Thank you for the insights. It is extremely intellectually honest of you, a no-code expert, to share with us an opinion about the limitations of no-code. I'm sticking to n8n for now because my workflows are very simple and I haven't reached the point of diminishing returns, but this is very useful information. I'm trusting the lumberjack so much more now.
Thank you so much for this! Anyone who claims that something is an ultimate solution to anything either doesn'nt know what they're talking about or lying.
My job here is to prepare you for that cliff and tell you whether what you're trying to do is beyond that cliff.
Hey! I loved this piece :) Also, I was wondering if you have a link to any writing about this Deep Research clone by 'Jim Lee'? A quick Google, xAI search, and Deep Research run have turned up nothing 🧐
Okay, this is interesting! I have above beginner understanding of code, I experimented with no-code workflows and building several apps. Even I didn't hit the wall, I already saw the wall in front of me, and started looking for alternatives, leaning towards more code based approach.
This post came just at the right time for this. If what you say is true, you just saved me a lot of experimenting and time! 🤩
"You don't need to understand every line! Not anymore." - until you need :) I think that article could be written recursively, with all the new abstractions coming along the way, right now it's LLM, so you can replace "no-code" with "LLM", and the article will be correct as well. I was vibe coding recently with the coding languages i'm not familiar with, like Go or Rust, and "the cliff is real and it's steep" once you are starting to build something bigger than 15 components and codebase grows. Suddenly, you need to dive into basics and core again to solve some problems and LLM's can create a lot of mess without understanding what's happening at the core, especially with larger codebases.
Great article! I can concur with your assessment of coding. I have been thinking about learning no code for a long time but haven’t been able to pull myself away from prompt coding. I can’t sit down with a blank page and start writing code, but I have projects with probably tens of thousands of lines of code by now. I use the same process as described in this article. 1. Prompt the LLM with what you are trying to accomplish. 2. Have LLM explain the code it wrote in a detailed docstring 3. Refine the code to precisely do what you want. It works well. Reading this article made me very reminiscent of my own journey with LLM generated code.
I think you both will enjoy this! I read this article tonight and in the last hour was able to go from almost no coding experience to having created my first very functional gensx python script that calls google search trends for the most searched topics and questions, and then uses chatGPT to assess for app ideas that could be created as well as creates a feasibility study on them!
Never even heard of GenSX before hearing this, but the combo of using ChatGPT to write the GenSX code, the GenSX documentation, the "Warp" terminal that I use and its ability to debug the script, I have it fully working! Check out this output:
Top App Concepts Based on Google Search Volumes and GPT-enhanced Feasibility Analysis:
Topic: language learning tools
Search Volume: 2170000000
Analysis: Developing a mobile app based on 'language learning tools' is a concept with both significant potential and challenges. Here is a brief analysis covering the technical feasibility, potential market size, and competition:
### Technical Feasibility
1. **Platform Development**: With current technology platforms (iOS, Android), it's entirely feasible to build complex, interactive language learning tools. These platforms support multimedia content, which is essential for language learning (text, audio, video).
2. **Content Delivery**: Cloud services enable efficient content delivery and storage. This is crucial for managing the vast amount of resources required for language learning, such as lesson modules, videos, and interactive exercises.
3. **Speech Recognition and AI**: Advances in artificial intelligence, particularly in speech recognition
Topic: mental health apps
Search Volume: 1660000000
Analysis: Developing a mobile app focused on mental health is not only technologically feasible but also highly relevant in today’s society where mental health awareness is increasing. Below is an analysis based on technical feasibility, potential market size, and competition.
### Technical Feasibility
1. **Platform Availability**: Modern app development platforms like React Native, Flutter, and native development tools for iOS and Android make it feasible to create high-quality, cross-platform mental health apps. These technologies support the integration of advanced features such as AI, chatbots, mood tracking, and data analytics.
