The surreal-ness of such an experience cannot be overstated. I made up everything about Sapiens First. The name, the colors, the logo, the website design, the theory of change, my title of “Executive Director”. It is not too different from the stories or drawings that I would make as a kid.

I practiced the talk in my room, walking around the dirty laundry on the ground. I rehearsed my lines, speaking them out loud, believing yet not knowing if others would listen.

What gets you up the day after a ghosted email or a botched funder pitch? Self-belief is the one thing a founder must have in her heart. In my time bringing Sapiens First to life, I have pitched my friends, funders, allies, even door-to-door strangers. But by far, the most important sale you do as a founder is to yourself.

Yet I still feel lost in the face of the AI crisis. Will the movement meet the moment? I don’t know, and I feel dread clutched in my heart – that the oligarchs might win in their endless quest for power – and that we might all suffer for it.

The words came tumbling out. I emphasized the extremity of the issue. That a radical situation calls for a radical solution, that this is the fight for our lives, to protect democracy and the future of humanity.

I felt the ideas resonate through the room, like the ring of an organ in a church. What is the Constitution but an idea? What is democracy or science or Christianity but an idea?

Ideas are brought to life by belief, and there is nothing like an ardent believer. People, many older than I; they asked me what they should do with their lives; they asked me for their vote of confidence in pursuing their own journey. I did my best. I felt the way that Olivia Rodrigo might at a concert, or the way John Green does at a book signing.

In the evening, I came back to my bedroom, dirty laundry still on the floor, exhausted, but hopeful, feeling a fiction coming to life.

I feel great urgency regarding the creation of superintelligence, which I think will certainly happen in the next five years. The question is, what do you do about that?

It turns out that people do not act very well in urgent situations. Asking someone to think about something very quickly typically means that they’ll default to their base instincts, biases, or otherwise avoid taking full stock of a situation.

There’s also the fact that rest matters. Let’s say you’re trying to drive across the US as fast as possible. If you floor it at 130mph the whole way, you’ll possibly get into a crash, get stopped by the cops, or break your vehicle – all of which would slow you down. It makes more sense to go at 70mph and just make sure you are doing that consistently. It also makes sense to stop at gas stations, because you need gas.

Bringing it back to working on urgent things, rest helps you sustain in the long-run. eg. sleep, exercise, doing things other than work mean you have a fresher mind for the journey ahead. It also means you have time to reflect – am I really doing the most important thing?

I generally find this counter-intuitive. If something is extremely urgent, shouldn’t we all be doing something until it’s resolved? That may be true for things on the timeline of 24 hours, but actually, few things are. Even if something has a deadline of two weeks, it’s probably good to take breaks and sleep. All the more so for things that have a shortish timeline (a few months to a few years).

“The tortoise with prudent wisdom hastens slowly.” – Jean de La Fontaine

The best advice I got early on was to have a ridiculous, overwhelming amount of confidence. Also, it’s essential to have an idea that you completely and utterly believe in. Founding a non-profit is not for the faint of heart.

If you have those things, then I really encourage you to do it. I’ve learned more here than in any other phase of my life. At the very least, you can spend 6 months prototyping and seeing if there’s appetite for your idea, then go back to a regular job.

Here’s some mostly practical advice about how to get started:

