[goncaloo]

Great article!

> What good is “change failure rate” if you can’t even jump-to-def across your code?

Exactly. Organizations often focus on the externally visible factors without considering the day-to-day of a developer productivity. If only we spent more time to refactor/maintain and general tooling instead of more status updates and unnecessary processes, imagine the productivity we could achieve..

At Google we had this simple problem yet it halted our entire team's productivity: IntelliJ started being insanely slow indexing files and auto completing stuff. It was a mixture of generated code and the monorepo that made it take tens of seconds for any kind of autocompletion. It's been there for years but there wasn't much priority to fix it, so things kept going (and probably still keep going) like this...

[zarkov99]

Man, I have seen this everywhere, there is so much low hanging fruit for organizations that take developer productivity seriously. It is insane the amount of friction (leading to burnout, disengagement, and ultimately turnover) that developers deal with, every day, just because no one in the organization can see past their quarterly OKRs and fix basic things.

[alasdair_]

Stripe has the concept of “paper cuts” that any developer can file via a simple web form. These are small annoyances that no one will have an OKR for but are bothersome nonetheless.

There is a team that focuses on fixing these smaller issues, in addition to a larger dev productivity team that works on the more gnarly problems like “building takes too long”. Importantly, the team DOES have an OKR for fixing paper cuts, and gets kudos for doing so.

[jrvarela56]

It baffles me how everyone complains that they can't find talent and that devs make so much money, but at the same time don't invest in at least trying making devs 2-5x more productive.

[BlargMcLarg]

The state of the industry could be way, way better. We just decided not to do so for.. some reason. We could answer it, but every time someone does, they get dogpiled by naysayers.

On the other hand, if the industry truly evolved, it would probably up the barrier to entry to a level a lot of developers would be out of a job.

[marcosdumay]

Unless we are very close to "peak software" where the most valuable problems are already solved (very unlikely), raising the productivity of some developers will not negatively affect other ones.

[BlargMcLarg]

I'm fairly certain we'll hit the point where we can no longer reasonably throw bootcampers at jobs and expect professional output by 3 months way, way before "peak software". We already don't expect that from loads of other fields, both arts and sciences.

[OkayPhysicist]

Frankly, we've already hit that point and nobody really wants to acknowledge it.

Is a college degree absolutely necessary? Probably not (but extremely valuable, nonetheless). But bootcamps are completely worthless. At best they can operate as a screen for the already capable, but they certainly don't market themselves as such.