"If you were a software engineer two years ago, you have a very different job than you do now post-AI."
What it was about
AI has broken the old signals HR relied on for hiring and development: resumes, gut instinct, annual reviews. Organizations now need skills-based assessment to verify real candidate capability, plus continuous, human-delivered coaching (not remediation) to keep people agile as jobs change faster than ever.
By the numbers
$150/month for 750 job applications/day
Cost and volume of AI mass-application tools like Lazy Apply
rec load reportedly up 10-20% depending on level
Recruiter workload increase attributed to AI-driven application volume, per audience response
Key notes
Move away from resume reliance toward skills-first hiring, using proctored AI skills assessments (cognitive, language, engineering, simulation) to verify claimed capabilities rather than trusting polished, AI-tailored resumes.
Watch for 'skill fishing' — candidates using AI to inflate resumes, read answers off a second screen during interviews, or even present an AI-generated face on video calls.
Deploy coaching as a growth benefit for all employees (not just remediation for underperformers) since flatter, AI-transformed organizations give managers less bandwidth to develop people directly.
The contrarian takeThe speaker argues traditional behavioral assessments are unreliable for identifying strong positive performers. They're good at flagging negative traits (a short temper, say) but have little predictive value for positive potential, challenging common HR reliance on these tools.
Take this back Monday
Do this for your team
Assign every new hire and newly promoted employee an external coach for their first six months, not just struggling performers.
Say this in your next leadership meeting
Resumes are broken by AI, so we should verify skills directly and give every employee ongoing coaching instead of waiting for annual reviews.
Watch out for
Relying on resumes alone in a world where anyone can paste a job description into ChatGPT and get a perfectly tailored resume, making keyword-matching a poor proxy for actual capability.
Over-indexing on pedigree (school, prior employer) in hiring, which causes companies to miss large pools of skilled candidates who don't fit the narrow profile.
Trusting traditional behavioral assessments as strong positive predictors: they're good at flagging negative traits (a short temper, say) but much weaker at predicting positive performance.
Fun fact · Matthew Hoffman
Matthew Hoffman helped scale DigitalOcean toward its IPO, where Harvard Business Review profiled the company's performance-development approach under his watch.