Matt McFarlane on why the salary number is only half the story

The episode

Matt McFarlane designs compensation for startups and scale-ups through FNDN. Before that he led the people function at Oyster while it grew from 250 to 650 people in a single year across 70-plus countries. Three years into his own business, he runs more than half a million dollars in annual revenue as a solopreneur, and he credits all of it to the brand he built by writing about pay. He knows compensation in Australian tech as deeply as anyone James speaks to.

His case is that most companies have no rationale behind the number. Someone negotiated it, or a mate in the ecosystem said that's what the role is worth. From there, James and Matt get into why AI salaries actually spiked and what tokens are doing to the payroll line, what equity is still worth in Australia, the CTO who promotes people for building agents that automate their own level, and how to get paid more without resigning.

What we get into

• 01Replacement value: the person with two years of context and relationships is worth more than the new hire who just out-negotiated them, and most companies bank that upside until the trust is gone.
• 02"The jump is too big" decoded. The gap between what someone is on and what the market pays is the exact size by which you've been underpaying them. Take the emotion out and pay it.
• 03Why AI salaries actually spiked: tokens are superseding payroll as the biggest cost in SaaS, so companies pay the people who understand the infrastructure. AI roles need a comp review every quarter, not every year.
• 04What equity is still worth in Australia: table stakes people don't understand, exit horizons stretching from 8 years towards 15, and the case for outsized grants to fewer, more important people.
• 05The CTO whose team gets promoted by building agents that automate the level they work at. He'll probably never hire a junior again, and expects eight or nine principals with stacks of agents underneath them.
• 06AI usage is not a performance indicator. Leaderboards reward consumption, and the real measure is whether an agent's output is worth more than its token cost.

Lines worth keeping

Tap to copy, then drop it into a post.

Why it matters

AI has repriced one slice of the tech market while everything else stays flat. This is how you find the gap on the people you can't afford to lose, and pay it before a recruiter does.