Waterjet cutter demo: 30,000 PSI meets aluminum

A robotic hand that’s smaller than an iPhone, yet can thread a needle, lift 20 kilograms, and weighs only 383 grams...and never gets tired! Autodiscovery’s new Revo-2 dexterous hand moves like us, grips like us, and is designed to handle the same objects we do. Which makes sense, after all, the entire world is built for human hands. Piece by piece, we’re cloning ourselves. Not just copying the shape, but improving it: steadier, stronger, more precise. And one day, these won’t look like robotic claws or metal parts. They’ll be made of silicon, designed to look and feel just like our own hands. That’s when it gets strange. Follow Endrit Restelica for more.

When it comes to a post labor world, robotics play a huge role in job automation and displacement but is talked about much less than AI. 👋🏻🤖 #robotics #automation #payattention #futureofwork #manuallabor #dexterity #AI #PostLaborEcononomy

🚀 At #AppliedMaterials, we don’t just build tools—we engineer the future of electronics. Our cutting-edge systems fine-tune wafer surfaces with extreme precision, adjusting temperature, pressure, plasma, and more to create the perfect material layers. It’s innovation at the atomic level—powering everything from smartphones to supercomputers. 🔬 Curious how we do it? Dive in: https://bit.ly/3KBVeW5 #Semiconductors #MaterialsEngineering #TechInnovation #WaferMagic #FutureElectronics

The first iPhone wasn’t built for high-speed automation. It was built to prove the idea. By 2025, the design looks totally different. Slim. Modular. Scaled for millions. Medical devices follow the same path. Pilot builds rely on semi-automated fixtures, manual steps, and flexible tooling. Great for proving a concept. Not for global demand. Production is a different game. Automation must be designed in from the start: - Material flow → automated feeding, nesting, packaging - Precision assembly → robotics, vision, in-line QC - Traceability → every step digitally captured for compliance - Throughput → scalable cells capable of running 24/7 without bottlenecks Companies that try to stretch pilot automation into production face the same risks over and over: -- Delays in launch -- Cost overruns -- Quality issues that stall approvals -- Lines that cannot keep up with demand The iPhone did not get sleeker by accident. It was redesigned for manufacturability. Medical devices must go through the same transformation, because production success depends on it. #MedTech #ManufacturingExcellence #Automation #SmartManufacturing #LifeSciences #ProductCommercialization #PilotToProduction

A robotic hand that’s smaller than an iPhone, yet can thread a needle, lift 20 kilograms, and weighs only 383 grams...and never gets tired! Autodiscovery’s new Revo-2 dexterous hand moves like us, grips like us, and is designed to handle the same objects we do. Which makes sense, after all, the entire world is built for human hands. Piece by piece, we’re cloning ourselves. Not just copying the shape, but improving it: steadier, stronger, more precise. And one day, these won’t look like robotic claws or metal parts. They’ll be made of silicon, designed to look and feel just like our own hands. That’s when it gets strange.

Electronic packaging is the invisible backbone of modern technology. Whether it’s shielding smartphones, managing heat in fighter jets, or protecting circuits on satellites, the way we package electronics determines performance, reliability, and even sustainability. This video explores how innovations in advanced materials, thermal management, hermetic sealing, and even microfluidic cooling are transforming the field. 🌍 Why it matters: packaging is no longer just protection — it’s an active part of the technology. 📺 Watch the video: https://lnkd.in/g2ndfYbz 📖 Read the full article: https://lnkd.in/geZnbcSQ #ElectronicPackaging #ElectronicsDesign #AdvancedMaterials #MilitaryElectronics #FutureOfTech #DefenseInnovation

A robotic hand that’s smaller than an iPhone, yet can thread a needle, lift 20 kilograms, and weighs only 383 grams...and never gets tired! Autodiscovery’s new Revo-2 dexterous hand moves like us, grips like us, and is designed to handle the same objects we do. Which makes sense, after all, the entire world is built for human hands. Piece by piece, we’re cloning ourselves. Not just copying the shape, but improving it: steadier, stronger, more precise. And one day, these won’t look like robotic claws or metal parts. They’ll be made of silicon, designed to look and feel just like our own hands. That’s when it gets strange. Follow The AI Colony for more. #ainews #breakingnews #aitools

