The funny thing, it's a +5000 persons company. If this Software Developer is instrumental for the company, you'd think he'd be worth $40/employee. So there are developers, just companies not willing to pay them enough (even though we established that it's not really that high).

I was mentioning small companies in my previous post (think 4-6 devs and maybe a dozen other employees). $100k/yerar can sometimes make it or break it. Tech developers remain inaccessible for these companies making them disadvantaged in this market.

This might explain the crazy seed rounds the eco-system has been going through in the last few years. $3 million seems be the bare minimum for any startup looking to do something technical now.

That's very silly, and false.

To back this up: I'm an unemployed Python dev in the U.S. with a credible resume and GitHub profile, actively looking for something remote. $140k would be absolutely fine, no interest in FAANG or shiny companies, in fact for various reasons I've been looking primarily at smaller places. I'm not really worried yet and I've had some luck, but it's definitely not the case that anyone skilled can count on just walking into a job. I've been rejected out of hand where I met the essentials of the posting and had no doubt I could do the job.

I'd like to hear the aforementioned spiel about the "advanced Python ecosystem," which to me could mean anything between "knows what a virtual environment is" and "ML/TensorFlow expert."

"Advanced Python ecosystem experience" in this case is understanding how to create and publish a Python package.

Do you mean being able to follow the tutorial at https://packaging.python.org/en/latest/tutorials/packaging-p...? I've never needed to do that before but it seems pretty straightforward. So far I haven't seen anyone specifically request that skill in a job posting and I'd consider that a strange question to ask in a skills test for a general developer position.

If you're claiming that it costs $200k to find a competent developer who can create and publish a Python package, that's risible.

At risk of ruining a great spiel, I’d assume it boils down to “it’s amazingly foolhardy to target the candidate pool of people who have already done what you’re trying to do, when there’s a much larger pool of people who haven’t already done that thing but are readily capable of doing it”.

Tech skills and domain knowledge are transferable and learnable. So if you’re targeting hires, maybe aim for people who excel in non-transferable skills and then just teach them (or pay them to learn) the domain skills.

There’s also the side benefit that if you hire a person and ask them to repeat something they’ve done before, you still need to figure out a growth path for them. But if you hire a person and ask them to do a new-to-them thing, you have the bones of a growth path baked in.