2. **Feature Integration**: Incorporating features such as video counseling sessions, AI-based psychological help, interactive mental health assessment tools, community support systems, and personalized mental health tips are technically achievable
Topic: best budgeting apps
Search Volume: 1320000000
Analysis: Developing a mobile app focused on the topic "best budgeting apps" involves creating a tool that helps users manage their finances, track spending, and save money efficiently. Below is a brief analysis of its technical feasibility, potential market size, and competition.
### Technical Feasibility
1. **Platform Compatibility**: Ensuring the app works seamlessly across various platforms (iOS, Android) is crucial. Cross-platform development tools like Flutter or React Native can reduce development time and cost.
2. **Feature Set**: Key features might include budget planning, financial goal setting, expense tracking, savings tips, and possibly linking bank accounts securely. Technologies like Plaid for secure financial data aggregation are technically viable but require navigating API integrations and data privacy laws
Topic: how to improve sleep
Search Volume: 967000000
Analysis: ### Technical Feasibility
Developing a mobile app focused on improving sleep is technically feasible with today's technology. The core functionalities such apps usually feature include sleep tracking, analysis based on sleep cycles, and personalized sleep improvement recommendations. These can be built using various sensors available in smartphones and wearable technology, such as accelerometers and heart rate sensors, to monitor user activity during sleep. Additionally, machine learning algorithms can be employed to analyze sleep patterns and offer tailored advice. Integration with third-party devices, like smartwatches or fitness bands, can further enhance tracking accuracy. Technologies required for such developments are well-documented and accessible, making the technical aspect not only feasible but also ripe for innovation, particularly with advancements in AI and machine learning.
Interesting paradigm, especially on how explaining workflow in lists/commands is better than diagram. Brb giving the star on Github. Thank you for sharing, sir!
Thank you for the insights. It is extremely intellectually honest of you, a no-code expert, to share with us an opinion about the limitations of no-code. I'm sticking to n8n for now because my workflows are very simple and I haven't reached the point of diminishing returns, but this is very useful information. I'm trusting the lumberjack so much more now.
Thank you so much for this! Anyone who claims that something is an ultimate solution to anything either doesn'nt know what they're talking about or lying.
My job here is to prepare you for that cliff and tell you whether what you're trying to do is beyond that cliff.
Hey! I loved this piece :) Also, I was wondering if you have a link to any writing about this Deep Research clone by 'Jim Lee'? A quick Google, xAI search, and Deep Research run have turned up nothing 🧐
My bad, I had a typo in the name, it's Jim Le, not Jim Lee.
Here's the n8n template link: https://n8n.io/workflows/2878-host-your-own-ai-deep-research-agent-with-n8n-apify-and-openai-o3/
Amazing, thank you :) actually one of my deep research runs that I thought had stalled out eventually did turn this up, haha.
Okay, this is interesting! I have above beginner understanding of code, I experimented with no-code workflows and building several apps. Even I didn't hit the wall, I already saw the wall in front of me, and started looking for alternatives, leaning towards more code based approach.
This post came just at the right time for this. If what you say is true, you just saved me a lot of experimenting and time! 🤩
"You don't need to understand every line! Not anymore." - until you need :) I think that article could be written recursively, with all the new abstractions coming along the way, right now it's LLM, so you can replace "no-code" with "LLM", and the article will be correct as well. I was vibe coding recently with the coding languages i'm not familiar with, like Go or Rust, and "the cliff is real and it's steep" once you are starting to build something bigger than 15 components and codebase grows. Suddenly, you need to dive into basics and core again to solve some problems and LLM's can create a lot of mess without understanding what's happening at the core, especially with larger codebases.
Great article! I can concur with your assessment of coding. I have been thinking about learning no code for a long time but haven’t been able to pull myself away from prompt coding. I can’t sit down with a blank page and start writing code, but I have projects with probably tens of thousands of lines of code by now. I use the same process as described in this article. 1. Prompt the LLM with what you are trying to accomplish. 2. Have LLM explain the code it wrote in a detailed docstring 3. Refine the code to precisely do what you want. It works well. Reading this article made me very reminiscent of my own journey with LLM generated code.
