5 Reasons Hardware Is About To Get "Easy"
IoT and edge computing are about to get interesting. REALLY interesting.
For the last several decades, it has been stated repeatedly that "Hardware is Hard."
But what if it isn't?
I would argue, if current trends continue, Hardware is about to get EASY.
Here is why.
Reason #1: Generative Electronics
It used to be the case that you needed a specialized engineer with skills in designing and creating PCBs if you wanted to explore the development of custom electronics. Increasingly, we are seeing a world where Generative AI can be used to instantly design electronics. Tools like Flux.ai have emerged to fill this opportunity. With advancements in LLMs and AI tools, this outcome seems a foregone conclusion.
Suddenly, hardware may not be so "hard" anymore.
Reason #2: Generative Robotics
Recently, I wrote that the potential for Generative Electronics, if projected to it's logical conclusion will enable a new field: Generative Robotics to emerge . If current trends continue, it is not implausible that nearly any robot you want to build can be rapidly designed and simulated in near real time.
Furthermore, multiple highly credible enterprises such as Figure, Tesla and Huawei are investing heavily. We are seeing increasing breakthroughs with rapidly learning robots which are able to use Q-learning to train themselves with greatly reduced data sets.
Put all this together, and we might be designing robots far closer to real-time than we might think (with help from our robotic friends).
领英推荐
Reason #3: LLM OS
Recently, a talk has been circulating which proposes simply rebuilding computers entirely around LLMs to form an "LLM OS." Such a shift would have very rapid and profound effects on how IoT devices and edge networks are created and configured. While still in early stages, this represents a major change in how hardware gadgets might be designed and even supplant Linux altogether in favor of baremetal minimal OS which at most run an LLM tasked with doing everything else.
Reason #4: Generative 3D Printing
It is not a huge leap to imagine that acceleration in 3D and 2D design tools using emerging enhancements such as "text-to-3D" or "image-to-3D" are going to directly impact what is possible in additive manufacturing, enabling far more interesting 3D printed designs to emerge.
Reason #5: On-Device LLMs
I recently shared an article outlining 5 interesting Open Source AI projects which offer different capabilities. These include:
These two innovations together mean that highly intelligent edge devices, IoT gadgets which control and steer their own actions is now a proven fact.
It is now possible to directly "talk" to your Raspberry Pi and ask it to do things, which will permanently change how IoT devices are programmed.
This isn't science fiction, this is THIS WEEK. This is NOW.
Founder and CEO at REEKON Tools
11 个月I think these will help but certainly not make it easy in the near term. There are so many edge cases in engineering that outside of making basic, very predictable products that a company has a history of making, these will help supplement specific (and perhaps tedious tasks) that an engineer has to do but the bulk of the lifting is still on the engineer. Similar to 3D printing not being a complete way to manufacture or even prototype a product, these AI tools will be helpful in expanding capabilities but not necessarily make the development and manufacturing of new products substantially easier.
SmartConnect IoT
11 个月From microcontrollers to microprocessors, a whole lot more capabilities come into play - processing power, storage, utilization of Open Source and standardized, proven packages for communications, hardware interfaces…. Thanks Rex St. John