Andy Powell

@KennethCassel highly recommended. great fun. bought mine at 170k miles, sold it at 200k, missed it since 2007

posting so I can quote this when I finally get an e28 M5

@JonnyBird At this rate, they’ll be ground up and washed down the storm drain in two weeks. 😕

Two days in with the bollards on Forsyth Street:

@aphysicist one piece at a time

request for startup: the apple/tesla of cnc/machine tools. gorgeous modern software at the core (kill cam), proprietary integrated controllers w/ api, onboard simulation, adaptive feedback, can work hard metals (sorry haas), networkable (think self-driving cars), etc.

@NickPinkston @rob_lh @jack_watson_hfw I'm sorry to hear about this, and I hope there will be an opportunity to learn from their experience.

@rob_lh @jack_watson_hfw Market saturation is not what killed FormLogic, but I will defer to the team to post an after action report when they're ready and have thought it through. In general, it's more this problem is very hard & complex both technically and operationally.

This sucks. FormLogic had some of the best technology in the CNC automation space that I've seen. I wonder how the post ZIRP VC era will treat companies building deep tech in general. AI is still in its mega hype period, but that will likely have a correction pretty soon.

@rezmos1 @searchbound Asking the real questions 💯

@searchbound Was CallRail around back then?

omg ha, just found this. years ago, I bought domain 'CallTracking․com' for $20k and me&friend built a stripped down call tracking software application. I then attended a dude ranch conference to try and generate some leads. I got absolutely zero interest 😂

@zanehengsperger Love that he's all-in on toolpath automation. We're working to make this happen.

Q: what are some business ideas you have in the space? A: Automated toolpaths....

Q&A thread with exited machine shop owner. Anonymous, 30+ years in industry, 50ish years old.

@Nico_Campolongo @istvan_csanady It's too soon to be certain, but in our initial testing it seems to be significantly faster and more reliable. We do need the actual contours rather than mesh-based approximations.

@apowellgt @istvan_csanady I can imagine how that quickly becomes slow and painful with OCCT 😅 is it better in terms of speed with Parasolid? Do you really need to do those operations with a CAD kernel though? Couldn’t you just mesh the part and use the mesh instead?

We need a CAD component provider to become the AWS of industrial software. The closest today is Siemens (Parasolid and DCM), but I think there is a huge opportunity to make it more successful. Why? Because CAD kernels are infrastructure, and they should be treated as such. The first kernel provider to realize this will dominate the market in the next decade. In this post, I’ll use Parasolid as an example, but the same applies to CGM, ACIS, and all high-end proprietary kernels. 👇 “Infrastructure | ˈɪnfrəˌstrʌktʃə | The basic physical and organizational structures and facilities needed for the operation of a society or enterprise: the social and economic infrastructure of a country.” Infrastructure is typically: •Expensive, hard, and time-consuming to build. •Very hard or impossible to replace without significant service interruptions. •Provide a platform for a range of different services built on top of them (which also means that replacing them comes at the cost of replacing all services depending on them). Parasolid checks all these boxes, but it behaves unlike the most successful software infrastructure providers. Here is what needs to change to ignite a new industrial software renaissance: 1. Make it more accessible Today, starting a project based on Parasolid is not like starting a project on AWS. You need to negotiate a contract, explain your use case (which is usually unclear at the beginning of a new project), and go through their sales team—before you’ve made any money. This introduces enormous friction for launching a new project. The infrastructure of industrial software should be accessible at a very low cost until projects launch or monetize. In an ideal world, I should be able to pull a GitHub repository, start hacking a sample project, and have a working application an hour later. 2. Make it easier Parasolid is fantastic, but it’s very low-level. To build an application, you need to write tens of thousands of lines of code just to get Parasolid off the ground. From memory management to graphics, there’s a ton of code that offers no competitive advantage, yet every company has to write it. This could be solved by making Parasolid more accessible and letting the community build this layer as open-source. There’s a wide talent pool eager to contribute to creating such a layer, and it’s a fun, collaborative project. 3. Make the documentation publicly available The documentation should serve as a marketing tool for Parasolid. I don’t believe that the functional or API documentation of a software component provides any competitive advantage to anyone. By making it accessible to the developer community, it would act as an “appetizer” for enthusiastic engineers, inspiring them to build exciting new projects on top of Parasolid. 4. Build a community Since the Parasolid ecosystem is currently closed, it’s impossible to build a vibrant community around it. A community would make it dramatically easier to build new applications on Parasolid, by providing support and developing a shared application layer. 5. Open-source as much as possible This might sound scary or insane to some, but… I genuinely believe the value of Parasolid isn’t in its parts and algorithms, but in its role as the industry-standard infrastructure, creating a shared language between industrial applications. Making as much of it open-source as possible would engage developers further and increase trust in the platform. To be clear: I don’t think the licensing model should change. Parasolid should still charge for licenses, and I strongly believe forking should not be allowed to avoid diverging from standardization. However, some form of guarantee that customers could continue using the platform uninterrupted, even if something happens to Siemens PLM Components, would dramatically increase trust and likely attract more customers. I can even imagine a scenario where companies like Autodesk use Parasolid as their main kernel.

