compared to the competition that is accidentally making another human amidst a drunken one-night stand and feeding them an infinitesimal equivalent to the energy and data center space of a continent.

So, one thing I did note is that a lot of the PETG vendors these days have been ratcheting up the speed that their normal non-rapid PETG filaments are printable at as well.

For example, Prusament PETG is good up to 200 mm/s and Elegoo’s Pro PETG is good up to 270 mm/s.

You can always mock some stuff up and try it out in PrusaSlicer to see how long it thinks it’ll take?

Wall thickness potentially depends on the size of the object? I guess 2mm would the the starting point, fill one with soil, see how sturdy it feels. Complexity for 3D printing is “free”, kinda. A lot of the best container designs incorporate ribs to strengthen them without using up too much material. Given that the joins are the weak part, you’d potentially want that a lot thicker.

You also want to look at “vase mode”. Some of the fastest printing objects you can get on a 3D printer are where you design around the constraints of vase mode and then you can use a fat nozzle with thick layers to print really fast.

You can always print plumbing instead of using PVC pipes? I’ve definitely seen self-watering pots such that they just have a pipe incorporated into the design such that it just sticks up along the corner. So, worse case, each module has a watering port. If you want to get fancy, you could make a manifold such that a single pipe sticks up in the middle and fills 4 reservoirs, although the fancier the plumbing the more likely you are to have one of them get dried out faster unless your filling routine tops them off.

Yeah, see if you know Python, then OpenSCAD is not a hard jump? One of the reasons why I really like OpenSCAD is that libraries like BOSL2 have parametric joiners and snaps and stuff. And you could totally write modifiers for FreeCAD or Blender to do it, sure, but it’s a lot less trouble to get it done with OpenSCAD. This way your end-result would let you input the size of the bed and it would figure out how many sections it needs, etc.

Lesee… 120x40x40cm is a lot of plastic to print, even with a single printer running all day all night.

What I’d suggest is that you make the wooden outside for the 120x40x40cm shelf and make 20cm x 20cm x 40cm units. At which point you can make bigger multi-part modules. It might actually make sense to keep the cups separate because you could adjust the holes and stuff based on the plant’s needs. Whereas the reservoir section is going to be happiest as a single tub. But the important part is that if you are a few modules short, just add a spacer for this season. And it gives you more time to experiment on the tub and allows you to swap that out mid-season.

This is pretty darn ambitious for a starter project. I say this as someone who is trying to get some fancy new 3D printable tomato cages going before the plants get tall and dangerous and I’ve been doing this for a while.

So you really probably want to de-complicate this, either by only making planters that are sized for the printer or finding a existing planter that’s the right size but not self-watering and designing just the self-watering part. You’ll probably learn a lot about the right way to do one this year and then next year you can attack the next generation of the planters.

The problem with printing in pieces is that you are going to have to make sure that the joints are strong enough for the weight of the soil. This is why using a ready-made outer container might help. In the same way, what you really want is something finger-ish or jigsaw-ish so that the pieces align themselves more easily and interlock.

You will probably want a fatter nozzle, otherwise this is going to take forever to print.

PETG seems to have worked fairly well for me for outdoor stuff? Coating or paint or whatnot is handy. You might want to look at the epoxy family? If you can print on the balcony, you might consider ASA which is totally fine for outdoor use with no paint.

FreeCAD is a bit of a learning curve? The thing that FreeCAD would make easier is a parametric model, where you say that you want a 400 x 400 x 300 planter. Except that if you are really serious about making large self-watering planters that are parametric, you are going to end up wanting to write code to make it all happen, which either means the Python in FreeCAD, the Python in Blender, or maybe just use OpenSCAD.

One avenue, which is also too big of an ask for this season, is making a multi-part model to cast the large pieces in concrete.

Another avenue would be to just design around the outside being wood and the 3D printed parts being brackets and jigs and connectors and the self-watering bits.