IBM Bob vs. Ornith LLM

IBM Bob really shines in the area of program creation, manipulation, conversion, etc. It is like a hard-working assistant, taking the burden of coding from you. It will also respond properly to questions about how to interact with certain software tools, how to create connections between software-tools using REST API (it will look up the documentation and find the proper syntax, etc.). And it is very quick at statistical analysis, anomaly detection (in data), etc. Don't treat IBM Bob as your regular ChatGPT-like general-purpose AI, asking it about showbiz or recent news. It will tell you politely that this is not its field of expertise.

Incentivized

• Analyzing existing programs, explaining clearly what it does.
• Rewriting programs in a more modern language and/or a much more modular design, adding lots of explanatory comments.
• Analyzing large amounts of data, verifying assumptions are either true or false.
• Improving my ability to write good prompts, by asking IBM Bob feedback about the prompt(s) I used.

Incentivized

• The setup of skills can be more simplified
• since the IBM Bob IDE is embedded in a VSCode based framework, we have to deal with the same UI (and missing guided workflows)
• The help pages are only available via URLs. I like to see an embedded help file which is structured and includes a search feature

It works well, for the most part, but asking for authorization for multiple steps or tool installation and use sometimes becomes bothersome. You can grant blanket authorizations but I do prefer to have a tradeoff there so not all is bad, but sometimes I get the impression that we get asked the same question several times during the same workflow.

Incentivized

ChatGPT just did not have the breadth and depth of insight for my coding needs. To get an approximation of what I wanted I had to restate my prompt multiple times, and it looked almost like it did not learn from one time to the other.

Incentivized

• It lessened the dependency on me as 'the programmer' in the team. I usually had to write even the simplest of shell scripts for monitoring purposes. I just taught one of my colleagues to use IBM Bob to do it in a fraction of the time. He created it with two or three decent prompts. That saved him having to contact me and saved me time in creating it.
• For the customer, this meant having the monitoring solution online within minutes, instead of days.
• I asked IBM Bob to document a complex script that I wrote in ksh and rewrite it in Python. I can read Python but am not "fluent" in programming it. With the help of IBM Bob, I created a plug-in replacement of this tool within 1 day, which would normally have taken me weeks.
• Also, the resulting tool runs 4 times faster than the original script.
• A customer wanted to know if their suspicions were correct that, at certain times of the day, network traffic was slower. I had the systems (8 systems talking to each other) collect data every 30 seconds and fed the resulting .txt files into IBM Bob for analyzing, asking specific questions about time periods, inter- or intra datacenter traffic, etc. I first verified the details and examples it found in the data (to make sure it is correct) and asked it to generate a user-friendly report with a few tables.
• That took less than 30 minutes to execute various queries until I was satisfied. It would have taken me days to analyze the data myself (and would probably missed some of the clues).

Incentivized