Alex

@emm0sh BRep or mesh doesn’t work well for text either so there can’t really be an effective feedback loop. There’s more effective ways to tokenize a CAD design and provide an LLM tools to do stuff

i can’t get claude to send an email. i’m not letting it anywhere near my 30,000 part assembly

@gabednconfused @ashvarro Truth in 24 is the only answer. The rest are wheeled Top Gun

Whats the best racing movie of all time?

@emm0sh Does the unit system support slugs?

WORC out on monday some of my favorite features: - unit handling — no more figuring out what kg·m²·s⁻³ is - sharing — WORC is free and web based so you can instantly share your analysis with anyone. no sign in, no installation, just sharing

@CalebChamberla6 The zero sum mentality is so damaging. Rather than slice the pizza thinner let’s just make more pizza. It’s a total lack of ambition

It's all theater now, but they'd pass it if they could. This zero-sum, anti-success, parasitic ideology is bad for everyone.

If Elizabeth Warren's wealth tax passes, I might have to sell OSH Cut just to pay the annual tax bill. We are growing fast and reinvesting almost every dollar to better serve our customers. This absurd class warfare would be a wet blanket on innovation and change.

@__WORC__ Enhance

@CATIAManikin @gak_pdx All tables should have three legs as well. Tired of tables rocking at restaurants

Trying to explain that using only three mounting holes is actually better than four, because four can overconstrain the alignment of the assembly (I broke a tap in the fourth hole)

@emm0sh Basically CAD doesn’t explain the negative externalities it generates. All features are created equally but have unequal pricing. Understanding DFM is a containment but not a true countermeasure

this is a case where the CAD companies couldn’t actually make the product better, and it wasn’t negligence on their part the problem with “show the engineer their mistakes in CAD” is that “mistake” is subjective. not only does it depend on the machine shop, it depends on the person within the machine shop the reason why shops like sendcutsend and protolabs can give you fast CNC DFM is because they have explicitly developed their guidelines and no matter who you speak to, they have a hard line on what they can and can’t do this isn’t true outside of these services. everything is a negotiation. i can get my local machine shop to cut hard internal corners but i have to buy them the EDM machine first and this takes us to the fundamental problem with CAD, or engineering software in general being useful per DFM feedback — it has to be driven by the downstream operators (people who are doing the actual CAM and cutting the actual parts) this is nearly impossible because it requires: a. your manufacturers giving you up to date and reliable information b. that information being objective (it often isn’t) c. ingesting it into CAD quickly/naturally d. it being presented meaningfully in your CAD application of choice d is impossible because c would require an interoperable format from ISO or ASME (i’m currently working on this with the ASME MBE committee, so good news there) there are scammy-ish startups out there like bananaz purporting to be doing this but for the reasons i’ve listed they are not (i’ve already tried involving them to solve c and they are not interested)

@emm0sh Cunningham's Law

i’m not opinionated because i think i’m right, i’m opinionated because it is the quickest way to find out i am not

@paulcjh_ Congrats Paul! Great to see. Would be good to catch up

I've been hard at work the past few months in SF working on some buy buttons. Today I'm launching my plug-and-play product line that lets you build a robot real fast. Right now to make a robot you have to stitch together a bunch of different PCBs with jumper cables and wait weeks for blackbox actuators to arrive from China. You spend lots of time debugging why your CAN bus isn’t working, why every actuator performs differently, and meanwhile your wires keep coming loose. So, I'm making the following: - RB1: A robot main board powered by an Nvidia Jetson. This handles power distribution, compute, and a bunch more. - WM1: 2-channel wireless radio for sending video/data making the RB1 remotely controllable over USB-C. - M1: A pancake BLDC motor machined in-house. - ACB3: An FOC control board with matching connectors to the RB1. (big brother to ACB v2.0) - A1/A1m: A planetary/cycloidal actuator powered by the ACB3. (this is on the site in a couple weeks) Everything is on sale for this week, and shipping begins this spring! As a big thank you to the supporters of ACB v2.0 (and thanks for patience in shipping delays) you can buy the ACB3 for 50% off. If you like this kind of thing and want to join please DM me, I'm working solo right now and need good folks to join!

@gabednconfused @Aldas001 F1 simply didn’t exist prior to drive to survive. Many such cases

I've never seen a start to a season like this, outside Mercedes the amount of negativity from almost every single driver about these cars is just surreal We have to give them a chance because they could provide amazing racing but something about them feels underwhelming

@emm0sh Text to CAD is designed to impress VCs not anyone that’s ever flown CAD for a living. It’s a cumbersome way to attempt to define anything . It only has appeal in isolated toy problems. Otherwise it’s an attempt to disprove a picture being worth a thousand words

@eaconner @GovSherrillNJ @NJTurnpike This feels like an expensive way to discover induced demand

.@GovSherrillNJ: The @NJTurnpike is lying about the need for an expensive two-span bridge across Newark Bay. These additional lanes will worsen local traffic and hinder Jersey City and Bayonne’s growth. If you want to address traffic, implement dynamic demand-based tolls!

@Aeden6007 @nick_mccleery I think it’s likely a function of the tyres changing and possible larger wheel size? Potentially the loaded wheel no longer needs as high a slip angle. Nick I think you have more VDG experience than me so I’ll defer to you.

@nick_mccleery What changed? The inverse Ackerman made sense to me for high speed, high G cars.

Great photo of surprisingly normal looking Ackermann geometry here. Not that long ago it was all this sort of thing

@JulianFried I enjoy that the 20% correct score lines up perfectly with what you’d expect from blind guessing

The correct answer is: D C C D B

“AGY has been achieved internally”(Artificial General Yapper) The models still can’t rotate basic shapes. No luck with o3-mini-high or Grok 3.

@nick_mccleery @gabednconfused This one is worth a community note

@gabednconfused Real Gs always just took the orange juice (or stayed at their desks, because the work is the fun bit)

@emm0sh @robirwinengr DS won’t approve their SW Add-On so it won’t matter.

@robirwinengr i reached out via dm but they aren't giving me access to try it out

Some interesting features shown here, sketching on screen to change 3d part could be a nice time save, but that could be fixed with snappier CAD software too

@lavitalenta It is almost Sunday

@EverestBrady @EdwardMehr However there needs to be an undercut for rolled threads for space for the run out of the tool. I mistakenly thought your undercut was for this.

@EverestBrady @EdwardMehr Nothing to apologize for! Rolled threads are stronger as the metal is cold worked rather than cut like single pin threading

Meet Mitch; we call him Moose. He engineered and built all the fluid systems for our new deployable robotic cell. This system can be deployed anywhere in the world and start manufacturing parts within a day of arrival. Let’s make this a thread: show your work. Talk is cheap 😉