• Incorporate. Non-profits take several months to incorporate! As soon as you have a rough idea of what you want to do, get two friends to be on the board with you and submit for the 501(c)(3). Heck, even if you’re not sure but think you might want to found something in the future, it is worth it to incorporate. You save on fiscal sponsors, which take 5-10% of your revenue.
• Idea. A great idea takes time to think about. I’d say the Ikigai framework applies: you want to do something that the world needs, that you enjoy, that you’re good at, and that is financially sustainable. It’s worth thinking about the business model early, since that could inform your strategy, even if you get by with savings, grants, or gifts from friends and family to start. For a non-profit, a business model just means some way of generating consistent revenue, eg. through a membership program, fundraisers, or by selling a product.
• Name. The name is exceedingly hard to choose. Choose a placeholder name at first so that you can incorporate and describe the project to people. Before you start looking at domain names, think about the purpose of the name: appeal to politicians, the general public, lawyers, Republicans, Democrats? That will determine what criteria you use. No name will be perfect in everyone’s eyes. Because Sapiens First is a grassroots organization, the core requirements were: it’s easy enough to spell; white space in search results; short enough that it can be used for chapters (eg. Sapiens First Berkeley); roughly related to the idea of AI & democracy.
• Funding. Getting funding is a lot like getting a job. Except you have to continue doing it indefinitely to keep your organization afloat. Grant-making and donations are not a fair & efficient marketplace (compared to, say, venture capital, where the funders have incentive to think rationally and avoid group-think). Alas. So that means that like many other things, funding is generally about relationships. Donors have to know, like, and trust you. You can get good at fundraising with persistence. Reaching out to people to catch up, sending cold emails to people who might be relevant, going to events in your area. Tailor the framing of your project depending on what you know about the person. Listen much more than you talk. When you’re deciding who to spend your efforts on, prioritize based on ability (ie. wealth), affinity (to your cause or you), access (people in your network, locale, or one degree away).
• Strategy. I’ve written about good strategy before, so I won’t rehash that. I do think it’s important to have a strong vision for your organization. Could you see your impact scaling 100x in five years, and what does that look like?
• Knowledge. Use AI a lot! It is so easy to learn these days. You can ask any LLM the exact questions you want answered and get great advice.
• Networking. Because your contacts may be helpful (eg. for funding, tips, encouragement), it’s good to start talking to people early. I recommend talking to low-stakes people who won’t be evaluating you, getting their goodwill, then moving on. However, if you’re extremely early-stage, reaching out to people to get their advice early can be useful. The 2-Hour Job Search and my post on job searching are relevant here.
• Go full-time. If you can go full-time, I recommend it. It helps to have a clear head, and having a good idea takes time. A lot of time! Focus is your superpower.
• Prioritize. Don’t put the cart before the horse. I do this a lot!
• Talk to users. If you are unsure how your idea will be received, talk to the people you’ll be working with. Talking to “users” in the non-profit sense means more: “do people care about this issue? What’s the right way to frame it?” Talking to users allows you to make a more informed pitch to funders. But also, when you talk to funders, you should find out what they want to hear / understand framing that works for them.

Some good resources:

• Start something that matters by the founder of Tom’s shoes, Blake Mycoskie. Encouraging, amusing, and down-to-earth. Good read for those unsure about pulling the trigger.
• Y-Combinator’s YouTube channel. Great content that teaches you to hone in on the core of your project, and also the importance of talking to users.
• Paul Graham’s excellent blog. Lots of people get inspired to found from this, and rightly so!
• Sam Altman’s excellent blog. Importance of self-belief, focus, and having great friends.
• Business entrepreneurship by Jim Collins. Collins’ writing is all extraordinary.
• Fundraising for Social Change. A little formal but useful! Don’t bother doing everything in this.
• Non-profit management plan. A good article I did not use but could be useful
• Non-profit operations playbook.
• How to start a nonprofit, an operational playbook. Lots of great resources linked here.

• Casio F-91W
• Claude Code
• My biking jacket from Kaohsiung Decathlon
• black tea
• Linux
• Quick dry sports towels
• Soy milk
• Tofu
• My house
• Second-hand clothes
• Friendship
• A kit for giving yourself haircuts

Price is determined by many factors, one of which is cost, but that’s not the full picture! There’s also “brand value” (which may or may not mean something to you); scarcity; how hard it is to produce; how hard it is to transport.

But the biggest source of inflated value to avoid is the lock-in effect. Apple users are the victims of this (no shade!). People like the look and feel of Apple products. But it’s so different from other products, from the OS to the charging cable to the green texts, that they can’t imagine getting another type of phone. Then they’re stuck with iPhones, even when the battery is intentionally reduced over time, and the prices remain sky-high even for old phones.

The lock-in effect doesn’t just apply to consumer products; it also applies to enterprise software. Some enterprise software is great, but also expensive. Slack is a good example – Slack can do a lot of things, so even though it’s somewhat expensive, people still recommend using it. HubSpot is an example on the other side. It seems like a great CRM, and maybe it’s the best, but I think the standards are so low.

Particularly now in the age of vibe-coding, it’s so easy to create your own custom software. I think this makes a lot of sense. In many cases, it takes just as much work to customize enterprise software as it would to make your own tool (there are people that dedicate their entire careers to helping people use bloated software like Salesforce, HubSpot, and the like!).

Granted, it’s hard to revolutionize a business that’s already running on one of those archaic pieces of junk. But I think that companies that can start afresh or make the transition to AI-first operations will soon outperform all competitors. And that goes for social movements like Sapiens First, too!

