José Siles | AI | Data

Soy oficialmente LinkedIn Instructor!

@polash_ai indeed!

@josesilesdata That's wild what AI can do now

🚨 THIS IS ABSOLUTELY INSANE 🚨 Someone just turned Claude Code and Codex into real CAD engineers. You describe what you want in plain English. The AI designs it as a fully functional 3D model. It’s called: text-to-cad. Before, building a robot used to require 6+ separate tools: → CAD software → URDF exporters → Kinematics solvers → 3D viewers → Slicers → Printer software text-to-cad replaces the entire workflow. It can: → Export STEP, STL, DXF, GLB and 3MF → Generate URDF robot descriptions → Solve functional robot kinematics → Slice models directly into G-code → Send print jobs to Bambu Lab printers → Run 100% offline with no backend One demo generated a complete 7-DOF robotic arm with working kinematics and a custom GUI almost entirely through prompts. It's 100% open source. Link in the comments 👇

@nahid_pro09 it's honestly amazing!

@josesilesdata tried it yet? does the kinematics actually work good?

Of course, arriving on time is always better. Plan the journey. Leave early. Expect delays. But confidence is not never making mistakes. Confidence is knowing how to recover from them without delivering a TED Talk about the traffic.

• Accountability • Emotional control • Professional maturity • Confidence under pressure • Executive presence Yes, apparently arriving five minutes late can now become a leadership assessment. What are interviewers really evaluating? Anyone can perform well when everything goes perfectly. They want to know how you behave when something goes wrong. Do you panic? Do you blame others? Do you make excuses? Or do you take responsibility and move forward?

You arrive 5 minutes late to a job interview. Low-value candidate: “I’m so sorry. Traffic was terrible, the train was delayed, Google Maps sent me the wrong way...” ❌ WRONG ❌ High-value candidate:

The Last 5 ㉖ "Take this response as 70 points and adjust it to 100 points" Improvement begins. ㉗ "Add the elements that are missing to make it the highest quality" The level of completion rises. ㉘ "Reconsider if there's an even better answer" You won't stop at the first response. ㉙ "Even out the overall quality of the response" Sloppiness in the latter half decreases. ㉚ "Keep refining right up to the end until it reaches 100 points" Strong in polishing to completion.

"Final Review" 5 Items ㉑ "Please do a self-check after answering" Reduces minor mistakes. ㉒ "Please check if there are any parts that look half-hearted" Uncovers sloppy sections. ㉓ "Please review if the quality drops midway" Ensures consistency throughout. ㉔ "Please confirm there are no logical leaps or insufficient explanations" Prevents omissions. ㉕ "Please read through once more before completion" Improves the final polish.

"Please summarize it ChatGPT" "Please createit, ChatGPT" ❌WRONG! That way ChatGPT generates AI slop. I compiled 30 "final words that keep ChatGPT from slacking off until the end." Copy-paste as-is OK👇

📎Repo link: github.com/earthtojake/te…