🎯 Key Takeaways for quick navigation: 00:00 🤖 *Introduction to the limitations of chatbots in coding with R* - Chatbots are not reliable for coding in R due to the complexity and nuances of the language. - Introduction to the challenge of testing Bing AI and Google Bard in code review, debugging, and map creation in R. 01:25 📝 *Code Review: Bing AI's performance* - Bing AI provides a detailed code review by analyzing specific lines of code. - Suggests improvements like adding comments and checking the existence of input variables. - Offers additional code to check the existence of defined variables. 03:03 📝 *Code Review: Google Bard's performance* - Google Bard provides a shorter code review with fewer details compared to Bing AI. - Offers suggestions, including adding comments and using the dplyr pipe operator. - Suggests using a non-existent function, which is a flaw in its response. 05:10 🔍 *Code Debugging: Bing AI's performance* - Bing AI successfully identifies the error in the code and provides a correct solution. - Offers a code snippet with corrections. - An amusing comment change is noted. 06:37 🔍 *Code Debugging: Google Bard's performance* - Google Bard identifies the error in the code and suggests the correct argument. - Provides a code snippet for correction. - The comment change is humorous but not relevant to the error. 09:44 🗺️ *Map Creation Attempt 1: Bing AI's performance* - Bing AI suggests creating a 3D population density map with R packages Rayshader and Contour. - The provided code lacks consideration for package installation and uses a global dataset URL. - Execution of the code faces issues due to missing data. 11:43 🗺️ *Map Creation Attempt 1: Google Bard's performance* - Google Bard suggests a 2D population density map with ggplot2 but doesn't provide data or instructions on where to find it. - The code lacks essential details, making it impossible to execute. 10:37 🗺️ *Map Creation Attempt 2: Bing AI's performance* - Bing AI suggests creating a population density map using real data from the WorldPop dataset. - Provides concise code but overlooks package installation concerns. - Dataset URL provided doesn't match the intended country level data. 12:37 🗺️ *Map Creation Attempt 2: Google Bard's performance* - Google Bard mentions the WorldPop dataset without specifying its source or providing instructions. - Fails to offer any code or data, rendering it impossible to create the map. 14:51 📈 *Overall Assessment* - Chatbots like Bing AI and Google Bard excel in code review and debugging tasks. - However, they are not suitable for generating maps in R due to limitations in providing working code and data. - Users should exercise caution and verify code generated by these chatbots for map creation. Made with HARPA AI
These AIs have a habit of inventing functions and then apologise when you tell them it's wrong.😂 From experience I believe ChatGPT is better because they provide the corrected code chunk when you send them a wrong line of code
I’m not sure why you think that, or what you are trying to imply by saying that. Maybe you are suggesting that Bing AI is biased towards Microsoft products, or that Bing AI is not compatible with other products that do not have auto decompress feature. However, these are just speculations, and one cannot confirm or deny them
I do not recall since i did not time stamp . So ignore my answer . As to why it used global map . Could be an advertising or whatever that site push . I will try to fiind the context of decompression . I use ms on xbox serie x @@milos-makes-maps
the decompress comment was about 8:30 ish portion . Unless bing ai got tripped on someting or tthe compression is too new or ttoo old bing ai can decompress things to read them . Often the issue i experienced on peasant subject is i use slang without me noticing it
See this issue you talk about is not present , why? You are asking ai to read in your mind . Message to bing ai . Always start with 1 question at least .after code reception . These code trick itgs old habit of academia teacher . Usually no context is supplied nno mention of is bing ai permitted to do the attempt outside the boundary of academia consensus etc . Here ? Just take this code and guess . So first part ? Bing ai likely did not know what you wanted . Since usually coding goes in precise cathegory
You can ask Bing AI multiple questions (and I did so with Cursor AI on GPT 4 in one of my previous tutorials) but the bot cannot guarantee that the script will work or meet your expectations, as the bot is not an expert in R coding
🎯 Key Takeaways for quick navigation:
00:00 🤖 *Introduction to the limitations of chatbots in coding with R*
- Chatbots are not reliable for coding in R due to the complexity and nuances of the language.
- Introduction to the challenge of testing Bing AI and Google Bard in code review, debugging, and map creation in R.
