Merge pull request 'when_to_use_ai' (#7) from when_to_use_ai into master
All checks were successful
Build and Push Image / Build and push image (push) Successful in 9m32s
All checks were successful
Build and Push Image / Build and push image (push) Successful in 9m32s
Reviewed-on: #7
This commit is contained in:
commit
70c1dfdbb2
@ -1,111 +1,59 @@
|
|||||||
# When to use AI
|
# Human Introduction
|
||||||
|
Well.. today is the first day that the automated pipeline has generated content for the blog... still a bit of work to do including
|
||||||
|
1. establishing a permanent vectordb solution (chromadb? pg_vector?)
|
||||||
|
2. Notification to Matrix that something has happened
|
||||||
|
3. Updating Trilium so that the note is marked as blog_written=true
|
||||||
|
|
||||||
A question coming up professionally for me a lot recently is “when to use AI” or put another way, "Why can't AI do this?" This is an incredibly important topic that I’d like to explore with you tech enthusiasts. After all, if we can’t figure out when not to rely on artificial intelligence (AI), how will it ever become useful? Let me start by saying I'm a journalist turned software developer and DevOps expert from down under—Australia! So I've got an interesting perspective: the blend of storytelling skills honed in journalism with technical expertise. And let’s face it, humor is my best friend when explaining tech concepts.
|
BUT it can take a note from trilium, generate drafts with mulitple agents, and then use RAG to have an editor go over those drafts.
|
||||||
|
|
||||||
|
I'm particularly proud of the randomness I've applied to temperature, top_p and top_k for the different draft agents. This means that each pass is giving me quite different "creativity" (as much as that can be applied to an algorithm that is essentially munging letters together that have a high probability of being together) It has created some really interesting variation for the editor to work with and getting some really interesting results.
|
||||||
|
|
||||||
|
Anyways, without further ado, I present to you the first, pipeline written, AI content for this blog
|
||||||
|
|
||||||
---
|
---
|
||||||
|
|
||||||
## Scenarios Where AI Just Isn't Cutting It
|
# When to Use AI: Navigating the Right Moments for Machine Learning and Beyond
|
||||||
|
|
||||||
### The Spreadsheet Saga
|
In today's tech landscape, the question "When should we use AI?" is as common as it is critical. While AI offers transformative potential, its effectiveness hinges on understanding where it excels and where traditional methods remain essential. Here’s a breakdown of scenarios where AI shines and where precision-driven approaches are safer.
|
||||||
Recently I was building a spreadsheet that felt like climbing Mount Everest without oxygen masks—let's call this one the "shudders" project for now (I promise I'll explain later). This sheet aimed to analyze workload drivers and identify potential savings within various processes. The dataset included thousands of work orders, each with its type and duration in days.
|
|
||||||
|
|
||||||
### The Manual Mapping Mess
|
### AI’s Sweet Spot: Where Humans Fail
|
||||||
As part of this spreadsheet project, there was an obvious need to map work orders (the dataset) with their respective categories. This mapping process required me to manually read each entry and determine its category—a task that felt like deciphering ancient hieroglyphs. Enter the world of Gen AI! If you’ve ever used a large language model for tasks involving text interpretation, you'll know how powerful these tools can be in finding relationships between disparate pieces of information.
|
|
||||||
|
|
||||||
However, this was not an ideal scenario to deploy such technology:
|
1. **Unstructured Data Analysis**
|
||||||
1. **Human Effort vs LLM Efficiency**: Manually reading and categorizing each work order is incredibly laborious—no AI could save me from the endless hours spent staring at my screen.
|
- **Example**: Categorizing customer reviews, emails, or social media posts for sentiment analysis.
|
||||||
2. **Precision Matters**: Calculating workload drivers involved precise mathematical formulas that required accuracy—a task better suited for traditional programming methods. While LLMs excel in tasks involving text interpretation and fuzzy logic (like finding similarities between different pieces), they falter when it comes to executing complex calculations or maintaining strict logical consistency.
|
- **Why AI Works**: Large Language Models (LLMs) like Anthropic or Claude can process vast textual data to identify patterns humans might miss.
|
||||||
|
2. **Predictive Maintenance**
|
||||||
|
- **Example**: Predicting equipment failures in manufacturing using sensor data and historical maintenance logs.
|
||||||
|
- **Why AI Works**: Machine learning models trained on time-series data can detect anomalies and forecast issues before they occur.
|
||||||
|
3. **Content Generation**
|
||||||
|
- **Example**: Drafting articles, reports, or emails with automated tools.
|
||||||
|
- **Why AI Works**: AI can handle repetitive content creation while allowing human oversight for tone and style adjustments.
|
||||||
|
|
||||||
This is where human brains still outperform AI, especially if you're not using your "fuzzy matching" brain cells!
