What you’ll learn:
- How AI is changing the electronics field.
- A few AI tools for electronics designers.
Many of us are using artificial intelligence (AI) in a multitude of everyday applications (whether we know it or not). “Self-driving technologies,” smart search engines, social media, weather tools, and most of our cell phone apps (face recognition, photo apps, etc.) have been artificially intelligent-ized. Most likely some of the phone-based customer services "might" be AI based already. We’re surrounded by AI.
But what is AI? Artificial intelligence is the "recreation" of human intelligence processes by software, and the application might dictate the AI complexity. It can be used for big-data analysis, sensor fusion (a variety of sensor data types combined), voice interpretation/translation, natural-language interaction, and so forth.
Is Skynet AI Coming?
It was shocking when NVIDIA CEO Jensen Huang announced (Word Government Summit in Dubai, February 2024) that "programming is no longer a vital skill"... because "AI will take over coding." But it was also a harsh wake up: AI is coming fast and furious and it's not going to spare anyone... or is it?
So far, the current AI "chatters" like OpenAI's ChatGPT, Anthropic's Claude, Microsoft Copilot, Perplexity, You.com (previously known as YouChat), Gemini (formerly Google Bard), META Llama AI, etc., provide amazing answers—if you know what you’re asking! (DISCLAIMER: some answers are still subject to hallucination). This list is going to grow significantly, and there are indications that China's Baidu “Ernie Bot” is improving significantly. Other AI tools are compiled here.
Current AI technology is amazing. It knows the entire web, it can elaborate answers at different degrees of complexity, write essays/poems, sing, paint, code, create video, distinguish people from our contact list, etc.
There are other "apps" are created to serve some other segments, especially on the "customer service" side: Tidio's Lyro chatbot, Kustomer's KIQ agent, ZOHO, Cognity, etc., with many more coming up. Some of these agents are already “pre-trained,” capable of learning the company's product to the core and multilingual capable. AI is also moving into the medical and law field, so why wouldn't the engineering side be considered?
Top AI Tools in Electrical Engineering
We have already heard of how AI is improving chip design, optimizing data processing, helping on wafer/circuit/PCB/surface inspection, algorithm enhancement, power efficiency, etc. The chip electronic design automation (EDA) field is not an exception to these enhancements:
- Synopsys.ai, offers a full stack, AI-driven EDA suite for the chip design process, including analytics performance and optimization.
- Siemens's Solido is an IC design and verification package that combines the power of AI with the scalability benefits of the cloud.
- Cadence also offers a full portfolio of AI-enhanced tools, with a wide range of applications. The company is integrating AI across the entire EDA workflow, using the JedAI Platform as a common repository, helping chips to run faster, use less energy, and cost less.
- ChipAI is an EU-funded project to enable the potential of photonics nanotechnology to deliver compact, high-bandwidth, and energy-efficient CPU’s using optically interconnected spiking neuron-like sources and detectors.
- ChipGPT is also making its way to enable higher-efficiency logic/chip design via natural interaction. It starts with prompts to the large language models (LLMs) to create initial Verilog code, which will be taken by an output manager to correct and optimize these programs before collecting them into the final design space. The final expectation is that ChipGPT will search through this design space and select the optimal design under the target metrics. For a more confusing landscape, Cadence also has a project code-named ChipGPT, which focuses on the design-cleaning process and enables designers to “interrogate the design, validate the design versus specification, explore and rectify issues, prompt analysis tasks, and receive explanations in their natural language.” Please notice, there is another “ChIP-GPT” that’s a managed LLM for robust data extraction from biomedical database records.
AI Tools for Electrical Engineers
AI technology is already getting inside all EE tool sets:
- CR-8000 (from Zuken) offers AI-powered PCB place-and-route technology
- Allegro X AI, which the company claims can accelerate PCB design >10X reduction in turnaround time, with analytics tools to optimize the designs for electrical, SI, and thermal performance.
- JITX is offering a different approach to design complex circuit boards by writing simple code and streamlining existing design processes into something highly automated. This "AI tool" can be integrated with other existing tools (Altium, KiCAD, Cadence, etc.)
- Cady reviews the netlist file and bill of materials (BOM), and it will inspect all of the circuits against their datasheet to make sure any pinout is correctly set up. It also generates an AI-powered electrical rule check (ERC) and design rule check (DRC).
- Quilter is for beginners that need a PCB layout assistance. It has the noble ideal of building “the equivalent of the software compiler for circuit boards.”
- TechSpec is the AI assistant for understanding hardware component datasheets. Currently, it has massive parts data and over 100 industry-specific books, providing effortless information on component analysis.
- ACE's Circuit Mind makes it possible to create a full circuit from requirements (or building blocks) as a hardware architecture, with predefined tradeoffs (cost, power, size). Then, from the trillion component choices, it will generate the schematic and layout in a few minutes with a full "realistic" BOM. This brings the additional benefit of exploring multiple circuit configurations, trading off parts availability (based on target volume) and other constraints.
- Flux for EE design includes schematic and layout, integrating multiple platforms (including cell phones and allowing for teamwork on the same design). It also comes with a "copilot" to get all of the hard design questions answered in the same interface.
- Circuit Tree can create your embedded circuit/placement/layout on the fly. It also enables exporting of the design to Cadence OrCAD, Mentor pads, Eagle, and Altium. It has a wide range of parts from Atmel, NVIDIA, NXP, STMicroelectronics, and many others. It will generate a real-time-cost BOM from Octopart, including the PCB.
Also, for the mundane product design, nowadays we can put a full datasheet/user manual into an AI chatbot and get answers right away about any part. You can even add other components into the picture and get somewhat "reasonable" answers.
AI as a Companion in Electrical Engineering
AI is not going to be working on a lab workbench (at least not yet, unless Sophia, Atlas, Eve, Phoenix, Astribot S1, Unitree G1, or any of their robotic incarnations come to live as a EE engineer). It’s evident all of the benefits that AI can bring into the EE design cycle (from a wide component search, alternative parts selection, tolerance checking, FEA, power analysis, signal integrity, safety/EMI/EMC/ESD/EFT hardening, regulatory compliance, security hardening, cost/manufacturing efficiency, etc.) are going to help create better products in the future. This new technology is going to be at our disposal—we can’t ignore it, and we must embrace it.
It's not about "if," but rather "when" all our design tools (computerized or bench equipment) will be AI ready and interconnected. We should be able to work in collaboration, side by side, with our AI companions.
In the end, AI is another tool (a gigantic Swiss semi-autonomous tool) that we (as engineers) need to master and use efficiently. As someone has said, "It is not AI that will take your job, but rather someone with AI skills” (anonymous). So, the challenge for anyone is to step up and learn to master the AI engineering agency department. Let's focus on how we can augment ourselves, be extra-productive, more forward thinkers, and create new products/markets with this amazing tool.
Read more articles in the TechXchange: Generating AI.
