Drake Morrison

If It’s Worth Arguing, It’s Worth Arguing With Whiteboards

2026-04-17T00:00:00Z

It’s easy to disagree with people. You just say, “That’s wrong” and decline to elaborate.

But that’s not very interesting. If you want to be making progress — instead of ragebaiting — it usually helps to find a way for your disagreement to be productive.

Productive disagreements start in the place where you are already in agreement. The places where your models of reality overlap. Without any overlap at all, it gets quite difficult to successfully converse. You need to be speaking a language your interlocutor can understand, for starters. That’s actually not trivial to establish. Defining your terms before a debate is done for a reason, and it can be quite hard. Especially when the thing you are arguing over doesn’t have a good term or definition in the first place.

That’s hard enough, but the more difficult thing in my view is to not get a false understanding of your interlocutor’s view. Sometimes a friend will say something like, “taxation is theft” and I nod along, as I backfill their claim with a possible set of observations and inferences. Yet when I interrogate the claim, I find the path my friend took to get there — from the observations they make to the inference chains that follow — are wildly different than the ones I constructed in my head. It’s actually really easy to think you understand someone in this way, and it’s harder to notice when you are going fast and making background assumptions before getting to the real thing you wish to argue over. Those background assumptions hold a lot of your worldview in them.

Going back to basics, I think it’s important to sort out your observations from your inferences. When you want to make progress on a disagreement, drag over a whiteboard and have each party make a list of relevant observations. Compare notes. Only then make lists of inferences from those observations, and argue over those. Often people compare entire claims about the world. That’s a lot harder to do than to individually compare observations or inferences separately.

People seem reluctant to do this. Perhaps because it means you have to put some work in to it, and it stops being a performance that gets you re-invited to parties. It could be the legibility, writing down your arguments exposes you to enduring criticism down the line. Or maybe it removes the ability to retreat into vagueness when pressed, which makes the lack feel much sharper. It can feel nice to let your friend and yourself off the hook a little, but my stance is this: If your argument is worth having, it’s worth having with whiteboards.

When you write down the basic building blocks that formed your worldview, it’s a lot simpler to compare them. Sometimes one party has a much longer list and it makes sense for the argument to turn into more of a situation where one person is teaching another. This is a much better outcome than the argument going nowhere because neither side knows how much evidence the other is actually sitting on.

It’s easy to disagree with people. But with a whiteboard, you can actually make progress. You may find you agree on much more than you expected, and can learn a little from each other. Writing down your observations and inferences has a tendency to clarify the few links where your disagreement lives.

It’s easy to disagree with people. With a whiteboard, you can actually make progress. You may be surprised, once both lists are up, and see that you agree on most of the observations and most of the inferences. The disagreement lives in one or two specific links — an inferential step one of you is taking that the other doesn’t see, or observational evidence one of you has been quietly relying on.

If you’re trying to clearly communicate, if you want your argument with someone to go anywhere, grab a damn whiteboard and make some lists.

If You’ve Never Bought a Tool You Didn’t Need, You’re Not Buying Enough Tools

2026-04-16T00:00:00Z

Having more tools greatly increases the space of potential plans you can consider. So by default you should be getting more tools a. Having too few tools tightly constrains your options, and often impacts the kinds of plans you consider in the first place.

Years ago I thru-hiked the Appalachian Trail, and every piece of gear mattered. I had a lot of things when I first started out that seem very foolish to me now, things I wouldn’t consider bringing again. For each piece of gear, I had to carry every gram of it 2200 miles, and so each gram had to earn its weight. I tossed or gave away almost everything that I wasn’t using every single day by the end of the second week. The only exceptions were some band-aids and similar items to prevent the worst outcomes. When I fell and split open my knee, I used toilet paper to pack the wound, and some duct tape to secure it. It worked great while I continued to hobble along. Every piece of gear had to pay for its weight, and part of that was finding creative uses for the pieces I did have, to cover more functions. I spent a lot of time optimizing within the strict constraints I had: weight, pack volume, distribution of frequently-used items with once-a-day items. I paid for this optimization with some increased risk and a greatly narrowed space of things I could do while on the trail. That was fine then, because the only thing I was doing was hiking.

Nowadays I’m not carrying everything I own on my back, hiking 20 miles every day to get to the next town and resupply before I run out of food. So the constraints are really quite different! I can buy and store random things like a cheap multi-tool just to test the idea of having a multi-tool around, and purchase a high-quality one once I’ve gotten a sense of how much it’s worth. When the cost of tools is so low, buying cheap tools can be a cost-efficient way to explore a new space of plans. The cheapness of tools is nice, but what I didn’t expect was how much this mindset of extreme minimalism affected my thinking.