@Nico_Campolongo @istvan_csanady This will get beyond my knowledge, but our process slices the model many times and also projects many 3D curves onto planes. An error in a long chain of calls will break the analysis, and it needs to run autonomously -- the user doesn't have the chance to adjust and recover.

@istvan_csanady @apowellgt But interactive modeling and geometry analysis are 2 different use-cases. I believe that for geometry analysis the failure rate should be lower since you’re not modifying the part, just analyzing it. If you care about responsiveness/speed then sure

@SeeBQ @BobcatCompany @GetGreenzie Got my Bobcat CT4058 here south of Atlanta and love it.

Visiting our friends, partners, and investors at @BobcatCompany in Fargo, ND. It really is the Google of North Dakota: incredible engineering, great product, deep expertise. Our @getgreenzie equipped autonomous mower will be at Equip Expo and it is looking good!

@ggraham As someone who enjoys projects and also wants to build a small barn, I'm intrigued. But compared to a professional crew, aren't most people (me included) negatively useful?

I think about this idea a lot.

@zanehengsperger Thanks for stopping by @ToolpathLabs! Huge week showing off our software and building community.

cool mfg finds from chicago this week - robotic arm to reduce labor - $500k boring mill - 5 axis CNC - CAM automation software

@NickPinkston I'll be there all week, repping @ToolpathLabs in the machine hall. We've come a long way since IMTS 2022, come say hey!

ROLL CALL! Who's all going to IMTS next week?!? This is the biggest manufacturing trade show in North America, and it only happens every other year. If you're in Chicago, it's well worth a visit no matter who you are. Like a modern day World's Fair. imts.com

@searchbound Looking for ranches to visit at this very moment.

Dude ranch season rn 💤 😴 Vidalia season rn 💤 😴 College Football season rn 🚀 ⤴ 💥 👀 🏈 🔴🔵

@JonnyBird It's gorgeous, isn't it? I sent solid returns to investors from that teller line. Good times.

Original C&S Bank in downtown Atlanta:

@mitchellh Is the second sim running now? What a great experience.

Good morning to everyone except the flight scheduler which assigned me a 5 AM flight. 🫠 Waking up at 4 AM and diving immediately into systems discussions and training flights is rough. Silver lining I’m done for the day at 11 AM.

@KennethCassel I got my 10yo daughter a discovery flight in a 172 and I sat in the back seat. I was also interested in flying, and started flight lessons shortly after that flight.

Our 10yo is obsessed with airplanes. I know they have discovery flights for adults but is there something like that I can do with him?

@CollinMickels @Rollup Nice! We're fans of @JuliaLanguage at @ToolpathLabs. (Though we're more Glenn and less JPL.)

@searchbound @ERodz3 @CallRail too kind, thank you

@ERodz3 outside of SearchIgnite, my fav is prob @callrail from @apowellgt

March 2009 I'd taken a consulting gig at Fortune 500 corporation in atlanta, running web analytics, paid search, lead gen ~ all that nonsense . I had just bought v1 iPhone & vividly remember taking these photos from my cubicle & thinking, "is this it? is this my view for the next 10-20-30 years?" I had previously worked at web startups, and began to feel I preferred that level of semi-structured chaos. but... but this was nice. ample pay. abundant healthcare. chefs prepping lunch every day. 2 hour lunch breaks. something, though, that began to grate on me: buzzing-in with my badge every morning. the gate wouldn't open unless I did. a small reminder that I'm chattel. that they're tracking me. that I'm a number, and one step away from another layoff. when my 6-month consulting gig concluded, they extended a hefty 6-figure salary to bring me on Full-time. it didn't take me long to reject that shit. my cubicle friends were dumbfounded. i didn't care. I left & eventually ran into my old friend @roger_barnette who was building a new software startup, SearchIgnite, that served the growing paid search marketplace. "Holy shit, I know how to build this" I remember thinking. I essentially hounded him until he brought me on to continue building this (early) SaaS model aimed at AdWords, Overture, FindWhat, and the like (thank you again, Roger). it was, while working there, that I began teaching myself how to code, how to build, how monetize websites, and most importantly (for me) how to discover expiring domain names. sometimes these steps don't seem to make sense, until they do.

@nick_mccleery We have plenty! 😀

Damn, need to invent some Julia-appropriate work for myself now