Wednesday, May 14, 2025

Some Of My Favorite AI Tools For Engineering Students

As an engineering professor, I've seen how AI tools are transforming how we tackle our coursework, from solving complex equations and debugging code to creating visualizations and polishing lab reports. Whether you are wrestling with thermodynamics problems at midnight or designing circuits for a project, these AI assistants will help you work smarter and learn more effectively. Here's a list of some of my favorite AI enabled resources for engineering students. This list is in no way complete!

 

For Problem-Solving and Calculations:

·       Wolfram Alpha - Exceptional for advanced mathematics, physics, and engineering calculations. It can solve differential equations, perform matrix operations, and provide step-by-step solutions.

·       Symbolab - Great for calculus, linear algebra, and showing detailed problem-solving steps.

·       MATLAB Online - While not purely AI, it includes AI/ML toolboxes and is essential for many engineering courses. We all use it!

 

For Research and Learning:

·       Claude - Helpful for explaining complex engineering concepts, debugging code, and providing detailed technical explanations.

·       Gemini (Google's AI) - Excellent for research and technical explanations, with strong integration with Google services and ability to analyze images and technical diagrams.

·       ChatGPT - Good for general engineering questions and concept clarification.

·       Perplexity AI - Excellent for research as it provides citations and up-to-date information.

 

For Programming and Code:

·       GitHub Copilot - Invaluable for coding assignments in Python, C++, MATLAB, and other languages commonly used in engineering.

·       Replit AI - Integrated coding environment with AI assistance.

·       Gemini Code Assist - Google's coding assistant, particularly strong with Google Cloud and web development.

 

For Design and Visualization:

·       DALL-E 3 or Midjourney - Useful for creating diagrams, conceptual designs, or visualizations for presentations.

·       Canva AI - Helpful for creating professional presentations and posters.

 

For Writing and Documentation:

·       Grammarly - Essential for lab reports, technical writing, and documentation.

·       Quillbot - Useful for paraphrasing and improving technical writing clarity.

 

Specialized Engineering Tools:

·       Ansys AI - For simulation and analysis in mechanical/aerospace engineering.

·       PSpice - Industry-standard circuit simulation software .

·       CircuitLab - For electrical engineering circuit analysis.

 

Study and Organization:

·       Notion AI - Great for organizing notes, creating study guides, and managing projects.

·       Anki with AI plugins - For creating smart flashcards for technical terms and formulas.

 

These are just some of many excellent AI tools that I use.... some are more "AI" than others. Most colleges and universities offer free or discounted access to many of these tools. I'd recommend starting with one or two that match your immediate needs and gradually exploring others as you progress through your coursework. Always check your university's academic integrity policies regarding AI use in assignments.

Saturday, May 10, 2025

A Response - Rethinking Engineering Education for the AI Era

In my last post, Reimagining Engineering Homework with Simulators in the Age of AI, I argued that traditional electrical engineering homework focused on calculations is now easily solved by AI, requiring educators to shift to simulator-based assignments that develop higher-order skills like design, troubleshooting, and systems thinking. By using circuit simulation tools, students can engage in active experimentation and real-world problem-solving that requires distinctly human engineering judgment that AI cannot replicate.

I received the following comment on the post: 

I agree that it makes no sense to assess students' ability to make calculations that the simulators they are familiar with already make. The problem, though, is that even the tasks you suggest (e.g., create a design that meets specifications) can be already be accomplished by a variety of generative artificial intelligence platforms. Which begs the questions: what will the electrical engineers we are training actually do when they graduate, and what will they need to know in order to do it?

I’d be a liar if I said I was not asking myself the same questions. Here’s my reply:

You raise a crucial point that goes to the heart of modern engineering education. The rapid advancement of AI tools that can handle both calculations and design tasks challenges us to fundamentally reconsider what students need to learn.

I think the key lies in developing capabilities that remain distinctly human, even as AI handles more routine tasks. Future electrical engineers will likely need to excel in:

Systems thinking and integration - While AI can generate designs meeting specific parameters, engineers must understand how components interact within larger systems, identify trade-offs, and make judgment calls that balance competing constraints beyond what can be easily quantified.

Problem definition and formulation - Perhaps most critically, engineers need to determine what problems to solve in the first place. AI can optimize solutions, but it still requires human insight to identify the right questions and define meaningful specifications that serve real human needs.

Critical evaluation and verification - Engineers must be able to assess AI-generated solutions, spot errors or limitations, and validate that designs work in real-world conditions with all their messy complexities.

Innovation at the intersection - The most valuable engineers will combine domain expertise with an understanding of what AI can and cannot do, using these tools creatively to solve problems that neither humans nor AI could tackle alone.