There were so many tiny things that I wasn’t considering, merely because the tool to do them wasn’t immediately available. When making simple plans I often use a simple algorithm and ask myself three questions: What do I have? What do I want? How can I use the former to get the latter? ¹But this misses a step wherein I can get more tools to widen the space of plans!

For tools I am proficient with — asking myself the question, “What do I have?” naturally includes them. They became an extension of me to some degree. But for tools I’m not proficient in, using them (or even learning) doesn’t cross my mind as part of a plan at all! Many tools are actually quite easy to learn, and the cost is low enough that I should have been making plans where I would consider them anyway, but I wasn’t. Having a tool expands your plan space; being proficient makes that expansion automatic.

If I get a rip in my socks, sewing them back up by hand quickly and easily rises to my mind as one possible plan to fix this problem. Yet if I didn’t have a needle or thread, or the very simple knowledge in how to use them, it wouldn’t be a consideration. I have witnessed this many times when a friend bemoans a rip in their favorite jacket, and manually sewing it back up doesn’t even cross their mind.

Or maybe it doesn’t cross their mind because they don’t know how to sew — not because they lack needle and thread. Yet I would bet if they had a needle and thread lying around, they might consider trying it before resorting to more expensive plans. The physical presence of a tool can be a cognitive prompt to use it.

Investing the time to become proficient in some general-purpose tools isn’t just useful for a current project, but it increases the types of projects you can now undertake. I’ve talked a lot about physical tools, but this applies to tools for thought like calculus, statistics, programming, Excel, Photoshop, etc.

Many people are blocked more by the plans they’re able to consider than by money. I think they should be spending more time on getting and learning to use new tools. Or phrased as a standard Umeshism a: If you’ve never bought a tool you didn’t need, you’re not buying enough tools.

From Alicorn’s glowfic!bella stories ↩

Carpathia Day

2026-04-15T00:00:00Z

(The better telling is here a. Seriously you should go read it. I’ve heard this story told in rationalist circles, but there wasn’t a post on LessWrong, so I made one)

Today is April 15th, Carpathia Day. Take a moment to put forth an unreasonable effort to save a little piece of your world, when no one would fault you for doing less.

In the early morning of April 15, the RMS Titanic began to sink with more than two thousand souls on board.

Over 58 nautical miles away — too far to make it in time — sailed the RMS Carpathia, a small, slow, passenger steamer. The wireless operator, Harold Cottam, was listening to the transmitter late at night before he went to bed when he got a message from Cape Cod intended for the Titanic. When he contacted the Titanic to relay the messages, he got back a distress signal saying they hit an iceberg and were in need of immediate assistance. Cottam ran the message straight to the captain’s cabin, waking him.

Captain Arthur Rostron’s first reaction upon being awoken was anger, but that anger dissolved as he came to understand the situation. Before he’d even finished getting dressed he had ordered the ship to turn around and make towards the last known position of the Titanic.

Even knowing they wouldn’t make it in time, Rostron ordered the heating and hot water turned off, to divert all the steam towards the engines in the hopes of getting more speed. He ordered lights to be rigged along the side of the ship so survivors could see it better, medical stations set up, opened the kitchen and got soup and coffee readied. He ordered extra shifts from the crew. Nets and ladders rigged down the sides; lifeboats swung out ready; three dining rooms converted to triage with a doctor in each. He ordered extra lookouts to help navigate the ice field — they had hours of dodging icebergs in the dark ahead of them as they sailed at speeds his ship was never rated for. The RMS Carpathia was only rated for 14 knots; and the engineers squeezed out 17.5. One steward recalled the captain announcing to the crew, “We are in danger. I am risking your lives. The Titanic is in trouble and is sinking and we have to go help them.”¹

When they arrived three and a half hours later, the Titanic had already sunk. Most of the passengers had already died. The crew did not let this dissuade them, and for hours afterwards they were finding lifeboats and rescuing people. The passengers of Carpathia helped where they could, offering blankets, warm drinks, and words of comfort.

The captain, crew, and ship were honored for their efforts, and the story is well-remembered as a model response to disaster. A little light in a dark time.

Carpathia represents the virtue of toiling in spite of knowing you’re probably going to fail. Trying your damnedest to eke out a little more, just a little faster, for the sake of people you don’t even know. Being willing to sacrifice comfort, space, warmth, all for the sake of something greater. Even if it wouldn’t work out. Even if it won’t be worth anything in the end. It will still be worth it to you. Knowing you tried. Making an extraordinary effort. It’s not the kind of effort you can sustain, but sometimes it’s worth actually trying that hard, in the time it matters most. So today, let us celebrate the passengers, crew, and captain of Carpathia — and consider what we would drop everything for.

