AI for Everyone

Kira Corbett

@TheRealKiraAI

Today's Topics

What is AI?
Getting Started in AI as a Web Dev
Exercise in AI
AI in the Real World

What is AI?

https://www.menti.com/ald6kki64xzo

Late 20th / Early 21st

Chess game without human intervention

Introduction to Imitation Games

Foundational moments of AI

Advancements in phsyical machine power

The ChatGPTs emerge

Term "AI" coined

GPT model released

Chatbot

I propose to consider the question, “Can machines think?” - Alan Turing, 1950

Machine Learning

Neural Networks

Deep Learning

Machine Learning - A spam filter that learns how to identify calls on your phone.

Neural Networks - Google photos classifying objects, faces, scenes

Deep Learning - Self driving cars

Intelligence - a(1) : the ability to learn or understand or to deal with new or trying situations : reason

also : the skilled use of reason

(2): the ability to apply knowledge to manipulate one's environment or to think abstractly as measured by objective criteria (such as tests)

- Meriam Webster

Artificial Intelligence - A branch of computer science studying and simulating intelligent behaviors in humans.

Machine Learning - A subset of AI that studies to "learn" by patterns.

Supervised Learning - Data with explicit labels.

Unsupervised Learning - Does not require explicit labeling.

Reinforcement Learning - Enables autonomous learning

Neural Networks - Mimics the human brain with artifical neurons

Deep Learning - Utilizes multiple layers of neural networks

Data Science - practice of implementing ML and data to lead to better insights.
Algorithms - step by step instructions to perform a task.

Generative AI - technology AI that creates content from large datasets it is trained on.

Hallucination - incorrect response or false information from AI system.

Natural Language Processing (NLP) - technology that understands human language e.g. speech recognition.

Prompt - input to get desired results (output.)

Sentiment Analysis - uses AI to analyze tone or format of text; also known as "opinion mining."

Training Data - information, examples, etc. given to AI to learn, identify patterns, or generate content.

Large Language Model (LLM) - engine that takes input and aims to predict a sequence.

Artificial General Intelligence (AGI) - the concept that AI achieves general human-level intelligence. Usually trained on one specific task, but not every task.

"The more tasks we want a single machine to perform, the weaker its general performance becomes."

AI will replace humans in the workplace. FALSE.

AI can think on its own. FALSE.

AI will replace humans in the workplace. FALSE.

AI can think on its own. FALSE.

Moon Highway's Instructors are AI. [ no comment ]

AI as a Web Developer

Prompt Engineering

AI as a Web Developer

Prompt Engineering

Pre-trained Models & GenAI

AI as a Web Developer

Prompt Engineering

Pre-trained Models & GenAI

Training & Tuning Your Own Models

AI as a Web Developer

Prompt Engineering

Pre-trained Models & GenAI

Training & Tuning Your Own Models

Building a Full Stack AI App

Prompt Engineering

Fine Tuning: Adapting pre-trained models

Prompt Engineering

Fine Tuning: Adapting pre-trained models

Prompt Tuning: Crafting & creating prompts for initial parameters (inputs)

Pre-Trained Models

What are you trying to accomplish?

Pre-Trained Models

What are you trying to accomplish?

TensorFlow.js
ML5.js
Brain.js

Training and Tuning Your Own Model

1. Ask a Question
2. Data
3. Choose a Model
4. Tune & Evaluate
5. Present the Results.

Unclear prompt: Explain how the world works but keep it short with not a lot of details.

This does not give much guidance on your purpose. How short is short? What is considered too many details? This is ambiguous.

Clear Prompt: In 2-3 sentences, explain how planet Earth rotates to a class of high school students.

Exericse: Prompting

EXERCISE INSTRUCTIONS

Choose 1-2 of the following types of prompting. In a word document or your notes, practice writing your own prompts with clear and concise prompt engineering.

1. text prompting - prompt ChatGPT to complete a text or question.

2. content generation - create new content given guidelines (instructions) for code, images, etc.

3. sentiment positive - analyzing something as positive

4. sentiment negative - analyzing something as negative

Recap for Prompting

Get creative (but keep it simple)
Provide context of your question e.g. an error message from code, your draft of an email. Reference materials.
Be clear, concise by reducing the amount of wording used.
Don't use sensitive data!
Fact-check; this is AI not some magic guru with all your answers with your life questions.
Adjust parameters as needed.

Great resource for your prompting in the future: https://www.codecademy.com/resources/blog/ai-prompt-engineering-tips/

AI in the Real World

AI for Developer Productivity
AI as an API

The Three C's

Clarity

Context

Constraints

Replace Search Engines

Example: How do I set up a static website hosted by GitHub and connect my personal domain to it?

Replace Search Engines

Developer Enhancements

Replace Search Engines

Developer Enhancements

Example: Create a simple UI landing page for a portfolio showcasing 3 projects.

Replace Search Engines

Developer Enhancements

Entertainment

Developer Specific AI Tools

Phind - Intelligent answer engine for developers specifically
Perplexity - AI powered answer engine
Cursor - Natural language AI tool built on top of VS Code
Github Co-Pilot - AI pair programmer
OpenAI ChatGPT - AI chat
Supermaven - AI code completion
V0 Dev - Generate UI with text prompts
VS Intellicode - AI assisted development tools

AI as Your API

Request weather data <> API returns weather data 
Display weather data <> Request local mountain ski conditions 

---> Display local mountain ski conditions




Prompt for ski conditions & weather <> AI Model returns results 

---> Display local ski conditions and weather

Thank you!

Kira Corbett

@TheRealKiraAI

Eve Porcello

@eveporcello

AI for Everyone

Today's Topics

What is AI?

AI will replace humans in the workplace. FALSE.

AI will replace humans in the workplace. FALSE.

AI can think on its own. FALSE.

AI will replace humans in the workplace. FALSE.

AI can think on its own. FALSE.

Moon Highway's Instructors are AI. [ no comment ]

AI as a Web Developer

AI as a Web Developer

Prompt Engineering

AI as a Web Developer

Prompt Engineering

Pre-trained Models & GenAI

AI as a Web Developer

Prompt Engineering

Pre-trained Models & GenAI

Training & Tuning Your Own Models

AI as a Web Developer

Prompt Engineering

Pre-trained Models & GenAI

Training & Tuning Your Own Models

Building a Full Stack AI App

Prompt Engineering

Fine Tuning: Adapting pre-trained models

Prompt Engineering

Fine Tuning: Adapting pre-trained models

Prompt Tuning: Crafting & creating prompts for initial parameters (inputs)

Pre-Trained Models

Pre-Trained Models

Pre-Trained Models

Training and Tuning Your Own Model

1. Ask a Question 2. Data 3. Choose a Model 4. Tune & Evaluate 5. Present the Results.

Exericse: Prompting

Recap for Prompting

AI in the Real World

The Three C's

Replace Search Engines

Replace Search Engines

Replace Search Engines

Developer Enhancements

Replace Search Engines

Developer Enhancements

Replace Search Engines

Developer Enhancements

Entertainment

Developer Specific AI Tools

AI as Your API

AI as Your API

Thank you!

www.moonhighway.ai

1. Ask a Question
2. Data
3. Choose a Model
4. Tune & Evaluate
5. Present the Results.