top | item 46975116

The Bottleneck: Why Faster Coding Doesn't Speed Up Projects

13 points| goloroden | 19 days ago |heise.de

7 comments

order

28304283409234|18 days ago

This is why I have been replacing my product owner and product manager with AI. For the life of them they have no clue about their product that I am building. AI is giving me far better problem definitions based on my observations on best practises and customer input than any PO ever did.

estimator7292|18 days ago

Yeah, what I really want is an AI project manager for all my personal projects. My PM at work is an absolute rockstar at herding all of us ADHD cats, and the support and organization they provide has radically improved my work.

Please someone give me an AI that plugs into my private GitLab server and organizes my work efforts

datadrivenangel|18 days ago

In most organizations, critical path time and effort spent on coding is probably 10-20% of overall time. Overall software development effort for a software development project is probably ~50% total, and if you're lucky half of that is coding, so if coding is now 1.5x to 5x as productive, then we'd expect software projects to take up to ~40% less overall time and regular project to be ~3-8% shorter for the same output...

bryanlarsen|18 days ago

One way to use AI to speed up your projects:

Let your Support/QA/Sales engineers use AI to create features / fix bugs. They're closer to the customer than your developers are.

Even if what they deliver is complete AI slop and has to be rewritten, that pull request is a far better specification of what customers actually want than the ticket you used to get.

Too bad you fired all your QA/support engineers a few years ago.

Our_Benefactors|18 days ago

It’s actually worse than slop, because they aren’t capable of producing slop, because they don’t understand the basics of version control or even how to talk about code concepts, so they require additional training and onboarding before they can even start producing claude-slop…