Things i'm seeing in tech recruitment + other general thoughts.

1) candidate cheating in assessments/interviews is at unprecedented levels right now. you'd think it's tools like interviewcoder or firstround, but nope. it's simpler: its candidates switching tabs, asking ai questions, and pasting answers. or they've got a second device going on.

2) if you're a candidate, pls trust me and assume assessment tools have serious proctoring capabilities. they're likely tracking long pauses, clipboard pastes, and i wouldn't be surprised if they're watching eye movements too.

3) recently saw an assessment tool launch a new feature - the candidate has to keep their phone camera on in addition to the laptop camera as they attempt the test. so you can see their hands, keyboard and screen throughout the test.

4) companies are starting to realize that take home assignments can easily be chat gpt-ed.

5) proof of work: with ai making it easy to generate code or spoon feed you answers, actual portfolio projects and real world problem solving become a lot more impt

6) interviews will shift a lot more towards system design, behavioural questions, and live problem-solving on unique problems. these are things that are harder to just look up or ai generate on the fly

7) sometimes i come across engineers that are technically strong but their communication (verbal/written) is weak. and i just wish they'd work on it. they'd be unstoppable and instantly 1.5x their market value. communication is really important.

5) tech hiring is still muted. but this is only true for the median engineer. if you are good at what you do, there's always going to be high demand for you

6) companies seem to not like candidates with crypto company background too much

7) engineers who havent been using cursor/windsurf/cline need to urgently begin doing so. because..

8) every cto as of today knows code gen ai is boosting engineering productivity. its not some future promised ai land anymore. this means leaner teams and a preference for 'ai-first' engineers.

9) several recruitment ai agent startups are popping up. especially for ai interviews and screening call agents. i've had prob 5 founders reach out to me to pilot their tool over the last 9 months.

10) the calling agents are getting ridiculously realistic. and latency levels are low. esp for blue collar hiring, its a no brainer to use ai screening agents

11) llms are actually great at matching resumes to jds. however doing it manually in chatgpt is cumbersome, so most recruiters aren't using llms for this purpose. solvable problem though. we for example have baked it into our in house built ATS.

12) things like JD creation etc are a solved problem at this point because LLMs can generate good ones. Though calling it a "solved problem" is not exactly right. I mean to say that the the baseline for what is a good JD has now gone up. which means if you want to write a JD that stands out and attracts candidates, you need to put in a lot more effort. This is true for so many other things. everyone can write great, professional looking emails today. the baseline has gone up.

13) every recruitment firm will slowly become a "tech enabled service"

14) there are a bunch of recruitment agencies focussed on ai talent. so ai researchers, ai engineers, mlops. if they can build a tiny by solid curated pool of such engineers, they can command a high % commission plus will be highly in demand.

14) recruitment firms can become "full stack ai companies" in the medium run. this a term coined by the yc partners i think.

15) to become a "full stack ai company" in the context of a recruitment firm, you basically own the service layer AND the ai layer, then sell outcomes. selling outcomes is anyway what agencies have always been doing since agencies tend to charge on successful placements.

16) if some point in the next 2 years, ai handles sourcing and basic screening, the real value for recruiters will be in deep candidate engagement, building a lot of relationships, understanding nuanced 'fit' beyond keywords, and strategic advising to hiring managers. recruitment firms will look a lot leaner over time.

17) i think we'll see a few more engineers quitting to become recruiters

18) the sheer number of college grads trying to break into software engineering is going to dip in the coming years. remember how ites firms used to massively drive up engineer supply? well, i've tweeted before that i believe ites as an industry is doomed. so, that demand will go down, and over time, the supply will naturally follow as those "mass hiring" opportunities dry up.

19) the implication of that last point is we're soon not going to be flooded with engineers entering the workforce. the ones who do choose this path will likely be those with a genuine natural aptitude and a real desire to become seriously good at software. overall, that's good news for india. this will probably play out over the next 3-7 years, with a lot of that talent shifting into other industries like hardware and whatnot where they'll be better utilized