Tibo's 2026 hiring rule, unlimited AI tokens for staff
Four lines, one of which turns inference spend into a standard employee perk.
Tibo posted what he calls Hiring 101, 2026 edition. Four lines. Hire smart people, pay them well, give them unlimited AI tokens, and leave them alone. His framing is that everyone comes out ahead, in his words you win, the product wins, the customer wins.
give them unlimited AI tokens
Sharing the playbook that turned years of failure into an $8m acquisition and multiple $100k+ MRR SaaS creations
The third line is the one that is new. Paying well and staying out of the way are old advice. Treating token spend as something nobody has to ask permission for is a 2026 idea, and it sits in the same budget conversation as laptops and software seats rather than in a separate approvals process.
What is behind it
Tibo attached the post to the playbook he says turned years of failure into an $8m acquisition and multiple $100k+ MRR SaaS creations. MRR is monthly recurring revenue. He gave no detail on the acquisition, no names, and no breakdown of which products sit above $100k, so take those numbers as a founder's own claim rather than anything confirmed. The post ends as a pitch for his newsletter, which he says has more than 75,000 founders and makers on it and runs with no sponsors and no ads.
The business read
Strip out the newsletter pitch and what remains is a budgeting position, not a management philosophy. An engineer who has to justify a per seat plan upgrade, or who babysits a usage dashboard, is spending attention on the wrong thing. The cost of that hesitation is probably larger than the bill it is meant to avoid, at least for a team small enough that the founder sees every invoice.
What Tibo does not say is where the ceiling is. Unlimited is easy to promise when a team is a handful of people and hard to hold when agents run long jobs overnight. There is no number in the post for what his own token spend actually is, per head or in total, and no description of how the policy is enforced or whether it ever got tested. On that, nobody knows yet.
It also lands against a backdrop of founders reporting real inference bills rather than theoretical ones, and the honest version of this advice is that the perk is cheap right now because models are cheap right now. The rule holds as long as that stays true. If it stops being true, unlimited becomes a line item somebody has to defend, and the fourth line about leaving people alone is the one that quietly goes first.
Steal the numbers, not the vibes
One email a day. Under three minutes. Real MRR, real pricing, real launches from founders building with AI.