Brian Ratliff

This is the future of Mechanical Engineering. It's already here. x.com/i/article/2120…

So Fable 5.2 just *gets* me. Built me a full glider.

I gave it a reference image of a glider body to go with the wing, and Fable in FaenoCAD created the body, wing-box, and vertical and horizontal stabilizer in a 3D printable assembly.

Unprompted, it cut a ballast compartment into the body and calculated an estimate for nose ballast to make it fly level.

This will fly. It'll print on my ungodly out-of-date Makerbot 2x. Probably need to add some trim tabs for the wings, but it's pretty fit.

I used to make gliders like this out of balsawood and index cards when I was a kid. I dropped the hobby when I went to school. Feels good to get back at it, enhanced by my engineering training. AI makes it even better.

Fun little project with Fable 5.2 building up a model glider wing in FaenoCAD. Fable 5.2 is a joy to work with. Very understanding of design intent.

I gave it general dimensions; Fable calculated the wing profile, taper, and added a washout twist.

Then it added a split and tab/key interface so this can be built on a 3D printer.

In addition, Fable was able to provide me with rough estimates to size a motor and how wing dimensions and weight traded off for motor power.

This is how agent-augmented CAD design changes the game. Research, calculation, design, manufacturing, all combine in the agent driven code-CAD. You can make this as elaborate as necessary to meet your goals.

Click a toggle to remove the tab/key joint and the model is ready for meshing. Same code, same source of truth--just a flag.

Interesting experience building in Solidworks through Claude Code.

Seems pretty capable. Could be integrated into the FaenoCAD family. The workflow is quite different.

Major difference is the script is never really the source of truth, which can be limiting.

At one time, I would’ve laughed at a conspiracy theory like this.

After my experience yesterday, with Solidworks literally taking >10 minutes to show and rebuild a drawing, *with both the part and drawing already up in my system*, I might be a believer.

Honestly, at one point I was thinking of moving my client over to FaenoCAD.

Maybe I should have.

@emm0sh, you asked me at one point how Solidworks dies. This is how it dies.

You have someone like me with clients who pay by the hour. I can’t reasonably charge someone for time I spend troubleshooting a reasonably simple part because of a back-end regeneration bug. I can’t, in good faith, charge for that cumulative hour I spent staring at a frozen Solidworks app, waiting for a drawing to regen. I can’t charge for the re-draft of a drawing page to implement a workaround.

My FaenoCAD system is open source, free of licensing fees, faster, better control, better UI, same outputs. So then, why Solidworks?

At some point, OCCT is good enough. Build123d on top of that—lots of improvements @Jern123d, amongst others, are making to these open source systems. Combined with Claude Code or Codex, it’s all good enough.

Parasolid might be the shit, but it’s not showcased well by Solidworks, and that’s how Solidworks dies.