Here is some advice I would give to myself on March 2nd when I decided to found Sapiens First:

• Patience and persistence. I’ve really resonated with this idea that you overestimate what you can do in a day, but underestimate what you can do in a year. I’ve noticed that being patient, consistent and doing a little bit each day has been positive for my mood. It also teaches you to practice a long-run view.
• Optimism. Starting an organization demands that you get lucky. You can’t force luck to happen, and your self-esteem can’t rely on good things happening to you. However, you can maximize your exposure to luck by being persistent and meeting lots of people. That helps me feel more optimistic and not pin my hopes on any one opportunity.
• Focus. Focusing on one thing at a time is essential. One summer in Berkeley, I decided I would only do one thing a day. If I had lunch plans, and someone called me to have dinner, I told them no, we’ll have to do tomorrow instead. I’d argue it’s good to do this for work projects. Today’s sole task is to XYZ (eg. revamp the website). If it’s important, it should be important enough to dedicate a whole day.
• Choices. Choosing what to focus on takes some thought. I’ve accidentally put the cart before the horse, for example reaching out to someone before I was ready to talk to them. Normally that happens because I felt desperate to make something happen. Which means I didn’t have enough patience and persistence.
• Networking. You never know where a connection might lead. For an organization like Sapiens First, essentially anyone and their mother might have a productive role for the company: a funder, a volunteer, a employee, a collaborator, an introduction to a politician, etc. This has to be balanced with meeting deadlines and such. But I rarely say no to a call, it just might happen for a couple weeks.

Getting the first grant feels great! The validation feels great (though I’ve come to believe that funders know much less than entrepreneurs about what is a good idea). More importantly, I feel that I can keep working on this project, and have the funds to support myself and our early activities in the early months.

Here are some things to think about before you start token-maxing:

• Vision. Have you thought about what you want to do?
• Strategy. Have you considered a wide variety of ways to approach the problem? In creating a strategy, you can certainly have a dialogue with an LLM (without using that many tokens per unit of human effort).
• Prioritization. Your grand vision may be easier to implement than ever with AI, but it will still take time, effort, and tokens to vibe code. You should try to build a minimal viable product that gets 80% of the idea across.
• Talking to Users. Hear users’ concerns, and see what they like / don’t like about your MVP. If it’s a product for yourself, really sit down and think, “what do I want out of this?”
• Iterate. Keep building with AI, but don’t build everything all at once.

Some thoughts that are more pro-AI:

• AI at every stage. All of the above steps can be aided with a chatbot. Talk through the vision, do research on potential strategies, weighs pros and cons of prioritization.
• Long and descriptive prompts. Use voice to text and really speak your mind about AI. The more context you give it about what you want, the better. If you’re not sure what you want, tell it that! It is intelligent enough to help you brainstorm.
• AI-native environments. In building a product / company, every choice should be made with AI in mind. For example, transitioning from Google docs to plain text is a way that you can make your AI more powerful. Also, avoiding software that cannot be customized programmatically. For example, at my last company, we used Hubspot as a CRM – automations in Hubspot are built through a drag-and-drop editor! Agonizing. I believe that for essential, day-to-day products like a CRM, it may be worth it to build it yourself (and that’s where vibe-coding comes in).

Choices are the hardest thing to make. I don’t believe we’re taught how to make choices. I learned how to write a five-paragraph essay, and even how to create a pro-con list, and once in robotics club I learned how to create a table of values. It all felt rather inane to me. I don’t feel that I truly made a decision until I graduated from college. But others are different – some of my friends turned vegetarian when they were just 10 years old.

For me, the hardest part about making a choice is the grief at missing out. By choosing to do one thing, you forgo an alternate future. To choose is not just to choose once, but to choose over and over again. To choose is to stand by your choice when others question your actions. While others’ opinions about your choices do not always matter, there may be part of you that wonders “what if”.

Though it can be hard, making a choice is actually liberating. You are freed from the cloud of indecision. You can delve into a new world, and now have greater commonality with other people who made that same choice. Of course, you may make the wrong choice. Most choices can be undone, though it doesn’t always seem like it in the moment. Best to be strong and wrong, as my old jazz professor once told me (see Office Hours with Ben Goldberg).

Speak less, listen more. People will seek you out again.

Write succinctly. Others will read more.

Less is more.

