Jakub Kotowicz

@barbinbrad @neilkanakia That's such a great UX/UI! Do they light up accordingly on saloon/table row hover?

we have ballooning! this was a community contribution from our good friend @neilkanakia the beautiful part of this is that it integrates like 🧈 into the shop floor, NCRs, inbound inspections, etc, etc, etc -- great for OEMs, great for parts manufacturers

@mr_sudo_ @AXL_Labs You dodged the main question. What kind of "DATA" you want to use to create truthful representation in that exploded view? For that demo you just used pre-made 3D model, right?

@KubaKotowicz @AXL_Labs you have a great point, the scene is only as good as the data we provide it, working on ways to get this context into the models before it generates

instant pull-aparts! see exploded views of everyday objects with liquid photos. created at @AXL_Labs

@mr_sudo_ @AXL_Labs Oh cmon… what kind of data? Repair manual with drawings? Schematics? 3d models or scans? Teardown videos?! That gamepad looks accurate-ish, but is the PCB layout real? If so, how does it know? If it’s not mostly accurate, then it’s party trick. Really cool UI/UX exploration tho

@KubaKotowicz @AXL_Labs DATA. the reality is the models need accurate references to compose these with higher precision

@theonecid @MKBHD It’s not glass, cause in SF luggage would get stolen every other ride

@MKBHD Why would they not give it a rear window if it can be driven by a human 😭

I’ve been in Texas for maybe 10 minutes

@repkord woundt it be more space efficient ot have double depth rack and have printers back to back? Or do you need back access that often?

"I HAVE THE POWERRR!"⚡️ Waited 2 weeks for this bad boy to arrive but now I can finally run rack 2 at full capacity (15 printers/rack) 5ft long, 28 isolated plugs, real time monitoring, she ain't your typical power strip. And no, there's never enough power in the shop.

@VishakhRanotra is it just me or their videos are way lower quality now? I loved their in depth car teardown, especially those hosted by Sandy himself or Cory, but lately most videos are quick and kinda shallow imho

Munro Live just did a tear down video of YASA’s axial flux motor. Very interesting piece of engineering. youtu.be/m507ryWhc6c?si…

@Scav Ofc. Few simple „obvious” functions for note taking app, simple interface, robust sync for coffee price, multiplatform. Did people switch already to some new, shiny things?

Is anyone actually still using Obsidian?

@schiffer_soft @PRUSA @josefprusa @tommy_prusa3d @PrusaGuy @Prusa3D Cool exploration, but not sure if app on a touchscreen isn't better for that specific use case (complex, multilayered UI). I use Stream Deck Mini for QC actions control and this is where it shines - tactile input for hands-on work, no need to fumble with keyboard/mouse

@Prusa Connect on a Stream Deck?? 👀 What started as a small idea has since grown and will soon be ready for testing. If you're interested, just let me know. 😅 Follow me for more updates...😊 #StreamDeck #Prusa #PrusaConnect #schiffer_soft

@emm0sh fucking legend. Would love to hear more insider stories from him

here's my conversation with Carl Bass. we talked about CAD, engineering software, AI, Autodesk and organizations i hope you enjoy it as much as i did

@Whyman5Whyman @simonkalouche @GoingBallistic5 FWIW second camera might be way cheaper than the cheapest humanoid or cobot arm. Heck, you might actually buy few industrial grade cameras and still be 50-100k up

@simonkalouche @GoingBallistic5 The robot still isn’t doing the same job. Also, how many facilities have cameras top and bottom?

The task is to flip the packages label side down if necessary, not move to a parallel conveyor Your example cannot do that First understand the nature of a problem before proposing a solution

@Taylor_64 @OpenAI If it was really unnoticed for 60 years, then it still outperformed humans that were searching for solution

@OpenAI Will this turn into another situation where we learn in a week that the model actually just copied the solution from a 60 year old paper that went unnoticed at the time?

Today, we share a breakthrough on the planar unit distance problem, a famous open question first posed by Paul Erdős in 1946. For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids. An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better. This marks the first time AI has autonomously solved a prominent open problem central to a field of mathematics.

@shinjipons @codyaims @kyleanthony @bscholl @jimbelosic You miss the point! It’s corny on purpose. As the service itself, it’s inviting, straightforward and a little bit silly - stark contrast to corporate behemoths that won’t even answer quote from normies. Almost “undesigned” on purpose!

