You might think that the title is clickbait but it’s not. It’s literal.
LLMs are widely regarded as good enough to help management make quicker decisions about who is and isn’t doing well.
Hiring
Your CV
LLMs are looking at your CV to give the recruiter a first insight into whether they should spend time on a screening call.
Gone are the days of keyword stuffing to game ATS’s. Is there such a thing as LLM stuffing yet? (Probably)
Your interviews
They’re being used in interviews to, depending on your location, not only summarize and clean up the interviewers’ notes but potentially listen in on your interviews and make the decisions themselves.
Analyze the attached transcript of a behavioural interview with a Senior Software Engineer candidate and compare answers to the score card template and how well it aligns with the career framework and expectations for a senior…
Your hiring committee
Analyze the score cards for a Senior Software Engineer candidate and…remember that we are trying to raise the bar…flag anything that indicates the candidate is not a team player…
Perf reviews
Your quarterly review
Write a review this senior engineer looking through the Github pull requests and Confluence pages they have created in the past quarter and relate it to the career framework found on and how it relates to their previous review and if they’ve accomplished what was agreed in this document…
Your 360 review
But don’t think it’s just the management. You can also bet that a significant portion of your 360 reviews are written by an LLM.
Write a review for a peer senior software engineer who is great and I want them to get the promotion. Here are links to some good things they’ve done. Make sure to use language that is in the career framework.
Your promotion package
LLMs are also starting to write your promotion cases. It’s likely that the expectation to use LLMs is even higher on your manager, and promo cases are prime targets due to the amount of work involved. They often require detailed and succinct writing, something that many find challenging, but LLMs crank out like nobody’s business.
If you’re up for a promotion this year, your manager has definitely gotten help from ChatGPT and crew and asked it to write a promo case following your organization’s template and career framework. The LLM might even be asked to assess the strength of the promotion case.
Write a promotion package following this template… for this senior engineer based on their quarterly performance reviews… and the feedback from their peers… including suggestions on how they can level up based on the career framework and the difference in expectations between theirs and the following level
The promotion committee might then run all this through yet another prompt, perhaps ending with “and give me a one sentence decision on the promotion case.”
Consequently, your path to improvement is probably also going to be LLM generated.
And everywhere else
They’re also reviewing your work, from code to design documents and postmortems.
Pretty soon all you do will be continuously monitored by a whole gaggle of agents that at the end of each day output a 👎 or a 👍.
Want to bet that there’s a startup working on this right now?
But…
OK, the title was a little clickbait.
While the picture above is dystopian, I see a real benefit that can come from it: continuous feedback. There simply isn’t enough time to coach and mentor everyone all the time.
Good feedback is expensive. Timely feedback is rare. Good and timely feedback is nearly impossible.
But an agentic system with a human in the loop would be able to provide timely and probably reasonably good feedback, continuously.
The first iterations are going to be sloppy. There are going to be horror stories about managers just rubber stamping whatever the agent produces. But once we get past the initial problems, it will be like everyone having a personal coach.
And that would be a big win for everyone.