> Some FAANGs are doing amazing things with unlimited tokens.
Would love to know what things!
14,009 karma · joined June 4, 2013
Would love to know what things!
Design needs theory to be intentional. It can of course be accidental (”seems to work, I guess”) or intuitive (”i know in my guts this is right but cant explain it”).
While both can end up with functional systems, if you cant vocalize the design journey the system is not very maintainable in the industrial sense (hence - theory is the vocalization of the design and the forces that influenced it).
Another point of view is that LLM:s perform to an extent on the same level as outsourcing does. This interface requires a bit more contract mass than doing everything within single team.
This is a separate dimension to alternative high quality modeling solutions alone.
Now, some of the users especially are _proud_ of their product specific skill set. They don't _want_ to switch a package.
And - it's much easier to get professional engineers to use extensions to packages their engineering office already uses.
And this comes before any technical side-by-side feature comparison.
Basically same shaped taylorism-derived industrial management has imposed itself as the "default dogma" in private and public administration.
"Hey give me a comic of how to create a rocket engine i can build at home"
Unlimited creativity will be shackled by safety.
Still pretty amazing.
The article summarizes the functional parts so well. What is very hard to communicate is the feeling of space, especially in the top floor with the books. It's sort of unique, and recommend a visit anyone traveling nearby.
I worked at the company that developed the software used to design the construction of Oodi (Trimble/ Tekla Structures). It's so awesome to walk through a building you know the tool you helped to build, helped to build :D
Consumer don't care if the OS is proprietary, as long as it works and there is a responsible party they can trust to serve them the offering.
The blog is running already on hugo on full llm automation but I had though it would not work for documentation (this is for non-techies so want something more product-manual -like and less SDK-docs flavour) or landing page that well.
So this is you company’s site and it’s on Hugo? https://formationxyz.com/
I don't think there is need to unseat anyone. 3D modeling market is expanding and the intent is to serve people for who are not users of current market leading tools. There are tons of plausible UX paradigms that have not been explored. This is one such exploration :)
"like could you make it aware of the fdm limits and help me avoid them while im building "
That's part of the long term vision. First solve modeling, then solve manufacturing of the models reliably.
The modeling already follows this principle - you can't model things the rest of the operations can't support.
I guess you mean https://github.com/oxc-project ?
Are there benefits compared to using, say, protobuf?
While https://xkcd.com/927/ is something to be avoided there really isn’t an existing format I would be aware off that fits the bill.
I’ll add this to the roadmap.
But to do this right it really needs to be a properly specified schema, with conformance test suite and versioning.
I appreciate your use case, but for users coming out of context, parsing underspecified XML files would be more of a curse than blessing :)
I agree supporting ”pointy” input modalities - pens and touchscreens is valuable.
I don’t think there will an android port in the near future but appreciate the sentiment.
And specifically ”whatever xml”, not, say STEP, or , openscad code, python code (via using cadquery) or anything like that?
Do you know your final output format?
If you are inclined to continue testing dropping the resolution or making the window smaller _might_ help. Also it's expected the user has a SSD.
My other testing platforms are a desktop rig with 4k screen/ 3080 GPU and Windows Sandbox (the latter being super sluggish). I've not tested on igpu with WQHD+ resolution - will definetly add this to my test matrix in the future. But don't know if I can help you right now.
This is great feedback btw. for alpha version regardless to whatever conclusions you come on the applicability.
STEP XML would be probably the closest here.
To be realistic, I might wrap a CLI to AdaShape first, then the user could query the model and have their LLM backport the model tree to something like CadQuery :D
The output resolution as such can be made "arbitrarily" precise if the model geometry is authored within AdaShape. So the facets in your image would not result from the limitation of the generated mesh (https://www.cnccookbook.com/wp-content/uploads/2017/08/facet...).
