...who will work with bots until they are replaced by bots trained and backstopped by people managed by bots (not uniformly to anybody's gain except the bot industry).
The link: ‘I Applied to 2,843 Roles’ With an AI-Powered Job Application Bot (404)
What it says: Covers AI Hawk's Auto Jobs Applier, an Open Source bot that goes through LinkedIn and auto-applies to jobs using OpenAI as necessary to generate (somewhat) company- and role-specific text about how aligned you are with their values and how excited you are to etc etc etc.
The takeaway: This was unavoidable, and could be seen coming a mile away (e.g. I wrote AI hiring filters don't optimize what you think they do last November and Relax: ChatGPT mostly breaks the parts of the Internet that are already broken in January 2023, and those were far from the first things I or anybody else wrote about it. It the financial logic of platforms: to be VC-viable you have to be or promise to be huge (i.e. monopolic or near-monopolic, so you can move forward to Phase II of enshittification), to achieve that scale you have to automate the slow, non-scalable human aspects of social interactions, which makes them gameable, which fundamentally changes their dynamics and informational characteristics. AI hiring filters shift hiring a bit further away from "can pass interviews" (already a poor proxy for performance, assuming you know what it is) to "can pass AI filters," which is now of course becoming "uses a competitive job-application bot." This is parallel to the previous social network dynamics, which moved reach from "says something meaningful to a lot of people" to "has the know-how and resources to game the algorithms or pay somebody to do it." Algorithmizing an interaction changes its systemic outcomes. That can be good or bad, but what you should never do is pretend it didn't.
So what do we do? The truth of the HR in 2024, from both sides of the search and from the first interview to falling to yet another painful company restructuration to optimize long-term strategic goals, is that it's not particularly effective in matching people with companies in a way that works best for either one of them. It can and arguably should be better, but it's way past time to acknowledge that this won't be done by scaling up the size or intensity with which we exploit the same sets of signals and data from which most relevant information has already been gamed away. It's back to the drawing board on this one, i.e. revolution, i.e. this isn't going to be an easy-to-sell project with a plausible, easy-to-visualize outcome.