“Joseph Zupicich, 94, AIDED TITANIC VICTIMS”. The Morning Call. 14 April 1987. ↩

Plans are Not Promises

2026-04-14T00:00:00Z

I have been pondering Wentworth’s Information vs Assurance a post, and I think I have another angle on a similar concept. Wentworth identified the problem of information being mistaken for assurance. I want to explore what it looks like to deliberately construct assurance, and what that costs. Conveying this concept in different communication-cultures is hard — plans get mistaken for promises, and promises for plans.

What’s the difference? Well, plans tell you something about the way the planner views the state of the world, the state they wish the world was in instead, and the process by which they plan to change it. If I tell you I’m planning on taking Route 9 to pick you up from the airport, this tells you something about my model of traffic, my schedule flexibility, maybe even my risk tolerance. Often someone sharing a plan with you is an invitation. Not just to the plan but also the frame that generated it — the worldview through which they see the world and measure risks and opportunities. There’s also often an implicit caveat that plans change. There is an impermanence to plans, indeed that’s part of their usefulness! You can update a plan on new info, change the trajectory you are aiming for, change who is involved. When a plan fails, there are questions about how to predict obstacles earlier or who could have acted differently. People may feel sad the plan failed to succeed, but rarely do people feel trust was broken. Plans are statements of intention, not a commitment to outcome.

Promises are different. People are rightly mad at you if you break one of those. There is weight and expectation built into a promise. A promise is a plan with an assurance your counterpart can put weight on. They can invest in the world where the promise is held, and hold you accountable if the promise is not. Contracts are a close analogy, synonymous in some cases. Breaking a contract comes with legal consequences, in part due to the attempt to recompense the recipient what they lost by investing — on your assurance — in a world that didn’t exist. A promise comes with the expectation that you will sacrifice to keep your promise, you will be willing to undergo some inconvenience to keep your word. What’s at stake isn’t just the trust of the people you made promises with, but also the possibility of that better world that could be invested in.

Keeping promises is hard though, what do you do when circumstances change? What if an emergency arises? Do you become like Wei Sheng¹ — who drowned rather than break his word — or do you lose a bit of trustworthiness and hope it’s ameliorated by the understanding of your counterpart. What happens when there is disagreement about reasonable circumstances? I would feel aggrieved if I put a bunch of resources into a course of action that doesn’t make sense anymore because you broke your promise over something I consider trivial, even if it’s not trivial to you. I might even understand, but that doesn’t change my loss! Does it matter how much I’ve invested? It gets tricky fast, especially when people have different expectations about what constitutes an exception they won’t hold against you.

Enter bonds. They are my new favorite mechanism for making the cost of promises legible. I recently had someone ask to borrow a $180 board game on a trip to Yosemite. I told them if they gave me $180, they could take the game, and I would give them back the money once the game was returned. This sort of thing made it super clear exactly the cost of the promise they were making, and if it failed, I would keep the money as recompense. It was so simple and clean! Expectations were very clear, we both knew the size of the promise and the cost of breaking it. With the cost legible and the bond in hand, I wouldn’t feel betrayed if they lost the game on their trip. I could invest in the world where I had the game. I had assurances. If a very costly promise is made, it matters a lot the promise is kept, and with bonds the cost can go to the person making that promise. If an emergency comes up and I feel like I need to break some promise, but am not sure if the person I made the promise to would agree the emergency is justified, I have to risk breaking trust and having a long-drawn out fight over it. With a bond, I can directly measure the cost of the emergency against the bond amount. It’s much simpler and clearer to everyone involved.

You don’t have to make bonds denoted in money — I think most are implicitly made with reputation — but money has many nice properties, not the least of which is money can be used for recompense should the promise fail. I once made bonds with my roommate denoted in chores.

Many people dislike the cold, calculating nature of bonds denoted in money. They like promises made with trust and reputational costs as the bond. I think this works sometimes in tight-knit communities where reputations are common knowledge. Not only does the promise breaker lose some reputation and trust from the community, but the community can enforce a claim against them to recompense the person they promised. This stops working as soon as any of it gets complicated — the group grows too large to establish common knowledge reputations, or different factions disagree on what’s a fair exception. It gets a lot harder to get reimbursed for a claim of damages when this happens, even if the promise breaker still loses reputation and trust.

Thinking of promises as bonds also reveals interesting things about the way people think of the bounded-nature (or not) of promises. Many promises are implicitly bounded to time or context. But not all are. If I say, “$5 says I can climb that tree” suddenly my friends want to know how many tries I get, how it only counts if one of them can witness it, etc. When there’s money on the line, suddenly people want to get into the details and operationalize the context. When someone asks to borrow my game, I’ll ask for a bond, but I’ll also ask when they’re expecting to return the game. It is wise to have exit clauses and expiration dates for contracts you make, and in a similar manner thinking about the bounded-nature of a promise helps to prevent it from becoming all-consuming.

