<aside> <img src="/icons/glasses_gray.svg" alt="/icons/glasses_gray.svg" width="40px" />
That ‘Internet of Things’ Thing: In the real world, things matter more than ideas.
I didn’t truly understand the meaning of IoT before this. Twenty-six years after the term was invented (which is also my age—I was born in 1999) and sixteen years after this article was published, we now live in a time when computers, as the article puts it, “have their own means of gathering information, so they can see, hear and smell the world for themselves,” largely through the power of AI. But have they really achieved the ability “to observe, identify and understand the world—without the limitations of human-entered data”?
There is still a significant gap between computers’ and humans’ capacity to understand things. From my perspective, computers and humans have a complementary relationship rather than a replacement one. Machines can use a wide range of sensors and can gather many categories of information, with strong advantages in recording and analyzing data clearly. Humans, on the other hand, are limited to five senses—vision, hearing, smell, taste, and touch—but these senses form an extremely sophisticated system spread across the entire body. In understanding the world, we rely on a more “comprehensive” or holistic mode of perception, rather than a line-by-line analytical process. The analogy of hunters is interesting here: hunters track animals using a subtle combination of senses that is hard to verbalize or translate into logic—almost like a kind of intuition or prophecy—which is something machines struggle to learn. So although many machines are modeled after human structures, they ultimately end up fundamentally different.
I recently had a conversation with a friend about the limitations of generative AI. He mentioned that even with image-to-image or image-to-video capabilities, the interface for generative AI still relies mostly on text prompts. Unless we move beyond this, generative AI may not be enough for creating design and art. This is because design—especially “style”—is essentially a feeling or atmosphere, something text alone cannot fully describe. Designers also aren’t necessarily good at translating those sensations into language. One possible way to push generative AI’s creative capacity would be to feed it hundreds of images to convey a single style or feeling instead of relying on words. But that requires far greater computational power, which leads to hardware limitations. Once again, one of the most important things I learned in this class is that everything virtual is physical somewhere else.
</aside>
<aside>
I tested with Arduino UNO R4 board with WiFiSimpleSender example in ArduinoMqttClient library example. A few notes: