Aka: gig economy requires you to have a skill set of 5-10 people, perform the job of those 5-10 people, and get paid a fraction of 1 person's salary.
You had a two week course in University on ML? That goes right on your curriculum as Skill. Had an Excel course in school? "Office suite knowledge" Created one endpoint in a C# .Net Core API? API programming check.
So our company hired some of these full stack freelancers. While the code "works", it is not anywhere in a acceptable performance range and full of bugs. A SQL server job that runs two hours to create some 1000 simple new datarows is just not okay.
But for the business end, server job performance is not a value they can "feel" unless there is some UI connected to it.
The point I want to make that it's a potentially harmful game in the long term for any company that reads the skillsets of a developer without ever asking 1 or two questions.
The end result is systems that are completely lob sided; nice frontends in the latest tech with insecure, slow, badly created backends that crumble under load. But it doesn't matter because you can get better performance throwing money at it in a pretty simple straightforward way, while hiring other programmers, refactoring etc are very unsure endeavors(it could end up worse, deadlines aren't made etc).
I have said it here many times before here; a lot of funded startups (especially) simply prefer to burn investor money rather than do things properly. I see aws/gcp bills of 20-100k per month that can be below 10k with a few weeks (sometimes even days) of effort from someone who knows what they are doing, but nope, not happening. Automatically resolvable mistakes like having no (!!) indexes in a mysql database and so running $8k/month db bills to keep it performant (multiple times often) is becoming a little be too normal for my taste.
usually in space IoT and AI but also some qt/c++ and web sometimes
i consider myself still a freelancer as I dont have employees. but i outsource work regurarily to other freelancers. ususally coders
my high rate comes from the fact that I alone can push projects to completion. i am not a wheel in the coq but i lead the effort and own failure and success
I work in IoT and a friend had to get a PhD to work in AI so it sounds super difficult for one man to specialize in so many things at once.
This is mostly by design by companies to keep engineering side detached from the business side and the borders are strictly enforced to have every engineer as a replaceable cog.
Is there a way to build that initial network as a dev or is it just a matter of luck, like the right contacts in university or being at the right place at the right time? Since just cold calling business people or approaching them at gatherings doesn't get you anywhere since nobody trusts strangers they just met.
i think the first "right" step I took was that I asked for a broader role in the startup I was working for. I didn't just want to be a senior dev anymore. I wanted to have client contact and also contact with potential suppliers/business partners. At one of our business partners back then, I met a guy who brought last year up to 200k of business for me. That was 5 years ago and we are friends since then, working together since 2 years.
So long-term relations are key. And patience. When I first met him he was just a great drinking buddy with whom I could talk about politics :P
I am pretty positive that this notion is one part of the problem. Seen in the SQL area: Code has been created to do a simple task. Code does it's job. Data gets bigger and bigger. Upon analysis we found out that Code is very suboptimal, no thought put into indexing, partitioning or operating on set-level overall.
SQL isn't complex to learn and to be honest, it's not difficult most of the time. But there are points in knowledge where people think: "That's it, i figured it out" and the general topic is sooo much more complex underneath. This hidden complexity leads to statements like yours, given off by relative beginners (not assuming you are one).