[cbm-vic-20]

"Like with typical Machine-Learning and AI projects, the job of the data-engineer is the least fun part and everyone wants to be a data-scientist, that works with a magically curated data-lake. This is the same with industrial data-integration. Most focus goes onto the projects that allow you to do great things with industrial data, however almost nobody is willing to work on the part of actually getting the data."

To me, this seems like the inconvenient truth of this latest ML wave. There's lots of activity around using various ML libraies and tools, but they all assume you have input in the right format, that someone is collecting the data you need in the first place, etc.

[DonHopkins]

Just like how everybody wants to be a game designer, so they can dictate to the engineers and artists how to implement all their glorious intricate designs they dreamed up while lounging in their leather armchair, without getting their own hands dirty doing any dull repetitive shit work in Unity3D or Photoshop, then complain bitterly when their minions didn't implement their vision properly to their exacting standards.

Real jobs like that simply don't exist. Not that there aren't thousands of bootcamps and trade schools that will take your money to train you for imaginary low-effort high-prestige fantasy jobs that you'll never actually find.

[voakbasda]

Knowing how to design a system is important even if you never get the chance to use that directly. Indirectly, it is useful when you need to call out a manager or colleague on their own bad designs.