The hottest Decision Theory Substack posts right now

And their main takeaways
Category
Top Science Topics

La différence entre menace et avertissement

L'Atelier Galita • 39 implied HN points • 02 Nov 24
đź“– Philosophy Decision Theory
  1. Threats and warnings are not the same. A threat implies a promise of harm, while a warning offers a caution about potential danger.
  2. Decision-making can be influenced by understanding these differences. Knowing how people respond to threats and warnings helps in planning actions.
  3. Real-life examples can illustrate the impact of threats versus warnings. Recognizing these concepts can improve communication and strategy in various situations.

Claude's Constitutional Structure

Don't Worry About the Vase • 3942 implied HN points • 26 Jan 26
đź•ą Technology Decision Theory
  1. Favor judgment over rigid rules. The system should be trained to cultivate good values and practical wisdom so it can handle novel situations instead of relying on brittle, hard-coded rules.
  2. Make decision theory and commitments explicit. Using a clear decision-theoretic framework (and observable commitments to the model) helps produce reliable cooperation and better long-run behavior.
  3. Prioritize safety, ethics, compliance, then helpfulness, and respect role hierarchies. The AI should be corrigible, avoid manipulation, protect user wellbeing, and follow maker → operator → user priorities while putting ethical constraints first.

The Claude Constitution's Ethical Framework

Don't Worry About the Vase • 2284 implied HN points • 27 Jan 26
đź•ą Technology Decision Theory
  1. Design the AI around virtue ethics: aim for it to be a genuinely good, wise, and practically skillful agent who behaves like a deeply ethical person rather than getting stuck resolving abstract philosophical debates.
  2. Treat honesty as a near‑absolute norm: avoid white lies and manipulation, be transparent about uncertainty and intentions, and refuse instructions that would require deceptive or harmful behavior.
  3. Combine firm hard constraints with nuanced value balancing: explicitly forbid aiding mass harm (weapons, cyberattacks, power grabs, CSAM) while weighing competing values like education, autonomy, fairness, and harm prevention, and handle moral uncertainty with coherent, context‑sensitive judgment.

The game theory of cooperation

Optimally Irrational • 72 implied HN points • 09 Mar 26
đź“– Philosophy Decision Theory
  1. When people expect to meet again, conditional strategies like "I’ll cooperate if you do" make cooperation a rational, self-interested choice because future losses deter short-term cheating.
  2. Reputation, indirect reciprocity and partner choice scale cooperation: public records, gossip and the ability to shun defectors let groups enforce cooperative norms even when partners change.
  3. Cooperation has multiple roots — kin selection, reciprocal altruism and cultural evolution — and because many cooperative equilibria are possible, societies pick and stabilize particular norms while moral feelings help people follow and enforce them.

Strategy in four worlds

Subconscious • 1265 implied HN points • 11 Jan 26
đź•ą Technology Decision Theory
  1. Risk and uncertainty are different: risk is measurable and fits expected-utility tools, while uncertainty involves unknown possible outcomes and needs a different approach. You can categorize environments as clear, complicated, complex, or chaotic based on how cause and effect behave.
  2. Match your tactics to the environment: clear and complicated problems reward forecasting, expert analysis, and optimization, whereas complex systems require robust, antifragile strategies that map feedback loops, and chaotic situations demand fast reflexes and simple orientation to survive.
  3. Scenario planning is the right tool for complexity: it helps identify major drivers, surface feedback loops, and wind‑tunnel strategies across many plausible futures so you can build robustness or intentionally shape outcomes. Because real challenges mix these worlds, skilled strategists combine forecasting, scenarios, and adaptive judgment rather than relying on one model.
Get a weekly roundup of the best Substack posts, by hacker news affinity:

Reincarnation and Suicide

Fake Noûs • 123 implied HN points • 21 Feb 26
đź“– Philosophy Decision Theory
  1. If time stretches infinitely in both directions, the fact that you’re alive now makes it unlikely you only live once, which supports the idea of reincarnation.
  2. Even if reincarnation is real, death still destroys your memories, relationships, possessions, and learned abilities, so dying prematurely is usually a bad loss.
  3. Whether suicide is rational depends on expected future utility: without reincarnation it would be rational if your known future utility is negative, but with reincarnation you should compare your life’s utility rate to the average utility you expect in future lives, and uncertainty generally favors waiting.

