12-05 x2

20260512121806-data_camp_ai_training.org

@@ -0,0 +1,161 @@
+:PROPERTIES:
+:ID:       ceeb2ad1-b091-49d8-8cd6-752f28a6fd86
+:END:
+#+title: Data Camp AI Training
+#+filetags: :technical:notes:career:ai:
+
+*TITLE: Introduction to AI for Work*
+
+* Summary
+** Learning Machines
+**AI Fundamentals**
+
+- Artificial intelligence enables computer systems to perform tasks typically associated with human intelligence, such as learning, reasoning, and decision-making.
+- AI has actually been around for many years and is already integrated into your daily life in ways you might not have realized.
+
+**Traditional Programming**
+
+- Before modern AI, engineers created intelligent systems by explicitly programming step-by-step procedures.
+- The challenge is that for many important tasks, we simply can't spell out the procedure—a human expert may excel but can't articulate the instructions for a computer.
+
+**Machine Learning**
+
+- Instead of programming step-by-step procedures, machine learning enables computers to learn from examples.
+- AI fundamentally works through pattern recognition..
+  - During Training: The system recognizes and learns patterns from examples.
+  - During Operation: The trained system receives new cases and compares them against the patterns it learned during training to make decisions.
+
+** Generative AI
+**The Generative AI Breakthrough**
+
+- While AI researchers have been making steady progress for decades, late 2022 was the tipping point when systems like ChatGPT became good enough for public release.
+- Generative AI systems generate new content rather than simply analyzing existing information.
+
+**Large Language Models**
+
+- Large Language Models (LLMs) are the most important type of generative AI for your work—the technology powering tools like ChatGPT, Claude, and Gemini.
+- They can communicate fluently in natural language. More importantly they exhibit common sense and logical reasoning capabilities.
+
+**How LLMs Work**
+
+- Large Language Models (LLMs) work like other machine learning systems. They learn patterns from data during training.
+- What makes them special is scale. They’re trained on huge amounts of text from the internet, books, articles, and more.
+- Because of this broad training, they can handle many different kinds of tasks — writing, summarizing, coding, explaining, etc.
+- When you give them a prompt or question, they use learned patterns to generate a fitting response.
+
+**Beyond Language Models**
+
+- Generative AI also includes systems that create and modify images, as well as video.
+- Image generation is a key capability for many professionals, including marketing, product design, and documentation.
+- Video generation is a new frontier in AI, but it's already being used for training materials and marketing videos.
+
+** The Opportunity
+**Understanding AI at Work**
+
+- Today's AI systems excel at doing tasks, not taking over entire jobs.
+- AI doesn't boost productivity on all tasks—only those within AI's capability boundaries where it provides significant value.
+- Even for tasks where AI helps, producing useful output requires human oversight and judgment.
+
+**The Real Opportunity**
+
+- AI is raising the bar for what any professional should be able to accomplish.
+- The real question isn't "Will AI take my job?" but "Will I be one of the people who can work effectively with AI?"
+
+**Significant Benefits**
+
+- When used effectively within its capability boundaries, AI helps you accomplish significantly more, at higher quality, while making work more engaging.
+- Most professionals aren't leveraging AI well yet. By being here, you're setting yourself up to be ahead.
+
+** How AI Can Help You
+**Execution**
+
+- AI excels at executing knowledge work tasks where you know what needs to be done and the work is at a level you'd delegate to a capable junior teammate
+- In practice: You provide clear instructions, AI carries it out, you review the output
+- This dramatically reduces time on routine tasks and frees you for higher-value work
+
+**Thought Partnership**
+
+- AI serves as an exceptional brainstorming partner when you don't know what needs to be done—like turning to a creative colleague to think through tough problems
+- Particularly effective for diagnosing unclear problems, exploring solution possibilities, and weighing difficult decisions
+- Make it a habit to bring AI to the table when facing complex challenges
+
+**Refinement**
+
+- AI helps you improve your work by providing high-quality, objective feedback—pointing out weaknesses and suggesting concrete improvements
+- In practice: Share your work, specify what feedback you need, and AI provides specific suggestions
+- Make it a habit to seek feedback from AI on your work
+
+**Continuous Learning**
+
+- AI can explain any concept clearly and immediately, adapting its approach until it clicks for you
+- AI is an effective teacher: adjusts to your level, welcomes all questions without judgment, and is always available
+- Note: AI cannot replace structured learning programs that require expert-designed progressions and hands-on practice
+
+
+** Working with AI Effectively
+**Core Collaboration Principles**
+
+- Think of working with AI as collaborating with a colleague rather than using a tool. This collaboration mindset underlies everything about working effectively with AI.
+- Communicate Effectively:
+  - Use clear, unambiguous language—no special phrases or magic words are required.
+  - Give AI sufficient detail to accomplish the task successfully.
+- Iterate: AI's first output is rarely perfect. Continue the conversation and guide AI toward what you need through multiple rounds of feedback.
+
+**Communication Framework**
+
+- The Ask: What exactly do you want the AI to do? Be specific and clear about the task or outcome.
+- The Requirements: What does the output need to satisfy? Include focus, boundaries, format, style, and other specific needs.
+- The Context: What does AI need to know about your specific situation? Why you need this, how it will be used, and relevant background details.
+- The Examples (optional): What does success look like? Show the AI what you want through concrete examples—especially valuable for visual/structural requirements ( format, layout) and qualitative/subjective requirements (style, tone, quality standards).
+
+**The Practical Test**
+
+- If you walked your request to a competent junior teammate, can they complete the task with the information you provided? If not, AI probably can't either.
+
+**Practical Tactics**
+
+- Content over polish: Focus on including the right information (ask, requirements, context) rather than perfecting the writing. AI handles spelling mistakes, broken sentences, and disorganized thoughts—what matters is having the key details, not perfect prose.
+- Start simple and build: Begin with a basic request and add more detail based on what you get back. Don't stress about getting everything right upfront.
+- Ask AI what it needs: If unsure what information to provide, ask directly: "What would you need to know to help me with this?"
+- Use different modalities: Dictate your requests, provide screenshots or photos, or mix text, voice, and images as needed.
+
+**Practice**
+
+- The best way to get better at using AI is to use AI.
+- Start with low-stakes tasks where you can experiment without pressure.
+
+** Working with AI Responsibly
+**AI's Limitations**
+
+- Knowledge fabrication: AI can confidently produce false information that sounds completely plausible
+- Recency ignorance: AI works with outdated knowledge from its training period
+- Biased outputs: AI can unfairly favor or underrepresent certain groups, viewpoints, or aesthetics
+- Sycophantic outputs: AI tends to tell you what it thinks you want to hear
+
+**Why These Happen**
+
+- These limitations result from AI's training data, training process, and pattern-matching approach
+- AI labs are actively improving these issues with each generation, but they haven't been eliminated
+- Your judgment and oversight remain irreplaceable safeguards when using AI
+
+**Review AI Outputs**
+
+- Maintain critical assessment and healthy skepticism with AI outputs
+- Verify factual claims (dates, statistics, citations, technical details, recent information)
+- Ask AI to search the web and cite sources, especially for factual or recent information
+- Watch for bias: Are perspectives missing? Would this be fair to all affected groups?
+- Watch for sycophancy: Is AI telling me what I want to hear?
+- Calibrate scrutiny to stakes (more rigorous for high-stakes decisions)
+
+**Seek Critical Perspectives**
+
+- Counter bias: Ask AI to consider different viewpoints ("What perspectives might be missing?" "How might this affect different groups differently?")
+- Counter sycophancy: Explicitly request critical feedback ("What are the weaknesses?" " What could go wrong?" "What assumptions might be wrong?")
+- For important decisions, consult people from diverse backgrounds who can offer genuine pushback
+
+**Privacy Risks**
+
+- Conversations may become training data; databases can be breached
+- Follow organizational policies and avoid sharing sensitive information
+- You can usually achieve your goal without exposing private details—use generic examples or anonymized data instead
+- Use privacy protections (enterprise versions, private modes, opt-out settings)

Career Concepts/20260101010049-concepts.org

@@ -14,7 +14,6 @@ In this node lies notes relating to theoretical concepts tying with maths and/or
 ** AI Training:
 - [[id:ceeb2ad1-b091-49d8-8cd6-752f28a6fd86][Data Camp AI Training]] 
 
-
 * Mathematical Concepts
 
 Part of being a software engineer is having a good grasp of mathematical concepts. Here are some notes on various mathematical concepts that I find useful.

all-files.json

@@ -249,7 +249,7 @@
       {
         "id": "ceeb2ad1-b091-49d8-8cd6-752f28a6fd86",
         "title": "Data Camp AI Training",
-        "url": "/20260512121806-data_camp_ai_training.html"
+        "url": "/Career Concepts/20260512121806-data_camp_ai_training.html"
       },
       {
         "id": "e448cd99-afee-4702-947f-644bb34dc1aa",