I'm looking forward to low-no-code-pro-code
Great piece, David! I've never really enjoyed using the no-code solutions as I've found them too limiting. Cursor, for me, has changed the game!
I think you both will enjoy this! I read this article tonight and in the last hour was able to go from almost no coding experience to having created my first very functional gensx python script that calls google search trends for the most searched topics and questions, and then uses chatGPT to assess for app ideas that could be created as well as creates a feasibility study on them!
Never even heard of GenSX before hearing this, but the combo of using ChatGPT to write the GenSX code, the GenSX documentation, the "Warp" terminal that I use and its ability to debug the script, I have it fully working! Check out this output:
Top App Concepts Based on Google Search Volumes and GPT-enhanced Feasibility Analysis:
Topic: language learning tools
Search Volume: 2170000000
Analysis: Developing a mobile app based on 'language learning tools' is a concept with both significant potential and challenges. Here is a brief analysis covering the technical feasibility, potential market size, and competition:
### Technical Feasibility
1. **Platform Development**: With current technology platforms (iOS, Android), it's entirely feasible to build complex, interactive language learning tools. These platforms support multimedia content, which is essential for language learning (text, audio, video).
2. **Content Delivery**: Cloud services enable efficient content delivery and storage. This is crucial for managing the vast amount of resources required for language learning, such as lesson modules, videos, and interactive exercises.
3. **Speech Recognition and AI**: Advances in artificial intelligence, particularly in speech recognition
Topic: mental health apps
Search Volume: 1660000000
Analysis: Developing a mobile app focused on mental health is not only technologically feasible but also highly relevant in today’s society where mental health awareness is increasing. Below is an analysis based on technical feasibility, potential market size, and competition.
### Technical Feasibility
1. **Platform Availability**: Modern app development platforms like React Native, Flutter, and native development tools for iOS and Android make it feasible to create high-quality, cross-platform mental health apps. These technologies support the integration of advanced features such as AI, chatbots, mood tracking, and data analytics.
2. **Feature Integration**: Incorporating features such as video counseling sessions, AI-based psychological help, interactive mental health assessment tools, community support systems, and personalized mental health tips are technically achievable
Topic: best budgeting apps
Search Volume: 1320000000
Analysis: Developing a mobile app focused on the topic "best budgeting apps" involves creating a tool that helps users manage their finances, track spending, and save money efficiently. Below is a brief analysis of its technical feasibility, potential market size, and competition.
### Technical Feasibility
1. **Platform Compatibility**: Ensuring the app works seamlessly across various platforms (iOS, Android) is crucial. Cross-platform development tools like Flutter or React Native can reduce development time and cost.
2. **Feature Set**: Key features might include budget planning, financial goal setting, expense tracking, savings tips, and possibly linking bank accounts securely. Technologies like Plaid for secure financial data aggregation are technically viable but require navigating API integrations and data privacy laws
Topic: how to improve sleep
Search Volume: 967000000
Analysis: ### Technical Feasibility
Developing a mobile app focused on improving sleep is technically feasible with today's technology. The core functionalities such apps usually feature include sleep tracking, analysis based on sleep cycles, and personalized sleep improvement recommendations. These can be built using various sensors available in smartphones and wearable technology, such as accelerometers and heart rate sensors, to monitor user activity during sleep. Additionally, machine learning algorithms can be employed to analyze sleep patterns and offer tailored advice. Integration with third-party devices, like smartwatches or fitness bands, can further enhance tracking accuracy. Technologies required for such developments are well-documented and accessible, making the technical aspect not only feasible but also ripe for innovation, particularly with advancements in AI and machine learning.
Interesting paradigm, especially on how explaining workflow in lists/commands is better than diagram. Brb giving the star on Github. Thank you for sharing, sir!
Also yeah not gonna lie, Conversation-Driven Development sounds much better than "prompt-driven" or "vibe coding" hahah
Interesting! I’m still getting started with the AI tools, but this is consistent with my experience so far.
Thank you for the hint with GenSX. Will try.