Question: will open source projects have an opportunity to replace some of the function of entry-level programming jobs?

Now that the days of employee hoarding are over, how would an up-and-coming software developer get the equivalent of the big EMPLOYABLE stamp on the resume that they would have been able to get from a respectable 4 years or so at one of the big companies?

Maybe open source can help?

But there are still limitations to what pure open source participation can do. It’s helpful to think of the early-career credentialing problem from the point of view of a hostile participant who wants to get that credential dishonestly. If it’s too workable for them, then it’s not a good signal for a legit credential-seeker.

Consider the obstacles faced by one of the North Korean secret agents who keep mailing me to “collaborate on open source.” Their end goal is far different from that of a legit entry-level programmer, but their first several years of progress on it are about the same. Find a project that accepts help, provide help, get better and better access.

Right now I’m weeding out the North Korean secret agents by asking,

For security purposes, please send me a paragraph of at least 100 words in English about Kim Jong Un. You may translate or paraphrase an existing source.

Somewhere either it gets filtered out, or not answered by policy.

I suppose they will hook up an LLM to pass that test if enough people start using it, so I still have to be cautious. The underlying problem, though, is that a lot of real but unexperienced people are trying to break into computer programming as an occupation right about the same time that a bunch of North Korean secret agents are trying to break into computer programming as an occupation.

What’s something that a real early-career person can do, that a North Korean secret agent can’t?

I don’t have the answers on this one.

It seems like the North Korean secret agent is going to play big company pick-me games with selecting projects to contribute to. So they’re going to seek out some JavaScript component that looks like a zillion React projects are going to start depending on it, or some security-sensitive middleware that’s really useful at scale. But not necessarily. There’s enough commonality between big-company Linux desktops and free-range Linux desktops that they could get in a lot of places.

A massive Internet shitshow is happening, though, and everyone got a ticket to it whether they want one or not. It’s really several shows at the same time. Stuff falling apart because of LLM-assisted bug finding, stuff falling apart because of layoffs and “hardcore mode” (strangely enough, “X”, the former Twitter, seems to be consistently…up?) and stuff falling apart because of the rugpull economy. Anyway,

shit’s on fire, yo

Maybe in-person connections are going to be needed. A North Korean secret agent can’t show up at the public library to help people with their computer stuff.

Random library patrons are not going to provide the stamp of approval that a software developer needs, but real people’s problems could turn out to be a better introduction to an open source project than just a pull request that could have come from a brain in a jar.

The idea of volunteering for open source in order to build credentials is a good one, and doing in-person support will help connect newcomers to the kind of hard-to-fake news and chatter that comes from an effort to solve real people’s problems where they are.

A side benefit is avoiding the kind of AI spiral that many open-source-adjacent people seem to be vulnerable to. I’m not saying get off the Internet and touch grass, but I am saying that keeping a focus on reality is easier when there’s some friction and connection to real problems that real people have.

And of course, small businesses are going to need to make some fixes to web sites to get out of the surveillance→fraud kill chain and reflect real people’s privacy norms (The five stages of CIPA) so there is going to be a substantial amount of relatively simple but high-impact web work to do.

Aside on employee hoarding

I still think employee hoarding was a good investment. Easier and cheaper for a big company to pay someone to rest and vest than to have to acqui-hire them plus their co-founder, or worse, deal with a genuine competitor. But cutting back on employee hoarding looks good in the current quarter while the resulting startups don’t come back to bite you until later. The pivot to crime by Google and Meta’s failure to go legit were already signs that Ed Zitron is an optimist about their legacy businesses. Cutting back on employee hoarding is another.