PaliPali

Over the years, my passion for design thinking and innovative frameworks has shaped my career, especially through my extensive experience in developing high-performance gaming laptops. This background has led me to believe that leveraging laptop technology as the foundational “brain” of humanoid robots could be a groundbreaking concept.

When I first entered the computer development field, I dreamt of the day we might create a humanoid robot akin to Astro Boy. Inspired by that vision, I purchased microcontrollers and servomotors, teaching myself to code in hopes of building a toy robot prototype. However, witnessing Honda’s impressive Asimo robot in the early 2000s made the dream seem distant, and I temporarily set it aside.

In recent years, the rise of AI has reignited interest in humanoid robots, and I’ve realized that perhaps, without knowing it, I’ve been contributing to their development. The core technologies we’ve refined in laptop design—high-speed computing with CPUs and GPUs, advanced camera systems, microphones, speakers, and wireless communication—align perfectly with the requirements of a humanoid robot’s head.

Furthermore, As the critical experience of integrating software, hardware, thermal, and power seamlessly. This coordination is fundamental to humanoid robot development, where complex systems must operate in harmony. Perhaps I’ve been on this path all along, preparing to develop the “brain” of future robots.

Currently, many humanoid robot companies originate from the automotive or robotic arm industries. Taking Honda’s approach as an example, they focused on developing legged locomotion and balancing systems before tackling the robot’s control center and AI. In contrast, if we start with the computational power and control systems perfected in laptops, we might achieve breakthroughs more swiftly.

copy from Honda website

AI development has three layers: large language model training, fine-tuning, and inference. I believe that for humanoid robots to thrive, they need computational power capable of real-time fine-tuning, a capability that mobile laptops may achieve in the coming years.

When we reach this milestone, strategic alliances with automotive and robotics manufacturers could pave the way for a new, innovative R&D model. This collaboration might finally bring humanoid robots into consumer markets and everyday family environments.

• -列子湯問–

周穆王西巡狩，越崑崙，不至弇山。反還，未及中國，道有獻工人名偃師，穆王薦之，問曰：「若有何能？」偃師曰：「臣唯命所試。然臣已有所造，願王先觀之。」穆王曰：「日以俱來，吾與若俱觀之。」越日，偃師謁見王。王薦之曰：「若與偕來者何人邪？」對曰：**「臣之所造能倡者。」**穆王驚視之，趣步俯仰，信人也。巧夫，顉其頤，則歌合律；捧其手，則舞應節。千變萬化，惟意所適。王以為實人也。與盛姬內御並觀之。技將終，倡者瞬其目而招王之左右侍妾。王大怒，立欲誅偃師。偃師大懾，立剖散倡者以示王，皆傅會革、木、膠、漆、白、黑、丹、青之所為。王諦料之，內則肝、膽、心、肺、脾、腎、腸、胃，外則筋骨、支節、、皮毛、齒髮，皆假物也，而无不畢具者。合會復如初見。王試廢其心，則口不能言；廢其肝，則目不能視；廢其腎，則足不能步。穆王始悅而歎曰：「人之巧乃可與造化者同功乎？」詔貳車載之以歸。夫班輸之雲梯，墨翟之飛鳶，自謂能之極也。弟子東門賈、禽滑釐，聞偃師之巧，以告二子，二子終身不敢語藝，而時執規矩。

人形機器人：植基於筆電技術的下一個科技前沿

多年來，我對設計思維與創新架構的熱情，深深影響了我的職涯，尤其體現在我長期參與高效能電競筆電開發的經驗上。正是這樣的背景，讓我逐漸相信：將筆電技術作為人形機器人「大腦」的基礎，可能是一個具顛覆性的創新概念。

當我初入電腦開發領域時，曾懷抱著一個夢想——有朝一日我們能打造出類似《原子小金剛》的擬人型機器人。受到這個願景的啟發，我買了微控制器與伺服馬達，自學程式語言，希望能親手打造一台機器人玩具原型。然而，當我在 2000 年代初期看到本田（Honda）推出的 Asimo 機器人時，那樣的夢想顯得遙不可及，我也暫時將它擱置一旁。

這幾年，AI 的崛起重新點燃了人形機器人領域的熱潮，我也驚覺：或許不經意間，我一直都在為它們的誕生鋪路。我們在筆電設計中所累積的核心技術——高速 CPU/GPU 運算、進階相機系統、麥克風、喇叭與無線通訊——恰好與人形機器人「頭部」的技術需求完美契合。

更重要的是，我們具備整合軟體、硬體、散熱與電力系統的關鍵經驗。這種跨系統協調能力，是人形機器人開發中不可或缺的基礎，因為它們必須讓複雜的子系統協同運作。也許我一直走在這條道路上，只是現在才意識到——我們正為未來機器人的「大腦」打下基石。

目前市場上多數人形機器人公司，來自汽車或機械手臂產業。以本田為例，他們的策略是先專注於雙足行走與平衡系統，再逐步開發控制核心與 AI。相較之下，若我們從筆電領域已臻成熟的運算能力與控制系統出發，或許能更快速實現人形機器人的關鍵突破。

AI 的發展可分為三個層次：大型語言模型的訓練、微調（fine-tuning），以及推論（inference）。我相信，若人形機器人要真正蓬勃發展，就必須具備能夠進行即時微調的運算能力，而這項能力，在未來幾年內，有望由行動筆電實現。

當我們達成這項里程碑後，若能與汽車或機器人製造商建立策略聯盟，將有機會開創出一種嶄新且創新的研發模式。這樣的跨界合作，或許能讓人形機器人真正進入消費市場，走進一般家庭的日常生活中。