Shipped something fun: From your video to captions in minutes, if you know how to load up colab ;-) A Hugging Face demo that takes your videos, pulls frames, and captions them using Apple's FastVLM-0.5B model. Runs in Colab, so you can open it up, see how it works, and tweak it yourself. Love how this embodies everything we're about here at Renesas Electronics and #RCar — open source, transparent, and actually useful. If you squint hard, you can imagine what else is coming through in the pipeline from us :-) You can literally see every step of the process and modify it however you want. Been testing it on talks, clips, even family videos — surprisingly solid for such a small model. 👉 Check it out: https://lnkd.in/gqbDdkJB What are you all building with vision models these days? Curious to know. PS: Credit to NVIDIA - I used their warehouse video to test this codebase and generate the output - https://lnkd.in/gd_bwcn6 NVIDIA Robotics

Apple is pushing into AI hardware with a Siri-powered robot, smart displays, and home-security devices. 🧠 🍎 Apple is preparing a big AI hardware push with a tabletop robot featuring a smarter Siri coming in 2027, a smart speaker with a screen next year, and new home-security cameras. These devices will run on a new OS and tie into Apple’s ecosystem, but critics warn the products seem late, lack unique features, and may struggle to compete unless Apple delivers something truly new. This and more on the Tech Field Day News Rundown with Tom Hollingsworth and Alastair Cooke. #TFDRundown #AppleAI #SmartHome #AIHardware #AIInnovation

🚦 Driving WS281x LEDs directly with AVR Assembly video https://lnkd.in/dEQ8esWJ For most makers, turning on the first LED of a WS2812/WS2816 strip isn’t a big deal — libraries like FastLED or Adafruit NeoPixel make it as easy as a single function call. But for engineers, it can be fascinating to dive into the protocol details and see how precise timing at the nanosecond level is handled directly in AVR assembly. ⚡ WS2812 first LED: ASM vs FastLED ⚡ A minimal routine in pure AVR assembly to light the first WS2812/WS2816 LED takes only ~200–300 bytes of flash. The same job with FastLED (C++ templates, chipset drivers, buffers) ends up around 10–15 KB. 👉 That’s about 40× larger. Assembly is ultra-compact and educational (cycle-counting at the nanosecond level) Sometimes it’s worth seeing both sides of the trade-off. Below is a minimal routine that sends out a 24-bit “white” frame to the first LED of a WS2816. At 16 MHz, carefully counted cycles of nop, sbi, and cbi instructions generate the strict timing required: T0H ≈ 200–320 ns T1H ≈ 520–800 ns Bit cycle ≈ 1.25 µs Reset > 50 µs ; first LED on ws2816 high white .org 0   rjmp init    ; reset vector -> jump to program start .org 0x68      ; (arbitrary start address depending on flash map) init:   sbi 4,0     ; set bit in I/O register: DDRB0 = 1 (set PB0 as output) loop:   ldi r17,24    ; number of bits to send = 24 (1 LED x 24 bits) sus:   ; --- BIT "1" routine ---   ; total time per bit = 1.25 µs ± 250 ns (valid range 1.05 – 1.45 µs)   sbi 5,0     ; set pin high = start of bit   ldi r16,14    ; delay for HIGH period (~700 ns, range 500–850 ns)   rcall ld_loop  ; wait loop for HIGH   cbi 5,0     ; clear pin low = end of HIGH   ldi r16,12    ; delay for LOW period (~600 ns, range 450–1000 ns)   rcall ld_loop  ; wait loop for LOW   dec r17     ; decrement bit counter   brne sus     ; if not zero, send next bit   ; --- Reset frame ---   ; needed to latch data into WS2816 LEDs   rcall ld_loop  ; here used as reset delay (>50 µs)   rjmp loop    ; repeat the process forever ; ---------------------------------------------------- ; Subroutine: ld_loop ; Provides a simple delay ; Each iteration ≈ 50 ns (at 16 MHz) ; Input: r16 = loop count ; ---------------------------------------------------- ld_loop:   nop       ; no operation (~62.5 ns each at 16 MHz)   nop   dec r16   brne ld_loop   ; repeat until r16 = 0   ret This exercise shows how “bit-banging” at nanosecond accuracy is still possible with an ATmega328, even without dedicated peripherals. It’s a great reminder of how much control you get when you understand the hardware at the instruction-cycle level. 💡 Even better: this exact code is already included in the online compiler/uploader at 👉 https://lnkd.in/dBDJn3rq (WebUSB-based). That means if you have an Arduino Nano and an addressable WS2812 LED, you can try it right away — no need to install any IDE!