HR in Review

There are a handful of non-obvious observations - the 'easy apply' critique, the bifurcation of TA into process-lean vs. consultative functions, and the social-contract breakdown argument - but the episode spends significant time on surface-level observations and conversational meandering that dilutes the useful content across 28 minutes.

“I think the most damaging thing that happened to TA Was, was easy apply”

“I think you're going to see a bifurcation where you end up with a process orientated TA function where you're going to have a very, very lean team”

A few genuinely reframed takes - questioning what 'cheating' even means when employers demand AI fluency, and the argument that TA 'forgot the consultant bit' - but the broader framing of AI disrupting TA, the volume-of-applications problem, and the employer-employee trust breakdown are widely circulated themes in the HR space.

“we always say talent acquisition is broken. Um, I don't think it's broken. I think we've forgotten what we are”

“if you're going to expect the uh, candidates to use this tech within their job for your improved efficiencies, etc... why wouldn't they want to use this tech?”

Toby Kolshaw is a credible practitioner who has run talent intelligence functions at genuinely large-scale organisations (Amazon, Philips) and has written a practitioner book on the topic; however, he is now predominantly in community-building and advisory mode, which edges him toward thought-leader territory with softer accountability to live results.

“Toby is the former head of Talent Intelligence at Amazon and Philips, founder of the Talent Intelligence Collective, and author of Talent Intelligence”

“my previous employers were 60, 70, 80,000, up to 1.57 million employees. If, if they put the wrong data center in place or they put the wrong software engineering team in place and they put a thousand headcount in the wrong place, it'll be painful, but they'll survive”

There are a couple of vivid anecdotal data points - 6,000 applications in three minutes, the Talent Intelligence Collective prompt competition - but no named companies, no measured outcomes, no cited research, and most numbers are illustrative estimates rather than verified figures; the 1,200-application example came from a newspaper article the host read, not the guest's direct experience.

“I was talking to one TA leader who advertised the job and uh, it went out, I think it was 2 in the morning, 2:03, um, and within the first three minutes they had 6,000 applications”

“I just ran a competition for the Talent Intelligence Collective where it was how well can somebody write a prompt to do a location strategy”

The host asks almost entirely open, orientation-level questions, volunteers newspaper articles and personal anecdotes rather than probing for depth, and actively redirects away from difficult threads ('let's talk about what is going well'); there is no meaningful pushback, no follow-up that forces specificity, and no productive disagreement across the entire episode.

“It's like Glastonbury Tickets, isn't it?”

“Gosh, when you say it like that, you know, who'd want to work?”