Markie Wagner, who runs Delphi Labs, says large language models are among the most important technologies since the iPhone, and predicts AI will take over large parts of executive, individual contributor and manager work within fifteen years.
Wagner and Trey Stevens developed the good quest idea in a Pirate Wires essay, defining it as work that makes the future better than it looks today and warning that harsh framings can erode people's sense that their work matters.
Wagner says a good side quest can be worthwhile but warns that calling a role a side quest can become a stop sign that excuses a weak main quest.
An AI founder on choosing work that improves the future
Markie Wagner argues that people should pick work by asking whether it leaves the future better than it looks today. She applied that test from childhood, building apps, bringing STEM programs to her schools, and then committing to AI before it became obvious. Her view is that AI will absorb routine work, so the valuable skills are judgment about what is worth doing and the grit to stay with it.
Takeaways
A good quest leaves the future better off. Wagner defined the idea with Trey Stevens in a Pirate Wires essay. A good quest makes the future better than it looks today, while a bad quest does little for the world or makes it worse. Her example is leaving a startup to fight peers for allocation in hot venture deals.
Side quests are fine until they excuse weak main quests. Singhal notes jobs often last only a few years and products may not survive, so people should judge their career as a whole. Wagner warns that the side quest label can become a stop sign. A good side quest can still be worthwhile.
She chose to build an institution that lasts. Reading about technology's greats convinced Wagner that a small handful of institutions shape the future, and building one was the most meaningful thing she could do. She became convinced AI would drive the major milestones of her lifetime and set out to be world class in it.
Large language models matter as much as the iPhone. Wagner calls large language models among the most important technologies since the iPhone. She says the hype around ChatGPT does not mean nothing real sits behind it, and she built her work and career around that belief.
Much of executive and manager work will be delegated. Wagner predicts that within fifteen years people will find it strange that workers once ran routine workflows by hand. Large parts of executive, individual contributor and manager work will go to AI, and people will become orchestrators of AI resources.
Learn the work before you manage it. Wagner advises people to learn how to do the work before managing it. At Delphi she spent her own hours on routine tasks such as copying code and doing outbound LinkedIn outreach, before handing such work to others or to AI.
Judgment, good quests and grit explain her success. Wagner credits honest judgment about which work is worth doing, a focus on good quests, and grit. She says that once she locks onto something she does not look away, and that she struggles to motivate herself on uninteresting tasks.
Wagner's path from hamster website to Delphi Labs
She got a family computer at four or five, answered questions on Yahoo Answers from about seven, and built a website about her pet hamster by tweaking templates.
Tap a name to go there · tap a line to see why they are linked
On this episode 2
Show 1
In seventh grade she built a charades app that reached tens of thousands of users. It taught her she could build what people wanted regardless of her age.
In high school she set up the first STEM programs at her public school and neighbouring schools through Project Lead the Way. The senior computer science class filled up and split evenly by gender, and alumni now work at SpaceX, Tesla, Microsoft and Google.
She applied only to Stanford and did AI research there and at Google and Waymo.
Founders began asking her for help adding AI features, including on a contract worth $250,000 in annual recurring revenue. Delphi grew during the pandemic.
She left Stanford to run Delphi Labs full time, working days of about eighteen hours.