Weeknotes no.17 – the work hamburger & what not to work on 🍔
a few months ago i upgraded myself on our token plan, since i was hitting the limits far too quickly.
we’re a small team and an even smaller developer force, but we’re managing way more systems than teams of the same size usually do.
for that scenario coding agents help soooo much. the bottleneck used to be writing code; now it’s the decisions around that.
while that gives us insane agility to work on anything on any given day and go from signal to execution in record time, it creates insane mental friction.
this month was the first where i also hit the token limits of the larger account. we’re at 96% with two work days remaining.
at the same time it’s worth mentioning we’re not tokenmaxxing or what the latest slop trends are. all those are real updates with real use.
roman ugarte from the grok bot team (i know, difficult company, but insane product) recently put it into words that really resonated with me:
[…] devs went home and used cursor for side projects; came back to work and demanded the same workflows at work
the dissonance between ai advancements, personal exploration, and the corporate inertia is growing by the day.
marie kilg recently quoted the work hamburger at an ebu event that we hosted (recording is available publicly). this image, originally coined by narayanan and kapoor, is a really good visualization of the changes by ai agents.
actually quite similar to the example of “work in between the work” from weeknotes 16.
back to work.