Good idea but bad execution on all counts. Overly detailed, illegible at small sizes, the link to the company is too soft. It could be more appropriate to use that logo for a video games studio focused on first-person shooters. The company is great no doubt but the brand does not do it justice.

We need to rebrand this company ASAP

@mehul @drnafizhamid Is the robot’s view cropped to this height by hardware (cameras FOV) or is it “cropped” on-board before leaving the device? Or do you get full 360 mapping in such debug clip?

@drnafizhamid This is great to hear Nafiz. Would you please press the camera icon in the app for robot to record a clip of those areas you mentioned and DM me your robot ID. We'd love to look at those clips from robot POV and debug on our side to see why it's doing it. Thanks!!

most humanoid demos are form looking for a function (problem) we invented machines to escape the limits of human form now we’re putting human form back into machines that’s odd humanoids will win when they’re built for a function that suites the human form until then…alas

@Ruth_Kant_ @simonkalouche Tbh if VLAs gets solved, you can mount cobot arm with a a camera and jetson and tell it what to do just as well. Question is: is this setup still cheaper than universal humanoid produced at bigger scale?

@simonkalouche Application specific bots are great if you can justify the initial setup cost for the specific production line, and possibly the limited resale value once the specific task has been finished. Humanoids just walk to whatever needs to be done and then do it, no setup required.

Unfortunately Frank, Rose, Bob and Gary the humanoids all just lost their jobs to a pair of faster, lower cost, more reliable robot arms. Superhuman robots will beat humanoids for every industrial application with scale.

@simonkalouche @shai_machnes Personally im still on the fence, but what if humanoid is 3x cheaper than faster, specialized one? Exactly cause it’s versatile, it can have enormous economy of scale. With factories producing single model. It might end up cheaper than even 6 axis arms even tho it’s more complex

Yes which are most industrial tasks across high volume manufacturing, logistics, agriculture, construction, food prep etc. Each of these industries will be dominated by industry-generalist superhumanoids: robots that are powered by a general brain with bodies optimized to be general purpose and superhuman at performing all or most tasks within that specific industry.

@EERandomness @simonkalouche If autonomy in all form factors is solved then the question is: will humanoid be cheapest form factor thanks to biggest scale? Cobot arm uses 1/4 of actuators, but if its’ production scale is way lower (cuz less versatile) it still might be more expensive

@simonkalouche This completely misses the point. If you have the capital and skill to implement a specialized robot, you probably already have one. Humanoid robots are versatile. The same humanoid can hand out fast food or build a car, no custom engineering required.

@CircumjovialLLC @OpenAI You do realize that this AI model is human achievement itself, right?

@OpenAI This is not a human achievement. We exist for human achievement. A world where we are all passive observers while AI models do all the achieving is a nightmarish dystopia and humanity can choose not to take that path.

@mehul @jimbelosic did you use SCS for prototyping or production? I guess it was more 3d printing and now injection moulding for production, but maybe some brackets/structure inside?

This so great to see... company built the right way to reindustrialize America and now firepower to scale ecosystem around it. Love... love to see! Congrats! @jimbelosic.

@ti_morse @jimbelosic @sendcutsend Why this episode is not on Spotify? I prefer it for long form as I can lock the screen on phone and its generally better at syncing progress between devices

My second interview with @JimBelosic, Founder of @SendCutSend. 0:15 Beating China in manufacturing requires scale 1:10 Automation is not a magic bullet 2:12 Scaling profitably 4:31 Delivering a great customer experience 8:44 Why SCS raised $110M 11:20 Expectations are promises 15:33 Increasing speed and capacity 18:46 Focusing on customers 22:20 Frugality and generosity 25:06 Creating The Anything Factory 26:33 3pm is the new midnight 27:49 Figuring out the next bottleneck 30:37 Finding buildings with enough power 35:21 Lowering prices 37:36 Gambling to turn on factories faster 39:03 Aim for perfection and you’ll end up at excellence 40:04 Maintaining a maniacal sense of urgency through impatience 41:30 Seeding SCS DNA at new factories 43:08 Solving challenges 45:32 Finding great capital partners 48:40 Focus on customer pain 51:59 Learning from Home Depot 54:16 Bringing manufacturing back to America 55:11 Building capacity creates demand 57:01 Handling demand surges 58:21 Being default skeptical and sensing bullsh*t 1:00:27 Getting back to building