There’s a relationship here I see no way around. Promises are only as strong as what you would sacrifice to keep them. Yet paradoxically, being willing to lose everything to keep a promise makes it very difficult for me to trust you, because you can’t give me everything as a bond. There isn’t a way for me to collect on that kind of collateral.

There’s a fundamental limit to how strong a promise can be, and if there are many things that are worth everything to you, any one of those things is worth breaking a promise over, and I’m back to not trusting you.

Bonds are not a panacea for all contractual coordination problems, but they do seem underutilized in the interpersonal space. I hope you find ways to include them more, and make the costs of your promises legible — else declare them plans.

By default, when someone tells me they are going to do something, I take it as a plan, a statement of intent. If you want me to take a commitment of yours seriously, show me the bond and I’ll tell you how much it’s worth to me.

Chinese parable about a man who made a promise to meet a woman under a bridge. When she didn’t arrive and floodwaters rose, he clung to the bridge pillar and drowned rather than break his word. As told in the Zhuangzi (Chapter 29, ‘Robber Zhi’). ↩

Meaningful Questions Have Return Types

2026-04-13T00:00:00Z

One way intellectual progress stalls is when you are asking the Wrong Questions a. Your question is nonsensical, or cuts against the way reality works. Sometimes you can avoid this by learning more about how the world works, which implicitly answers some question you had, but if you want to make real progress you have to develop the skill of Righting a Wrong Question a. This is a classic, old-school rationalist idea. The standard examples are asking about determinism, or free will, or consciousness. The standard fix is to go meta. Ask yourself, “Why do I feel like I have free will” or “Why do I think I have consciousness” which is by itself an answerable question. There is some causal path through your cognition that generates that question, and can be investigated. This works great for some ideas, and can help people untangle some self-referential knots they get themselves into, but I find it unsatisfying. Sometimes I want to know the answer to the real question I had, and going meta avoids it, or asks a meaningfully different question instead of answering it. Over time, I’ve stumbled across another way to right wrong questions that I find myself using more often.

In programming there is the idea of return types a — constraining and declaring the type of data structure that a given method or function will return. Having a return type means I can rely on this function not to mess up something else later down the line that requires a particular input. I can more easily interpret what the function is doing (or supposed to do), and can use it more effectively in my program.

I find this a useful metaphor when thinking about how to make progress on a difficult question. It is a different angle than the “go meta” approach; it routes around the self-referential nature of asking why I think some question is answerable and reframes the question to focus more on what shape the answer must take. To ask, “What’s the return type of, ‘do I have free will?’” is very useful. Suppose I ask this question in a group of friends and someone laughs and tells me, “Yes! Of course you have free will.” Have they answered my question? Do I still feel confused? I know I would. I don’t even have to ask friends, I can just imagine friends giving me different kinds of answers, and compare this to my feeling of confusion. Would this shape of answer actually dissolve a my confusion?

Often to answer confusing philosophical questions, I need a finer-grained picture of reality. I need a better causal chain of the dynamics involved in the question, a higher dimensional conception of the problem. This is useful because it lets me verify my understanding. I can follow directly the chains of cause and effect, I can see the logic clearly. A better map gives me the ability to tell if the answer is correct.

If I can’t tell whether some answer is correct or not, I either haven’t defined the question well enough, or my map of reality is sorely inadequate. If I can’t reframe a question to have a return type it reveals my confusion is deeper than I suspected and I need to take a step back. I need to step down and focus my uncertainty a into something more basic and foundational that I can get a handle on.

Consider: “Is the universe deterministic?” What’s the return type of this question? A binary yes/no works, but it wouldn’t resolve my confusion. How about, “How could I tell if the universe is deterministic?” The return type for this question is more answerable, and gets at my real confusion better. So how could I tell? With a detailed enough causal map of reality I could answer this question, so I would start there.

What about, “What is consciousness?” What’s the return type of that question? How could you tell if your answer was correct? How could you tell if your answer was wrong? Imagine a friend saying, “I’m conscious because subjectively it feels like I am conscious” Are you less confused? How about, “Everything is conscious, everything has a subjective experience of being itself” Can you verify this answer? If there was a detailed theoretical bridge between phenomenology and cognitive science, pinned down with experiments, would that enable you to answer the question? Is that then the return type?

This is related to the idea of operationalizing your bets. When you make a prediction market, figuring out how to resolve the market is a big question, and the strategies market makers use to turn abstract, weird, questions into resolvable markets are techniques you can use yourself for personal intellectual progress. “How do I resolve this market?” is almost the same question as, “How do I answer this question?”

When thinking through difficult philosophy, asking yourself how you are going to verify the answer is an underrated tool. Not only does it help you answer the question, it acts as a diagnostic tool for discovering just how confused you are, and where the edge of that confusion lies. The next time you find yourself in some philosophical quandary with your friends consider asking yourself, “What’s the return type?”