The most valuable part of this build may be the decision record, not the dashboard. A risk number is useful only if you can trace it back to the contract, quote, currency and assumptions that produced it. Your point about a plausible number under the wrong assumptions really stayed with me: AI can make a system look finished long before it has earned your trust. Recording each adjustment and its rationale gives you a way to test the process later, including the decisions that felt sensible at the time but did not help.
Oeste, thanks. Its a lot of tokens. The last 3 weeks, I have spent 5.1 billion tokens with Z.AI, which is my favorite go-to model provider from an intelligence/cost perspective. I use a mix of GLM 5.3 and GLM 5.3 Flash.
Testing, debugging and so on is probably about 70% of the time. This system is not trivial.
Hi Tom! As a loyal follower of mine, you should know the answer already :-) OpenAI and Anthropic have decided to ban their models in China/HK so they are not even on my radar. I use Z.AI at the moment - see my comment to Oeste above.
Thanks for sharing what you are building. I must say that it looks quite impressive. Are you planning on allowing other people to use your system once it closer to complete? I would be interested.
You are welcome! I am considering commercialising the system, but first I need to prove it works for myself. That said, I do listen to input on requested features.
The most valuable part of this build may be the decision record, not the dashboard. A risk number is useful only if you can trace it back to the contract, quote, currency and assumptions that produced it. Your point about a plausible number under the wrong assumptions really stayed with me: AI can make a system look finished long before it has earned your trust. Recording each adjustment and its rationale gives you a way to test the process later, including the decisions that felt sensible at the time but did not help.
Exactly!
Great work!
Do you have an idea of how much it cost you in tokens and dollars?
Of those 300 hours—how many were spent on testing?
Oeste, thanks. Its a lot of tokens. The last 3 weeks, I have spent 5.1 billion tokens with Z.AI, which is my favorite go-to model provider from an intelligence/cost perspective. I use a mix of GLM 5.3 and GLM 5.3 Flash.
Testing, debugging and so on is probably about 70% of the time. This system is not trivial.
Sounds like a mammoth task but I’m sure you’ve learnt a lot in the process! Keep up the good work. Did you use ChatGPT or Claude?
Hi Tom! As a loyal follower of mine, you should know the answer already :-) OpenAI and Anthropic have decided to ban their models in China/HK so they are not even on my radar. I use Z.AI at the moment - see my comment to Oeste above.
Of course! Sorry, early Sunday morning brain.. :)
Thanks for sharing what you are building. I must say that it looks quite impressive. Are you planning on allowing other people to use your system once it closer to complete? I would be interested.
You are welcome! I am considering commercialising the system, but first I need to prove it works for myself. That said, I do listen to input on requested features.