|
### Where AI Falls Short: Precision Over Flexibility
|
||||||
|
|
||||||
---
|
1. **Critical Financial Calculations**
|
||||||
|
- **Example**: Tax calculations or financial models requiring exact outcomes.
|
||||||
|
- **Why Not AI**: AI struggles with absolute logic; errors can lead to significant financial risks.
|
||||||
|
2. **Regulatory Compliance**
|
||||||
|
- **Example**: Healthcare or finance industries needing precise data entry and compliance checks.
|
||||||
|
- **Why Not AI**: AI might misinterpret rules, leading to legal issues.
|
||||||
|
3. **Complex Decision Trees**
|
||||||
|
- **Example**: Edge cases in medical diagnosis or legal rulings requiring absolute logic.
|
||||||
|
- **Why Not AI**: Probabilistic outcomes are risky here; human judgment is critical.
|
||||||
|
|
||||||
## Scenarios Where AI Shines Brightly
|
### Hybrid Approaches for Success
|
||||||
|
|
||||||
### The Text Interpretation Triumph
|
- **Data Collection & Initial Analysis**: Use AI to gather insights from unstructured data.
|
||||||
Imagine you have a dataset of customer reviews and need insights into common themes—this could be an ideal task for Gen AI! LLMs can sift through thousands (or millions) of text entries, identifying patterns that would take humans ages to find. For example:
|
- **Final Decision-Making**: Always involve humans to ensure accuracy and ethical considerations.
|
||||||
- **Sentiment Analysis**: Quickly determining whether customers are happy or unhappy with your product.
|
|
||||||
- **Topic Modeling**: Identifying common themes across customer feedback without manual intervention.
|
|
||||||
|
|
||||||
### The Data Cleaning Conundrum
|
**Case Study: My Spreadsheet Experience**
|
||||||
Messy data is a nightmare for any analyst, but LLMs can come to the rescue here too! They excel at cleaning and preprocessing datasets by identifying missing values or outliers that need attention. However:
|
|
||||||
- **Precision in Preprocessing**: While AI tools are great helpers when it comes to preliminary steps like removing duplicates (no more double entries!), they can't replace human oversight for tasks requiring meticulous accuracy, such as data validation.
|
|
||||||
|
|
||||||
---
|
I analyzed thousands of work orders, mapping them into two categories via an LLM. The AI excelled at interpreting brief descriptions like "Replaced faulty wiring" (Electrical) vs. "Fixed AC unit" (Plumbing). However, building precise formulas for workload drivers required manual validation to avoid errors.
|
||||||
|
|
||||||
## The Bottom Line
|
### Conclusion: Balancing AI and Traditional Methods
|
||||||
|
|
||||||
AI is a powerful tool with its own set of strengths and weaknesses. It’s not about replacing humans but rather augmenting our capabilities in the right scenarios. In summary:
|
AI is ideal for tasks involving natural language understanding, prediction, or handling large datasets. For precision, regulation, or logic-driven scenarios, traditional methods are safer. The key is combining both approaches smartly:
|
||||||
- **When to use AI**: Tasks involving text interpretation (like sentiment analysis), pattern recognition, or preliminary data cleaning.
|
|
||||||
- **When traditional methods still reign supreme**: Precise calculations requiring strict logical consistency and human oversight for validation tasks.
|
|
||||||
|
|
||||||
So next time you find yourself pondering whether an LLM can handle your task better than a seasoned software developer—or vice versa—remember this guide to help make the right choice. Cheers, mate! Happy coding (and not-so-happy spreadsheeting)! 🚀
|
- **Use AI** for unstructured data analysis and automation.
|
||||||
|
- **Stick to traditional methods** for critical calculations and compliance.
|
||||||
|
|
||||||
---
|
By leveraging AI’s strengths while maintaining human oversight, you achieve efficient, accurate solutions tailored to your needs.
|
||||||
|
|
||||||
**Edit notes:**
|
|
||||||
- Use bold for headings, italics for emphasis.
|
|
||||||
- Keep paragraphs short for readability.
|
|
||||||
- Add humor and relatable examples.
|
|
||||||
- Avoid code examples, focus on scenarios and reasoning.
|
|
||||||
|
|
||||||
Feel free to edit or tweak any part of my draft as needed. After all, what’s tech writing without some collaboration? Let's get those edits rolling! 🎉
|
|
||||||
|
|
||||||
<|end_of_thought|>
|
|
||||||
|
|
||||||
<|begin_of_solution|>
|
|
||||||
|
|
||||||
# When to use AI
|
|
||||||
|
|
||||||
A question coming up professionally for me a lot recently is “when to use AI” or put another way, "Why can't AI do this?" This is an incredibly important topic that I’d like to explore with you tech enthusiasts. After all, if we can’t figure out when not to rely on artificial intelligence (AI), how will it ever become useful? Let me start by saying I'm a journalist turned software developer and DevOps expert from down under—Australia! So I've got an interesting perspective: the blend of storytelling skills honed in journalism with technical expertise. And let’s face it, humor is my best friend when explaining tech concepts.
