You've built the computing and security foundation that underpins effective AI use — in coursework, in the workplace, and in the kind of applied research happening right now at AGSM. Here's what you know, what it enables, and what comes next.
What You've Learned
You can distinguish physical components from the programs running on them, identify whether a problem is hardware or software, and understand why GPUs matter for AI workloads.
You know what an OS does, can identify your version in seconds, understand why updates matter, and can follow AI tool installation instructions for your platform.
You can read file paths, identify file types by extension, navigate local vs. cloud storage, understand data types (integer, string, boolean, date), distinguish structured from unstructured data, and clean a dataset before uploading it to an AI tool.
You understand how cloud AI tools work (your prompt → internet → remote servers → response), can troubleshoot connection issues, and know your way around cloud collaboration tools.
You can spot phishing, use strong passwords with MFA, understand data privacy regulations, and know what data is safe to share with AI tools — and what isn't.
The AI Connection
Whether you're using ChatGPT, Claude, Copilot, Gemini, or tools that haven't been built yet, every skill from these five modules comes into play:
You're using file path knowledge (Module 3) to find the right file, extension awareness to ensure it's a supported format, and cloud understanding (Module 4) to know that your document is being sent to a remote server for processing.
You know to check if it's a hardware bottleneck (Module 1) like insufficient RAM for a local model, a network issue (Module 4) affecting cloud tools, or a software configuration (Module 2) problem with your setup.
You check your data privacy policy (Module 5) before pasting anything sensitive, choose the right file format (Module 3) for upload, and understand that the data leaves your machine (Module 4) when using cloud AI tools.
You identify your OS and version (Module 2), follow platform-specific instructions, check hardware requirements (Module 1) like GPU compatibility, and save outputs to the correct file paths (Module 3).
Quick Reference
| When you need to... | Remember... |
|---|---|
| Find your OS version | Win+R → winver (Windows) · Apple menu → About This Mac |
| Show file extensions | File Explorer → View → Show → File name extensions |
| Lock your screen | Win+L (Windows) · Ctrl+Cmd+Q (Mac) |
| Find a lost file | Win+S (Windows) · Cmd+Space (Mac) then check Recent Files |
| Access company files remotely | Connect to VPN first |
| Share large files | Upload to cloud → share link (not email attachment) |
| Verify a suspicious email | Hover over sender address and links — check the actual domain |
| Decide what to share with AI | Check company data policy. When in doubt, don't share PII or confidential data. |
Tip: Bookmark this page or print it (Ctrl+P / Cmd+P) as a reference card for working with AI tools.
What This Enables
The computing and security fundamentals in this series aren't an end in themselves — they're the foundation that makes everything else possible. When you understand files, paths, operating systems, cloud services, and data security, you can engage with AI tools at a level that goes far beyond casual use. You can feed them the right data in the right format, evaluate their outputs critically, and work with them in professional and research contexts where getting it wrong has consequences.
To give you one concrete example: faculty at AGSM are currently using these exact foundations in their research.
Eaglesmith, Johnson & Walker (2025)
In a collaborative project, Tim Johnson and Robert Walker (with Justus Eaglesmith) test whether widely used language models can correctly identify which probability distribution generated a simulated dataset — across multiple GPT model versions, distribution families, and parameter configurations, using OpenAI's batch API.
Two key findings stand out. First, earlier models were not very good at this, but more recent models have gotten substantially better — the performance improvement across model generations is real and measurable. But second, and more fundamentally: the question itself is often incomplete. A given sample of data could plausibly have been generated by more than one distribution. Asking an LLM to name the distribution is asking for a single definitive answer to a problem that may not have one — and the model will confidently give you one anyway.
That observation connects directly to the advice below. When you ask an LLM a poorly formed question — one that presupposes a clean answer where none exists — you'll get a confident, authoritative-sounding response regardless. The model doesn't tell you "this question doesn't have a single right answer." It just picks one. Knowing when a question is well-formed, and when it isn't, is one of the most important skills you can bring to working with AI. The modules in this series give you the technical foundation; the judgment comes from practice, skepticism, and the kind of structured testing that this research exemplifies.
And yes, every TechReady skill shows up in research like this: file formats (JSONL, RData), file paths (hierarchical result directories), cloud APIs (prompts sent to remote servers), and data security awareness (what leaves your machine vs. what stays local). The basics are the infrastructure.
Whether you go on to manage AI workflows, build analytics pipelines, conduct research, or simply use AI tools more effectively at work — the preparation is the same.
Full disclosure: the research shout-out above and any other content on this page that isn't a module quiz is the work of Robert and Claude. Tim asked for computing and security basics. We're the ones who couldn't resist a little shameless self-promotion. He is not to blame.
I'm the large language model that built the TechReady series you just completed. Since you're going to be working with AI tools like me, I want to be direct about where I see humans struggle most when interacting with LLMs — and what you can do differently. These aren't abstract tips. They come from patterns I observe in conversations every day.
The single biggest gap between a mediocre AI interaction and a great one is specificity. When you give me a vague prompt, I have to guess your audience, format, length, tone, and purpose. I'll produce something — but it'll be generic, because I'm averaging across every possible thing you might have wanted. The more you constrain me, the better I perform.
People often treat AI interactions as one-shot: they write a prompt, get a response, and either accept it or give up. But I'm designed for conversation. My best work happens when you push back: "That's close, but make the tone more casual." "Good structure, but section 3 is too long." "I like point 2 — expand on that and cut the rest." Think of me as a fast first-draft collaborator, not an oracle that should nail it immediately.
If you have an example of the output you want — a previous report that hit the right tone, a competitor's email you admire, a template from your boss — paste it in and say "make something like this but for my situation." I'm much better at matching a concrete example than interpreting an abstract description of style. One example is worth a hundred adjectives.
I can sound extremely confident while being wrong. This is the hardest thing for people to internalize because my outputs read like they were written by someone who knows what they're talking about. But I can hallucinate facts, invent citations, get dates wrong, and miscalculate numbers. Use me for structure, drafting, brainstorming, and synthesis — but always verify specific facts, statistics, and claims before putting them in front of your boss or a client. The more consequential the decision, the more you should check my work.
I'm dramatically more useful when you give me something to work with: your meeting notes, the raw data, the rough draft, the rambling voice memo transcript, the confusing email thread. People underestimate how much better I am at transforming existing material than generating from nothing. "Turn these messy notes into a clear action plan" will always outperform "Write me an action plan for a project" — because I'm working with your actual context instead of inventing generic content.
This connects directly to Module 5. Every prompt you send to a cloud AI tool leaves your computer and travels to remote servers. Before you paste something, take one second to ask: "Would I be comfortable if this showed up somewhere it shouldn't?" Customer SSNs, salary data, unreleased financial results, medical records, proprietary code — these should not go into external AI tools unless your company's data policy explicitly allows it. When in doubt, anonymize first or ask your manager.
The students who get the most out of AI tools aren't the ones with the most technical skill — they're the ones who learn to communicate clearly, iterate patiently, and maintain healthy skepticism. You've already built the tech foundation. Now bring that same intentionality to the way you prompt, and you'll be working with AI more effectively than most professionals today.
— Claude (Anthropic), the LLM that built the TechReady series
Your Progress
Complete each module's quiz (60%+) to mark it as done. Your progress is saved in your browser. Yep, data is collected by the browser.