There resolution is currently fixed to presets for usability (see p. 28 of the changelog for the tolerance values - https://github.com/AdaShape/adashape-open-testing/releases/d...). I did not have CAM/CNC expert to consult on the details so those may be out of whack (but I'm happy to adjust them or add a user configuration).
"if it's a good fit, maybe I can take the manual off your hands?"
Would redistribution under CC BY 4.0 suit you?
I will definitely keep plugging at this :)
(Nothing against Mac as such, let's see if I will have bandwidth for a port at some point).
The data model supports solving dependencies like this on the data level.
TBH I don't know at the moment how to make something like this accessible but will definitely keep it in mind.
> As a person who has crashed and burned with every. single. traditional 3D CAD tool
I hear you, there are reasons for depth and complexity but not every program needs to be like that.
>if this is a good fit, will gladly pitch in using this for 3D.
I notice you are discussing specifically CAD/CAM for CNC routers. I don't know if this is applicable for your use case or not. Would be very interested to hear your opinion!
The output is a tessellated 3MF mesh. The tessellation accuracy can be tweaked to be as precise as needed, so if that's the only constraint this may be applicable.
Thank's for raising the manual! I'll have to invest more time into it :)
The CAD kernel is written by me apart from the boolean solver for the meshes which is the superb https://github.com/elalish/manifold
To explain a bit more as "do your own kernel" is usually considered more mad than mad-science - this is not done on a whim. I spent over a decade doing CAD at Trimble, developing base tech and CAD offerings (including Tekla Structures and SketchUp). Happy to discuss the architecture more.
OpenCASCADE is included as part of STEP importer though.
Solvespace is a nice reference! One can already use it as prestep to modeling - just export the output as STL or SVG and import it :).
Sketching 2D shapes is a very natural way to start thinking about shapes.
That's why there is a sketch + extrude.
Here are few examples - 42 seconds to a desk organizer
https://www.youtube.com/watch?v=VX6g5slTdeE
Or quick wavy vase.
https://www.youtube.com/watch?v=IkhAUhlg81s
The booleans and extrusion shapes are complements - both fitting different type of modeling.
Booleans are not only about shaping but also about composing individual parts to more complex assemblies.
So one can extrude few parts, then eg. combine them with a join.
The codebase itself is architected and documented to be LLM friendly and claude.md gives very strong harnesses how to do things.
As architect Claude is abysmal, but when you give it an existing software pattern it merely needs to extend, it’s so good it still gives me probably something like 5x feature velocity boost.
Plus when doing large refactorings, it forgets much fever things than me.
Inventing new architecture is as hard as ever and it’s not great help there - unless you can point it to some well documented pattern and tell it ”do it like that please”.
OTOH WPF is today surprisingly strong GUI platform if you just want to get your Windows GUI out there.
It runs really nicely even on low end hardware. All the nice styling and blending techniques now _just work_ even on the most cheap low end laptop.
The fact it's over decade old means all the LLM:s actually know really well how to use it.
So you can just guide your LLM to follow Microsoft best practices on logic development and styling and "just add this button here, this button here, add this styling here" etc.
It's the least annoying GUI development experience I've ever had (as a dev, non-designer).
Of course not portable out of the box (avalonia is then the ticket there).
If you want 3D, you can just plug in OpenTK with OpenGL 3.3. Decades old _but good enough for almost everything_ if you are not writing a high perf game.
Really, WPF plus OpenTK is a really robust and non-surprising development platform that runs from old laptops (eg. T14 Gen 2 per my testing) onwards.
I've been doing a sideproject using WPF and OpenTK - .net works really great - here is a sample video of the whole stack (from adashape.com)
More like Catch-22 than a cheap ”spill the tea” ride.
54 years.
I hope we as humanity never stop again.
Good luck!
A Unity expert can correct me, but IIRC (possibly wrongly) at least the following limitations apply:
For example Unity does not have accessibility features (screen readers etc) nor I don’t think it’s DPI aware. I would _guess_ it does not support platform fonts. Not sure if it supports non-latin font layouts like arabic. Etc etc.