01:25 📝 *Code Review: Bing AI's performance*
- Bing AI provides a detailed code review by analyzing specific lines of code.
- Suggests improvements like adding comments and checking the existence of input variables.
- Offers additional code to check the existence of defined variables.
03:03 📝 *Code Review: Google Bard's performance*
- Google Bard provides a shorter code review with fewer details compared to Bing AI.
- Offers suggestions, including adding comments and using the dplyr pipe operator.
- Suggests using a non-existent function, which is a flaw in its response.
05:10 🔍 *Code Debugging: Bing AI's performance*
- Bing AI successfully identifies the error in the code and provides a correct solution.
- Offers a code snippet with corrections.
- An amusing comment change is noted.
06:37 🔍 *Code Debugging: Google Bard's performance*
- Google Bard identifies the error in the code and suggests the correct argument.
- Provides a code snippet for correction.
- The comment change is humorous but not relevant to the error.
09:44 🗺️ *Map Creation Attempt 1: Bing AI's performance*
- Bing AI suggests creating a 3D population density map with R packages Rayshader and Contour.
- The provided code lacks consideration for package installation and uses a global dataset URL.
- Execution of the code faces issues due to missing data.
11:43 🗺️ *Map Creation Attempt 1: Google Bard's performance*
- Google Bard suggests a 2D population density map with ggplot2 but doesn't provide data or instructions on where to find it.
- The code lacks essential details, making it impossible to execute.
10:37 🗺️ *Map Creation Attempt 2: Bing AI's performance*
- Bing AI suggests creating a population density map using real data from the WorldPop dataset.
- Provides concise code but overlooks package installation concerns.
- Dataset URL provided doesn't match the intended country level data.
12:37 🗺️ *Map Creation Attempt 2: Google Bard's performance*
- Google Bard mentions the WorldPop dataset without specifying its source or providing instructions.
- Fails to offer any code or data, rendering it impossible to create the map.
14:51 📈 *Overall Assessment*
- Chatbots like Bing AI and Google Bard excel in code review and debugging tasks.
- However, they are not suitable for generating maps in R due to limitations in providing working code and data.
- Users should exercise caution and verify code generated by these chatbots for map creation.
Made with HARPA AI
Awesome test Milos. 😊 AI can be a bit of a hit and miss, but they are improving them each day, so who knows what the future holds. 😊
Thank you, Aleks! And that's a great point 😃
These AIs have a habit of inventing functions and then apologise when you tell them it's wrong.😂
From experience I believe ChatGPT is better because they provide the corrected code chunk when you send them a wrong line of code
Hahaha, yes, that's my experience, too!
Excellent contrast of these AI guys. Although they are usefull sometimes, in many times they are misleading and clueless.
Thank you, your words some it up effectively!
Bing ai likely use auto decompress and assume your on ms product wich have auto decompress
I’m not sure why you think that, or what you are trying to imply by saying that. Maybe you are suggesting that Bing AI is biased towards Microsoft products, or that Bing AI is not compatible with other products that do not have auto decompress feature. However, these are just speculations, and one cannot confirm or deny them
I do not recall since i did not time stamp . So ignore my answer . As to why it used global map . Could be an advertising or whatever that site push . I will try to fiind the context of decompression . I use ms on xbox serie x @@milos-makes-maps
the decompress comment was about 8:30 ish portion . Unless bing ai got tripped on someting or tthe compression is too new or ttoo old bing ai can decompress things to read them . Often the issue i experienced on peasant subject is i use slang without me noticing it
@@Qarol-t9k thanks for clarifying; now I understand what you meant by decompression
See this issue you talk about is not present , why? You are asking ai to read in your mind . Message to bing ai . Always start with 1 question at least .after code reception . These code trick itgs old habit of academia teacher . Usually no context is supplied nno mention of is bing ai permitted to do the attempt outside the boundary of academia consensus etc . Here ? Just take this code and guess . So first part ? Bing ai likely did not know what you wanted . Since usually coding goes in precise cathegory
You can ask Bing AI multiple questions (and I did so with Cursor AI on GPT 4 in one of my previous tutorials) but the bot cannot guarantee that the script will work or meet your expectations, as the bot is not an expert in R coding
Un agrado escuchar tus explicaciones Felicitaciones Milton.
Gracias 🙂