I use ai “voraciously” coding every day. Even the best models are not very good. I found that if something is hard to do it’s because it’s not documented properly, and therefore it’s not in the data anyway so the model can’t out, just guess. If it is documented properly then I don’t need its help so much.
I barely get 1 line of code that I’d actually ship unchanged. It’s really useful but still has a long way to go
I’m old too and have been doing this just as long.
What’s the most impactful task for the business right now? What’s the feedback from the sales team? What tickets are coming into customer support? What are our competitors doing? What do the marketing team need? You’re across all this? What about the other 250 developers in the business?
At some point it doesn’t scale. So maybe you need someone further up the pipeline identifying opportunities and looking at risks, talking to customers and aggregating the data so that we can be focused on the most important thing.
Maybe you’ve only worked in very small companies over those 20 years but it’s a naive take on things.
When we eventually nail agi, I think we will look at llm’s as nothing more than the interface to ai, how we interact with it, but we won’t consider it to be ai.
Well, there is quite some difference between what the agile manifesto describes and what the 5000-page Agile Handbook (TM) of some international consulting firm describes.
Dave Thomas, one of the creators of the agile manifesto, has a talk where he expresses his disappointment with consultants twisting the manifesto into a cult-like industry.
He calls this the agile industrial complex and says that anyone using agile with capital A is not referring to the manifesto but the industry and are selling you that.
Well, for me it isn't? I don't do project work, I am a maintainer at heart. If I am doing my job well, anybody who I think of as my customer will be happy if they need to interact with me as little as possible, because my well documented processes and tools are on-point :D
