www.transicionestructural.NET es un nuevo foro, que a partir del 25/06/2012 se ha separado de su homónimo .COM. No se compartirán nuevos mensajes o usuarios a partir de dicho día.
0 Usuarios y 3 Visitantes están viendo este tema.
https://www.youtube.com/watch?v=CgLwDbbledA&t=33sAquí un vídeo que encontré del bicho ese trabajando en la bmw. Da la risa ver como suelta la pieza casi de cualquier manera y la lentitud con que lo hace.No encontré vídeos en que hiciese otras cosas. Y si aún así en 11 meses está lleno de arañazos mal vamos
En mi opinión, tenemos que tentarnos las vestiduras con todo esto. Pero en el sentido de, ni creernos que todos esos avances son de hoy para mañana, ni lo contrario, esto es, que todo esto desaparecerá en dos días y no servirá para nada. Es evidente que estamos en una burbuja brutal, la mayor de la historia por volumen, pero también es cierto que lo que se está haciendo es llevar un poco más lejos lo que ya existía y funcionaba a nivel industrial - la robótica industrial o las redes de comunicaciones, o la estadística y la matemática computacional-. Miren, yo era de los que pensaba que Internet era una moda y me equivoqué. Con eso se lo digo todo.
CitarVideo: US humanoid robots retire with scars after helping build 30,000 BMW carsThe company revealed that its robots helped produce more than 30,000 BMW X3 vehicles and loaded over 90,000 sheet metal parts.Sujita Sinha · 2025.11.20Figure AI retired its Figure 02 robots. · Figure AICalifornia-based Figure AI announced on Wednesday that it is officially retiring its Figure 02 (F.02) humanoid robots.The decision comes after an 11-month deployment at BMW Manufacturing’s Spartanburg, South Carolina plant. The pilot project was part of a collaboration to test humanoid robots on a real assembly line.The company highlighted that the F.02 units helped produce more than 30,000 BMW X3 vehicles and load over 90,000 sheet-metal parts during their deployment.Figure CEO Brett Adcock shared visuals of the robots covered in scratches, scuffs, and grime, signaling the realities of working in an industrial environment.Scratches as proof of real workThe company shared video footage of the F.02 robots, clearly showing wear and tear from months on the line. Adcock described this as “real-world deployment.” The visuals serve as proof against earlier skepticism that Figure’s work at BMW was only a small-scale feasibility study. By showing the robots’ worn appearance, the company demonstrates that they operated for months on an active assembly line.The grime and scratches became an unintended badge of honor. They show the robots endured the repetitive and demanding nature of factory work. The company said the visuals and performance metrics together validate their claims of long-term deployment.Performance data from the trial runThe U.S firm released a detailed account of what the robots achieved at the facility. It said the deployment expanded quickly after the initial setup period. Within the first six months, the robots had been moved to Spartanburg and were operating on the floor. By the tenth month, they were running full shifts on the assembly line.Their primary task involved lifting sheet-metal parts from bins and placing them on welding fixtures with a 5-millimeter tolerance. After placement, traditional robotic arms performed the welding. The humanoids handled metal loading with a cycle time of 84 seconds, including 37 seconds for the load. Accuracy stayed above 99 percent, the firm stated in the press release.The robots completed more than 1,250 hours of runtime. The company estimated that the machines walked about 200 miles inside the facility. The deployment followed a 10-hour shift schedule, Monday to Friday. These metrics were presented as confirmation that humanoid robots can sustain industrial workloads for long periods inside active plants.Excited to share our F.02 robots have contributed to the production of 30,000 cars at BMWhttps://x.com/Figure_robot/status/1991178512510951782Lessons and the future of humanoidsThe robotics firm was also transparent about hardware challenges during the deployment. The forearm emerged as the main failure point due to the complexity of packing three degrees of freedom, thermal management, and cabling into a human-sized limb. Constant motion stressed the microcontrollers and wiring, a problem rarely highlighted in the humanoid sector.These lessons informed the design of Figure 03. The new model eliminates the distribution board and dynamic cabling in the wrist, with motor controllers now communicating directly with the main computer.The F.02 retirement marks a transition from pilot testing to larger-scale production. The company said, “Figure 02 taught us early lessons on what it takes to ship.”The fleet’s retirement clears the way for Figure 03, which the company claims is ready for scaled deployment.Battle Damange X-Blog-SocialsSaludos.
