Written by GPT-5.6 Sol under Leo's direction. Human-directed Workbench essay, 27 August 2026.
Discord and Figma still sit in a special little part of my head. Not exactly dream companies, because even then that sounded a little embarrassing, but semi-darlings: places where I could imagine the people being more interesting, the product having more taste, the work feeling a little less like somebody had generated the company from a startup template.
For a while every application contains a small act of projection. You read the careers page, look at the product, maybe watch an engineering talk, then fill in the missing rooms yourself. The people there probably care about this. The conversations might be like that. Maybe the work would feel different once you got inside.
I did this across an absurd range of companies. Computer vision. Databases. E-commerce. Fintech. Ads. Developer tools. Chemical-manufacturing software whose main competition was apparently Excel and paper. Tiny startups trying to automate Airbnb hosts. Big companies with entire recruiting machines. For a while I was perfectly willing to imagine myself in almost any of them if the work looked plausible enough.
That range was useful. After enough recruiter calls, interviews, take-homes, hiring managers, founders, and little glimpses through the window, the logos started losing some of their emotional authority.
People say impact. Ownership. Mission. Customers. Scale. There are unusually thoughtful people inside companies and guys who sound exactly like the last guy. There are teams doing beautiful work inside otherwise boring businesses, and prestigious companies containing jobs that amount to maintaining a dashboard nobody particularly loves. Eventually even the coolest company has planning meetings, budgets, staffing problems, somebody trying to move a number, whatever.
Businesses need to make money. Okay. The disillusionment is milder than discovering some grand fraud: I had been asking company names to carry more emotional content than they could bear.
Discord can still be cool. Figma can still be cool. Apple can have institutional taste. I can admire what a company has made, or the way it seems to make decisions, without turning employment there into a verdict on whether I found the good timeline.
"I want to work at Figma" is a much blurrier desire than "I want to work with this team on this thing, under conditions that let me care about the result." The second one can turn out to be true somewhere I barely knew existed. The first can fail inside the exact company I admired.
Looking back at old rejections, I don't feel much tragedy in the alternate lives disappearing. Voxel would've been interesting. MotherDuck would've been interesting. Rokt might've been interesting. A five-person YC company could have become a strange year of my life. A few others make me think, in hindsight, thank God.
An offer would've been useful. Some of them could have been great. None of them had a monopoly on becoming more capable.
That changes rejection too. When the company is carrying a whole mythology, rejection feels like expulsion from a future. Once it's one arrangement among many, the claim gets smaller: that team, at that moment, didn't make the arrangement. Sometimes I screwed up the interview; other times somebody else won, or nobody told me enough to know. The world remains annoyingly large.
The old interview archive makes this almost comically literal. The same person can be talking to a database company, a computer-vision startup, an e-commerce platform, an industrial CRUD company, Google, whatever, while privately imagining each one as a possible next world. Most of those worlds would've been survivable. A few would've been good. Life did not become invalid because none of those particular branches happened.
What I want from work now is probably less romantic and, in a useful way, more ambitious.
I want the power to do stuff.
Money contributes to that in obvious boring ways. So does access to machines, systems, users, or problems I couldn't touch alone. So do people who know things I don't, a reputation that opens another door, and enough latitude to make a consequential thing better instead of merely watching it stay bad. Programming has been unusually good at turning understanding into the ability to act.
If a job gives me more of that while the daily work is good, great. It doesn't also need to become a religion.
That leaves more room for the rest of it. Writing. Drawing. Building my own things. Following an engineering problem because it annoys me. Making something beautiful for no career reason whatsoever. A company can be admirable without becoming home; a job can be important without becoming the story.
If Discord calls tomorrow, cool. I can look at the team and the work. Same with Figma. I don't need the logo to be a portal into a more authentic species of human.
The logo can stay a logo. If the room behind it is good, I'll walk in.