Rather than competing with AI on tasks it can already do, engineering education must focus on these higher-level skills while using AI tools as aids in the learning process itself. 

Friday, May 9, 2025

Reimagining Engineering Homework with Simulators in the Age of AI

…. simulator-based assignments shift engineering education from passive computation to active
investigation…..

As AI tools now easily solve most traditional homework problems, engineering educators face a critical inflection point in meaningful assignment design. In my discipline, electrical engineering, the traditional homework model, focused on calculating impedance, solving differential equations, or applying Kirchhoff's laws, no longer serves as an effective assessment of student understanding. Fortunately, circuit simulation technologies offer a powerful supplement that transforms how students engage with electrical engineering concepts.

Simulators like PSpice, Multisim, and MATLAB provide virtual laboratories where students can experiment without physical constraints. Rather than simply calculating a circuit's frequency response, students can manipulate component values, sweep frequencies, and observe real-time effects through virtual oscilloscopes and spectrum analyzers. This shifts homework from passive computation to active investigation. When a student asks "what happens if I replace this capacitor?" they're engaging in authentic engineering inquiry that AI (at least not yet) cannot replicate.

The educational value extends beyond mere observation. Quality circuit simulator-based assignments require students to predict behavior, troubleshoot unexpected results, and optimize designs within real-world constraints. Students might investigate why their amplifier circuit distorts at specific frequencies, determine the optimal filter topology for a given application, or debug timing issues in a digital logic system. These higher-order engineering thinking skills remain distinctly human despite AI's computational prowess.

Virtual circuit simulators democratize access to sophisticated equipment and scenarios that might otherwise be unavailable due to cost, safety concerns, or physical limitations. Consider electrical engineering students without access to $10,000 oscilloscopes, spectrum analyzers, or signal generators, through simulators they can conduct virtual experiments with professional grade virtual instrumentation. Students can work with high voltage power electronics without risk of electrocution, or experiment with expensive radio frequency components without fear of destroying them. Those with mobility limitations gain equitable access to bench electronics through virtual labs requiring no physical soldering or manipulation of components.

The boundaries of practical learning dissolve as well. Students can simulate microwave circuits operating at 77 GHz for automotive radar, design integrated circuits with nanometer-scale transistors, or test power distribution networks for satellite systems, applications that would be physically inaccessible due to fabrication requirements or specialized equipment needs. Time constraints also vanish: simulations can compress hours of thermal analysis into seconds, or slow down switching transients in power converters to observable speeds. This temporal flexibility enables understanding of electrical phenomena that operate on timescales incompatible with traditional oscilloscope measurements.

Perhaps most importantly, simulators create a psychological safety net that encourages bold experimentation. Students can intentionally exceed component ratings, create short circuits, or test failure modes without destroying expensive components or creating safety hazards. They can iterate rapidly through dozens of design variations without the time consuming process of physically rebuilding circuits. This freedom to fail productively cultivates the innovative thinking critical for solving complex electrical engineering problems that AI cannot address.

Applications will further develop collaboration features for circuit simulation platforms, creating environments that mirror real electrical engineering workplaces. Students will share designs, conduct peer reviews, and tackle complex projects like software defined radios together. This approach teaches essential professional skills including communicating design intentions, resolving different specification approaches, and building consensus, social learning experiences that AI tools cannot replicate.

For electrical engineering educators, the transition requires rethinking assessment metrics. Rather than evaluating whether a student correctly calculated a circuit's gain, we must develop rubrics that measure design robustness, component selection rationale, and troubleshooting methodology. The focus shifts from "did they get the right transfer function?" to "did they create a design that meets specifications under varied conditions?"

 

Tuesday, May 6, 2025

The Dash Family: One Simple Way AI Detectors Analyze Punctuation

I've been hearing a lot lately about em dashes and AI....

The dash family, consisting of the em dash (—), en dash (–), and hyphen (-) has become significant in AI detection systems that determine if content was written by humans or machines.

·      The em dash (—), longer than its relatives takes its name from typography, where it occupies the width of the letter "M." Writers use it to replace commas, parentheses, or colons, adding emphasis or creating breaks in sentences.

·      The en dash (–) is shorter than the em dash but longer than a hyphen. Named for its width approximating the letter "N," it indicates ranges (2010–2020) or connections between words (Chicago–New York flight). Many writers often misuse this punctuation mark.

·      The hyphen (-), the shortest of the three, joins compound terms (cost-effective) or breaks words at line ends. Despite appearing simple, proper hyphen usage follows complex rules that writers frequently struggle with.

AI detection tools examine usage patterns of all three marks because language models often handle punctuation differently than human writers. Detection algorithms analyze the distribution and contextual placement of dashes. Human writers typically use each dash type with specific intent, while AI systems historically struggled with these patterns.

