Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

We've reached new heights of institutional incompetency from what I've seen.

There is a dilbert comic that said similar criticism about spreadsheets but it applies to BI tools and AI tools. It was something like.

"The spreadsheets in this presentation are of course riddled with errors and incorrect information. It doesn't matter because unless they reinforce a decision that upper management has already made no one will ever look at them again." - Dilbert Comic Paraphrase

I've seen this in display in the real world. Various reports and dashboards in states of broken.

1. Totally broken. Are not updating at all for months and sometimes years and no one seems to have noticed. People are actively using them for processes/decisions/workflows.

2. Broken in a large but not obvious way. The data is not updating but is pivoting on a date/time so it changes every cycle it runs but just rearranges the same data.

4. Endless other things I'm sure others here have many stories too. 3. Various formulas and "math" that is completely incorrect and outputting made up fantasy numbers.



> institutional incompetency

Bad choice of words. They are not incompetent.

It's just that the data challenge has become exponentially harder as the world moved away from centralised EDW and ERP systems.

And the level of investment hasn't caught up.


The level of investment went from having a team and internal software services, to a $60/month SaaS subscription(s) in many places.

The web of SaaS many companies run on will never be as coherent as a purpose built ERP, as every service is generalized and abstracted concepts don't map perfectly to your business.

You see it often, where a business will inherit the models of the SaaS they use instead of what makes sense to their business, as an attempt to map the two entities better.


If i can look at a report that I've never seen before and tell something is very wrong with the numbers in less than a minute. What do you call the people that are familiar with the data just going along with the incorrect data?

Value extractors?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: