the issue of code reviews is a problem in standard node-red, the flows.json does not lend its self to git management. I wrote an integration between node-red and vs-code, it's in the node-red forum and the comment bellow.
It fixes this problem, and also fixes a problem of integrating with test tools and code coverage tools.
It's the old 80/20 rule, 80% can be done very quickly with easy to configure out of the box nodes, but it's the last 20% of functionality that takes all the time.
Dropping out to vs-code for this 20% improved my productivity a lot.
TBH I don't know how I would do a code review with n8n, i will need to do more investigation.
Here i was trying to show the strengths of both, unfortunately Agentic AI is graph based and node-red is flow based. Apples and Oranges.
As a hack, I have got LangGraph working inside of a node-red flow. I will write it up as a medium article in the next few days, as it might be of help to others. It's very codey.
I think production engineers will have to master the Langchain ecosystem, when we creating mission critical AI based systems. I think node-red has good nodes in place to record decisions and data as part of an AI ethics program.
Flowise have a good set of nodes around langchain, but I bet it was a massive and expensive undertaking to build.
Also Dashboard 2 is something n8n is sorely missing and is needed to monotor live systems.
You are correct about temp and seed, but for some models it still does not guarantee identical answers. Only models explicitly designed for deterministic output can do that, and I am not sure the commercially available models are built like that. I might be wrong.
The thing I was trying to say (not very well) was AI is going to make mistakes.
When we had systems based on relational databases, it was possible to recreate the conditions where the mistake was made. But in a world of LLMs and rapidly changing graph databases, this is almost impossible. So we should be recording everything so we don't make the same mistake twice.
Someone said to me yesterday, "The best doctors have only killed one person." Apparently, it's a common saying in UK healthcare. But it illustrates the point.
I currently work for a bank that is hot on AI ethics; I initially thought it was a bit of HR hype. But as I come to understand the inner workings of these systems, I am thoroughly convinced "recording how decisions are made is just as important as the decisions themselves."
thank you for your kind words, I obviously was not clear in how I wrote the article, AI might have picked it up. But I am old school and type everything my self. It's part of the feynman learning process I use.
However I have updated the article so it is much clearer, the UI that is missing from n8n is not WISIWIG, but a management dashboard, to be viewed once the system is live. I have added the following:
"To be clear here, I did not mean the UI that you use to create flows with, both systems are drag and drop flow creators. The issues I come across is visualizing what is going on inside of a system once it is created and up and running."
this is from a post on the node-red forum, and a solution i made.
Once you start to get big and complicated with node-red you will have this issue. But the rest of node-red is fantastic. It also covers a fix for git code reviews and integration with external testing frameworks.
Funny enough, I choose not to use AI to review my stories, as I find it results in an article that is not in my own voice. As i am Dyslexic, I will update this typo.
Sorry, I should have made myself clearer, I did not mean the UI that you use to create flows with, both systems are drag and drop flow creators. The issues I come across is visualizing what is going on inside of a system once it is created and up and running. I normally would like to have a dashboard showing me what is going on inside of a system. The node-red comes with Dashboard 2 ( a vue based set of components) for displaying whats going on in a live system. But I could not find a similar tool for n8n.
But then n8n has things I am missing from node-red, especially wrappers around LangGraph. I have to code those manually, in a node-red function node and it's not a simple task :(
They both have cloud and self hosting options, I use self-hosted a lot.
I find n8n is more suited to tying into cloud services. I found n8n a bit more difficult to setup when i wanted to host things on my file system.
2. Read three news lists that relate to work, HN is one of them, 15 min max, use timer
3 I keep a list of small housekeeping tasks (not real housekeeping, but small digital tasks), i start by ticking off one of those
4 At end of day i write down what i need to do next, I use that as the full focus task of the next day. I don't expect to be in full flow in first hour of task
I only read email twice a day, and most of my chat channels are on mute. I check them once an hour
this has worked well, been doing this for 30+ years