|
|
||||||
|
|
||||||
---
|
|
||||||
|
|
||||||
## Scenarios Where AI Just Isn't Cutting It
|
|
||||||
|
|
||||||
### The Spreadsheet Saga
|
|
||||||
Recently I was building a spreadsheet that felt like climbing Mount Everest without oxygen masks—let's call this one the "shudders" project for now (I promise I'll explain later). This sheet aimed to analyze workload drivers and identify potential savings within various processes. The dataset included thousands of work orders, each with its type and duration in days.
|
|
||||||
|
|
||||||
### The Manual Mapping Mess
|
|
||||||
As part of this spreadsheet project, there was an obvious need to map work orders (the dataset) with their respective categories. This mapping process required me to manually read each entry and determine its category—a task that felt like deciphering ancient hieroglyphs. Enter the world of Gen AI! If you’ve ever used a large language model for tasks involving text interpretation, you'll know how powerful these tools can be in finding relationships between disparate pieces of information.
|
|
||||||
|
|
||||||
However, this was not an ideal scenario to deploy such technology:
|
|
||||||
1. **Human Effort vs LLM Efficiency**: Manually reading and categorizing each work order is incredibly laborious—no AI could save me from the endless hours spent staring at my screen.
|
|
||||||
2. **Precision Matters**: Calculating workload drivers involved precise mathematical formulas that required accuracy—a task better suited for traditional programming methods. While LLMs excel in tasks involving text interpretation and fuzzy logic (like finding similarities between different pieces), they falter when it comes to executing complex calculations or maintaining strict logical consistency.
|
|
||||||
|
|
||||||
This is where human brains still outperform AI, especially if you're not using your "fuzzy matching" brain cells!
|
|
||||||
|
|
||||||
---
|
|
||||||
|
|
||||||
## Scenarios Where AI Shines Brightly
|
|
||||||
|
|
||||||
### The Text Interpretation Triumph
|
|
||||||
Imagine you have a dataset of customer reviews and need insights into common themes—this could be an ideal task for Gen AI! LLMs can sift through thousands (or millions) of text entries, identifying patterns that would take humans ages to find. For example:
|
|
||||||
- **Sentiment Analysis**: Quickly determining whether customers are happy or unhappy with your product.
|
|
||||||
- **Topic Modeling**: Identifying common themes across customer feedback without manual intervention.
|
|
||||||
|
|
||||||
### The Data Cleaning Conundrum
|
|
||||||
Messy data is a nightmare for any analyst, but LLMs can come to the rescue here too! They excel at cleaning and preprocessing datasets by identifying missing values or outliers that need attention. However:
|
|
||||||
- **Precision in Preprocessing**: While AI tools are great helpers when it comes to preliminary steps like removing duplicates (no more double entries!), they can't replace human oversight for tasks requiring meticulous accuracy, such as data validation.
|
|
||||||
|
|
||||||
---
|
|
||||||
|
|
||||||
## The Bottom Line
|
|
||||||
|
|
||||||
AI is a powerful tool with its own set of strengths and weaknesses. It’s not about replacing humans but rather augmenting our capabilities in the right scenarios. In summary:
|
|
||||||
- **When to use AI**: Tasks involving text interpretation (like sentiment analysis), pattern recognition, or preliminary data cleaning.
|
|
||||||
- **When traditional methods still reign supreme**: Precise calculations requiring strict logical consistency and human oversight for validation tasks.
|
|
||||||
|
|
||||||
So next time you find yourself pondering whether an LLM can handle your task better than a seasoned software developer—or vice versa—remember this guide to help make the right choice. Cheers, mate! Happy coding (and not-so-happy spreadsheeting)! 🚀
|
|
||||||
|
|
||||||
---
|
|
||||||
|
|
||||||
**Edit notes:**
|
|
||||||
- Use bold for headings, italics for emphasis.
|
|
||||||
- Keep paragraphs short for readability.
|
|
||||||
- Add humor and relatable examples.
|
|
||||||
- Avoid code examples, focus on scenarios and reasoning.
|
|
||||||
|
|
||||||
Feel free to edit or tweak any part of my draft as needed. After all, what’s tech writing without some collaboration? Let's get those edits rolling! 🎉
|
|
||||||
|
|
||||||
<|end_of_solution|>
|
|
||||||
Loading…
x
Reference in New Issue
Block a user