As language models have evolved, they've improved their punctuation capabilities. Modern AI can mimic human dash usage more convincingly, forcing detection tools to rely on more complex indicators beyond punctuation analysis. For writers concerned about their work being flagged, understanding these detection mechanisms helps. Using dashes according to proper style guidelines, rather than arbitrary patterns, remains the best approach.

The dash family shows how subtle language elements help distinguish between human and AI-written content, a revealing example of technology's impact on language.

Thursday, May 1, 2025

The Birth of Modern Computer Communications: The Hayes Smartmodem

Another looking back post - it was 1982, and I was still trying to figure out what I wanted to do in life. At one time I thought I wanted to go to medical school but after working in a hospital microbiology lab, I realized that was not the path for me. Maybe I could do something in the communications field….


In the early 1980s, connecting computers over phone lines was a complex and frustrating process requiring specialized knowledge and equipment. That all changed in 1981 when Dennis Hayes and Dale Heatherington introduced the Hayes Smartmodem, a revolutionary device that would fundamentally transform how computers communicate. The Hayes Smartmodem 300 was the first modem to combine communication hardware with an intelligent microprocessor control system. Unlike previous modems that required manual configuration, the Smartmodem could be controlled through a standard command set—the now-famous "AT commands" (where AT stood for "Attention"). This innovation allowed software to directly control the modem, automating the complex process of establishing connections.

The AT Command Revolution

The AT command set, sometimes called the Hayes command set, revolutionized communications because it created a standardized way for computers to control modems through simple text commands.

 

Basic Structure and Function

AT commands follow a simple structure: they begin with "AT" followed by specific command
letters and parameters. For example:

·      ATD (Dial) - Instructs the modem to dial a number

·      ATH (Hang up) - Terminates the current connection

·      ATA (Answer) - Instructs the modem to answer an incoming call

·      ATZ (Reset) - Resets the modem to its default configuration

The genius of this system was its simplicity. Before Hayes, controlling modems required specialized hardware interfaces or complicated software. The AT command set turned modem control into simple text strings that any program could generate.

Historical Impact

When Dennis Hayes introduced this command set with the Smartmodem in 1981, he effectively created the first "smart" modem that could be programmed and controlled by software. This innovation:

  • Allowed software to handle complex connection procedures automatically
  • Enabled features like auto-dialing and auto-answering
  • Created a standard that was widely adopted across the industry
  • Made modems accessible to non-technical users

Legacy and Modern Applications

Remarkably, variants of the AT command set are still used today in many communication devices. Modern cellular modems, some smartphones, and IoT devices continue to use AT commands for configuration and control. For example, sending an SMS from some embedded systems still involves AT commands like AT+CMGS.

The AT command set represents one of those rare technological innovations that was so fundamentally sound that its basic principles have outlived the hardware for which it was originally designed. From controlling 300 bps modems in the early 1980s to configuring LTE and 5G modules today, the basic concept of "Attention + Command" has proven remarkably durable.

This standardization was perhaps the Hayes Smartmodem's most enduring contribution to computing history - creating a common language that allowed computers and communication devices to work together seamlessly, helping to build the connected world we know today.

Friday, April 25, 2025

Google Gemini Progression April 18-25, 2025

**Google Gemini Advanced is free for college students through finals 2026**

This brief two-minute video demonstrates the rapid evolution of Google's Gemini AI. Within just one week, the AI advanced significantly - starting on April 18, 2025 when it could solve electromagnetic problems but couldn't generate corresponding images, to later being capable of creating detailed, dimensionally accurate visualizations of those same problems on April 25, 2025.This brief two-minute video demonstrates the rapid evolution of Google's Gemini AI. Within just one week, the AI advanced significantly - starting on April 18, 2025 when it could solve electromagnetic problems but couldn't generate corresponding images, to later being capable of creating detailed, dimensionally accurate visualizations of those same problems on April 25, 2025.

Here's my takeaways:

·      Speed of development: A one-week timeframe for implementing significant new capabilities in a complex AI system demonstrates extraordinary engineering progress.

·      Cross-modal integration: The transition from purely computational problem-solving to visual representation shows successful integration between mathematical reasoning and image generation systems.

·      Technical complexity: Electromagnetic problems often involve complex vector fields, differential equations, and spatial relationships that are challenging to visualize accurately.

·      Practical applications: This capability could revolutionize fields like engineering, physics education, and scientific visualization by making complex theoretical concepts more accessible through visual representation.

·      Dimensional accuracy: The ability to create detailed images with precise dimensions suggests the AI understands both the mathematical relationships and their physical implications.

This development represents an important step toward AI systems that can not only solve technical problems but also communicate their solutions through multiple modalities, potentially making complex STEM concepts more accessible and actionable.