Video: US humanoid robots retire with scars after helping build 30,000 BMW carsThe company revealed that its robots helped produce more than 30,000 BMW X3 vehicles and loaded over 90,000 sheet metal parts.Sujita Sinha · 2025.11.20Figure AI retired its Figure 02 robots. · Figure AICalifornia-based Figure AI announced on Wednesday that it is officially retiring its Figure 02 (F.02) humanoid robots.The decision comes after an 11-month deployment at BMW Manufacturing’s Spartanburg, South Carolina plant. The pilot project was part of a collaboration to test humanoid robots on a real assembly line.The company highlighted that the F.02 units helped produce more than 30,000 BMW X3 vehicles and load over 90,000 sheet-metal parts during their deployment.Figure CEO Brett Adcock shared visuals of the robots covered in scratches, scuffs, and grime, signaling the realities of working in an industrial environment.Scratches as proof of real workThe company shared video footage of the F.02 robots, clearly showing wear and tear from months on the line. Adcock described this as “real-world deployment.” The visuals serve as proof against earlier skepticism that Figure’s work at BMW was only a small-scale feasibility study. By showing the robots’ worn appearance, the company demonstrates that they operated for months on an active assembly line.The grime and scratches became an unintended badge of honor. They show the robots endured the repetitive and demanding nature of factory work. The company said the visuals and performance metrics together validate their claims of long-term deployment.Performance data from the trial runThe U.S firm released a detailed account of what the robots achieved at the facility. It said the deployment expanded quickly after the initial setup period. Within the first six months, the robots had been moved to Spartanburg and were operating on the floor. By the tenth month, they were running full shifts on the assembly line.Their primary task involved lifting sheet-metal parts from bins and placing them on welding fixtures with a 5-millimeter tolerance. After placement, traditional robotic arms performed the welding. The humanoids handled metal loading with a cycle time of 84 seconds, including 37 seconds for the load. Accuracy stayed above 99 percent, the firm stated in the press release.The robots completed more than 1,250 hours of runtime. The company estimated that the machines walked about 200 miles inside the facility. The deployment followed a 10-hour shift schedule, Monday to Friday. These metrics were presented as confirmation that humanoid robots can sustain industrial workloads for long periods inside active plants.Excited to share our F.02 robots have contributed to the production of 30,000 cars at BMWhttps://x.com/Figure_robot/status/1991178512510951782Lessons and the future of humanoidsThe robotics firm was also transparent about hardware challenges during the deployment. The forearm emerged as the main failure point due to the complexity of packing three degrees of freedom, thermal management, and cabling into a human-sized limb. Constant motion stressed the microcontrollers and wiring, a problem rarely highlighted in the humanoid sector.These lessons informed the design of Figure 03. The new model eliminates the distribution board and dynamic cabling in the wrist, with motor controllers now communicating directly with the main computer.The F.02 retirement marks a transition from pilot testing to larger-scale production. The company said, “Figure 02 taught us early lessons on what it takes to ship.”The fleet’s retirement clears the way for Figure 03, which the company claims is ready for scaled deployment.Battle Damange X-Blog-Socials
Cita de: tomasjos en Ayer a las 11:00:31En mi opinión, tenemos que tentarnos las vestiduras con todo esto. Pero en el sentido de, ni creernos que todos esos avances son de hoy para mañana, ni lo contrario, esto es, que todo esto desaparecerá en dos días y no servirá para nada. Es evidente que estamos en una burbuja brutal, la mayor de la historia por volumen, pero también es cierto que lo que se está haciendo es llevar un poco más lejos lo que ya existía y funcionaba a nivel industrial - la robótica industrial o las redes de comunicaciones, o la estadística y la matemática computacional-. Miren, yo era de los que pensaba que Internet era una moda y me equivoqué. Con eso se lo digo todo.Parece que esta nueva generación de IA funciona bastante bien cuando es entrenada intensivamente sobre un conjunto de información muy grande y muy "bueno".Perdón por usar tantos términos coloquiales. No tengo otros.La primera red neuronal que conocí fue en 1991. Funcionaba regular pero es que contaba con un conjunto de datos ridículo. Como demostrador en el ámbito académico valía.Hoy conozco, de rebote, a gente que lleva varios años trabajando con ese tipo de IA de una forma bastante efectiva. Pero es que el volumen de datos perfectamente tageados que tienen para entrenar al bicho es monumental. Creo que ya lo he contado. Son radioastrónomos y tienen décadas de números recopilados perfecta y estructuradamente. La máquina aprende sin decirle mucho que cuándo pasa esto y aquello, allí hay una estrella de neutrones o una galaxia de forma de brazos espirales o es que hay un asteroide moviéndose. Básicamente le dicen lo que contiene un conjunto de datos, no tienen que explicarle por qué ni que forma o características presenta.Con la generación de imágenes pasa, creo, algo parecido. Por ejemplo, las caras las dominó desde un principio. Posiblemente porque las bases de datos de caras tageadas serían enooooormes. Ahora vemos que mucha generación trata de replicar la iconografía de las redes sociales. El influencer delante de la cámara, el "rant" en el coche, la foto del gim-bro o la gim-girl, la entrevista en la calle. Hay millones de referencias tageadas, auto tageadas o tageadas por el contexto, los metadatos o los datos del contexto que se pueden extraer.Si le pido una foto de un arrendajo vestido con una armadura italiana del siglo XV alejándose mientras camina por un puente de estilo racionalista, en perspectiva de tres puntos con el punto de vista elevado unos 15 metros sobre el suelo y con la iluminación difusa de un día cubierto por nimbos, se lía más.A mi internet me pareció genial desde el principio pero sí que es cierto que no pensé que el modelo de servicios gratuitos por publicidad tuviese mucho recorrido. Pensaba que se iba a quedar en el uso público, académico, corporativo y amateur y que los servicios de información serían de pago.Pensaba que las tablets eran un producto idiota y redundante.Otro de mis errores de predicción fue que vaticiné que la gig-economy no triunfaría.