Skip to content
Ernest's Coda
Main
Continuous Learning
Other stuff
Georesearch Role Progress
LTF2004
Geocentric Datum Update
Sample Report Templates
1on1 with Paulina
Plateau STC
Share
Explore

AI Intro Course

AI Fluency
Need to fully utilize LLMs.

What is AI fluency?
Not memorizing best prompts, it’s developing practical skills in doing prompts
Fluent AI use is:
Effective
Efficient
Ethical
Safe

Usual Interactions with AI
Automation - AI completes specific instructions e.g. summarize, plan trips, etc
Augmentation - AI and human collaborates and create work together, AI not being a tool but a partner. e.g. formulating thoughts in complex problems. AI helps you do you work better
Agency - AI works on your behalf e.g. auto reply to email AI tools, establishing AI knowledge for it to create the task you need for it to do

AI is not just a tool but a partner or collaborator.

Core competencies to use AI as a partner not just a tool
The 4D Framework
Delegation - focusing on big picture, what exactly to do and how exactly AI can help you and what are the tasks you should do yourself.
Understanding the doal and the problem to be solved
Know what AI systems can and cant do well
Decide how to divide the work between yu and AI
Description - communicating clearly with AI systems, about giving more details for AI
what your final output is
how you want ai approach the task
how you want ai to behave
Discernment - evaluating AI outputs and behavior with critical eye
Is the output useful and correct?
Is the AI taking the right approach?
Is the AI behaving as desired?
Diligence - responsible AI use
Ensuring accuracy and taking responsibility
Honesty and transparency
Ethical use and critical awareness
⁠
image.png
⁠
⁠

Exercise 1: Applying the 4Ds (Research Project)
You’re using AI to help analyze a large dataset for a research paper.
Delegation - AI can help in doing the manual and repeated computations that should be done for the raw data of the research
Description - Introduce the problem that the research is trying to answer, explain what kinds of data were gathered for the analysis and what is the expected outcome of the study.
Discernment - For the processed raw data, I will do random testing of data and see if there are no anomalies in the AI-processed raw data. Looking for related literatures might help as well.
Diligence - As the researcher, I will ensure to look, record and double-check the sources that AI will use for some parts of my study. Lastly, I will also include which parts are done with the help of AI.

Generative AI Fundamentals
Fundamentals
What is Gen AI?
Traditional AI
Generative AI - LLMs claude, gemini, chatgpt
Path to Gen AI (what made it possible)
Algorithms - neural networks, transformers, emerged in 2017
Data - Articles and websites can be used for LLMs to learn, code and multimodal content
Computation - modern GPUs, TPUs
⁠
image.png
⁠
⁠
Behind the scenes of LLMs
⁠
image.png
⁠
⁠
⁠
image.png
⁠
⁠
AI Context window contains:
Your prompts
AI Responses
Any other info you've shared
What makes GEN AI powerful?
process vast info during training and learns complex patterns
adapts to new tasks through in-context learning
demonstrates emergent capabilities from scale

Capabilities and Limitations (LLM)
Capabilities
Limitations
Training data, LLMs have knowledge cut off date
Hallucinations can happen sometimes due to this limitation
AI Context window
amount of information AI can store during conversation
Non-deterministic
asking exact same questions can be answered differently and this is natural and can be regulated using Temperature settings
LLMs still have limit access to all kinds of knowledge
Looking Ahead
RAG are now being explored for AI to be smarter
⁠
image.png
⁠
⁠

Delegation
understanding the problem trying to solve
understanding the capabilities of available AI tools
breaking down complex work into smaller parts
making strategic decisions about who does what
choosing the right mode of interaction

Cornerstone of good Delegation
Problem Awareness (about your expertise)
What does success look like?
What kind of thinking and work is needed to get there?
AI fluency begins with and depends on your expertise
Platform Awareness
Understanding the unique strengths and limitations
which models perform best for the work you have in mind
which systems prioritize speed, creativity, depth, accuracy
Task Delegation
what could usefully be automated
where would augmentation create more value
Want to print your doc?
This is not the way.
Try clicking the ··· in the right corner or using a keyboard shortcut (
CtrlP
) instead.