One. Fog and Clarity. You’re bored and you don’t know what to do (fog). You decide that you will clean your room (clarity). You go to your room and you see that it’s so messy, you don’t know where to start (fog). You decide that what you need to do is first, sort things into piles (eg. dirty laundry, trash, unwashed dishes) then deal with the piles in turn (clarity). You then create all the piles, and now have to figure out what to do with the dirty laundry (fog), but the answer is easy (you put it in the laundry bin) (clarity), and then next is the trash and so on. You are able to go through the piles easily and without much thought, because the moments of fog are relatively easy to deal with. After you’re done sorting through the piles, you realize that your room is clean, and you have a momentary feeling of victory! Followed by the realization that you have to figure out what to do next (fog).

Two. Another way to describe a moment or period of fog is crisis. A small moment of crisis is just boredom. My old music teacher used to say how important it is to be bored / stuck, because (just like you have to be scared to be brave) you have to be stuck to innovate. Many people avoid crises, particularly existential crisis. For example, you may be unhappy at your job, but you refuse to quit, because if you quit, you would have to go through this deep, soul-searching period where anything is possible and you have no idea what to do. Some people are forced into crisis. For example, this happens when most people graduate from college, or retire, or their kids leave their house, or they have a debilitating injury. While these moments are hard in the short-term, I believe that there’s a way to engage with a “crisis” so that it has a transformative effect on your life.

Other ways to describe this concept: see the forest and see the trees; explore vs exploit (machine learning); productive & reproductive work; tangible & intangible progress.

Here are some observations about this concept:

• There seems to be a lack of vocabulary / lexicon in how to support someone through a moment of fog. If someone is graduating, your first question is then, “what kind of job are you looking for?” That assumes that they’ve resolved this rather large question, that they know exactly what they want to do next.
• Clarity is not always good. The reason is that people who have “Clarity” on what they’re doing might be doing a horrible thing to themselves / people around them. How many despots or autocrats have utmost clarity / conviction in what they’re doing? How many Big Oil executives turned a blind eye when concerns about climate change were brought up in the 70s? Cliff moments (eg. the death of a loved one, a disruptive moment) can force such people to take a more expansive world-view, if they’re ready.
• Clarity is the only way to get things done. The fog can also be called, “analysis paralysis”. When you’re in the fog, you’re not directly working on anything.
• Being in the fog longer can lead you to greater clarity. Here’s a great quote by Abraham Lincoln: “Give me six hours to chop down a tree and I will spend the first four sharpening the axe.” The equivalent in life to “sharpening the axe” can be reading about philosophy, productivity systems, taking care of your body, etc.
• The fog may take most of your thinking time. Decisions take a lot of energy, as does prioritization.
• The fog can be particularly long when you’re doing something you’ve never done before.

My general feeling is that people spend too little time sitting in the fog. I include myself in that – though I’m good at taking a step back, I don’t often make a plan or clearly prioritize and identify what I’m doing. But it can be overwhelming, confusing, discombobulating, but also beautiful.

In today’s world, learning how to program is like learning how to use an abacus. Fun, maybe. But also entirely obsolete because of advances in AI. Today’s engineers are high-level architects. Their job is to have taste, good judgement, and handle multiple AI agents. A particular company has essentially no need to hire grunt-level engineers because one engineer can do more work than ever before (and feeding/instructing AI is much easier than doing so for a human).

In fact, this same logic should apply to a host of other industries. We have seen an incredible rise in the number of lawsuits because it is now easier than ever to sue with AI. If our government were competently run, we could also improve the functioning of our legal system more efficient through AI-augmented judges. Accounting, finance, marketing, scientific research, data analysis will all be similarly affected in time.

The hold up is not in the raw ability of the technology, but rather in the incumbent organizations that are not set up to take advantage of these boons. It has long been known that big government and large companies are not good at innovation, and it is not for lack of resources. But I digress.

At first glance, you might think that specialists will be out of a job. Indeed, we no longer have professional abacus-users, and if someone were stubbornly attached to that profession, they might have trouble finding work in today’s society. But I actually think that most people will adapt. Being a specialist may be a point of pride for many, but it is rather demeaning to be treated as a person with only one narrow set of skills.

The short-term answer for those worried about their jobs in the age of AI is to become a generalist. Generalists think about the why, the big picture, the strategy and vision, the values and principles. Great generalists will use AI to create world-changing social enterprises, billion-dollar companies, and improve the lives of others. AI can and should be used as a tool, but for what is the most important question.