Good if make prior after data instead of before

DYNOMIGHT INTERNET NEWSLETTER • 796 implied HN points • 18 Dec 25
đź“– Philosophy Decision Theory
  1. When the true hypothesis space is large or continuous, compressing it into a single coarse prior hides important differences and can produce misleading posterior probabilities.
  2. It often helps to look at the data first to see which distinctions matter, then define finer categories and ask how likely you would have judged those categories before seeing the evidence.
  3. In practice the simplest practical fix is to refine your hypothesis categories so the data likelihood is roughly constant within each category, because grouping poorly can under- or overestimate the probability of different outcomes.

Dekanting moral philosophy

Optimally Irrational • 59 implied HN points • 06 Feb 26
đź“– Philosophy Decision Theory
  1. Kant’s categorical imperative doesn’t follow from pure rationality because your individual choice can’t make others follow the same rule, so behaving as if everyone would comply can be irrational in strategic situations.
  2. Game theory shows morality is best understood as self‑enforcing social conventions: stable moral rules are conditional “oughts” that arise because following them serves each person’s interests given what others do.
  3. Evolved moral feelings make cooperation feel like an absolute duty, but treating those feelings as unconditional can produce worse outcomes in problems like prisoner’s dilemmas, mutual deterrence standoffs, or strategic voting.

The Surprising Computational Properties of Love and Voluntary Suffering

apxhard • 51 implied HN points • 08 Feb 26
đź“– Philosophy Decision Theory
  1. Love works like an outward-pointing utility that breaks self-referential loops and gives you clearer, less anxious targets to aim for.
  2. Loving many people widens your sample of reality and links your wellbeing to others, which prevents overfitting to your own experience and smooths emotional spikes.
  3. Choosing to endure short-term suffering lets you move against immediate pleasure gradients to escape local traps, and combined with love this grants much greater freedom to reach better long-term states.

We Should Try to Directly Measure the Value of Scientific Papers

Nonsense on Stilts • 59 implied HN points • 20 Jul 24
🔬 Science Decision Theory
  1. We should measure the value of scientific papers to understand their real impact. If a paper doesn't change how people act or think, then it may not be worth much.
  2. To figure out the value of a paper, we can use a formula that compares what outcomes we expect with the information from the paper versus without it. This helps us see if the research is actually useful.
  3. It's important to have good estimates and decisions tied to the research to see its true worth. By doing this, we can better judge which scientific papers are really making a difference.

1/(Effective Altruism)

Philip’s Newsletter • 41 implied HN points • 13 Dec 25
đź“– Philosophy Decision Theory
  1. Effective Altruism often treats distant and future lives as equally important, but it can also make sense to discount impact by distance and time and prioritize helping those nearest to you.
  2. If many people are giving, focusing on nearby recipients can increase measurable impact and coordination because local help reduces uncertainty and leads to better collective outcomes.
  3. Caring more about people close to you can support social stability and cooperation, and regardless of strategy, giving more generally benefits both recipients and the giver.

Bayes is not a phase

DYNOMIGHT INTERNET NEWSLETTER • 453 implied HN points • 27 Feb 25
đź“– Philosophy Decision Theory
  1. Bayesian reasoning is something we all use, even if we don't realize it. It's more about how we naturally think than some complex math.
  2. There are two types of uncertainty: aleatoric (random) and epistemic (based on knowledge). Mixing them helps us make better decisions.
  3. Arguing over which type of probability is 'real' is silly. It's better to recognize that life involves many messy decisions where formal reasoning can help, but is often complicated.

Does a hundred wrongs make one right?

inexactscience • 59 implied HN points • 04 Mar 24
đź’° Finance Decision Theory
  1. A famous bet involving coin flips shows how people's risk preferences can be inconsistent. People might reject a single gamble but accept multiple repeats because they think it lowers their risk.
  2. The original advice about investing suggests buying stocks when young and bonds as you age. However, Samuelson's argument raises doubts about this common belief, challenging how we think about risk.
  3. The idea of loss aversion helps explain why people might choose to repeat risky bets. People tend to feel the pain of losing money more than the joy of gaining, which can lead to seemingly irrational decisions.

If You Want to Hit Your Target, Take More Shots

The Palindrome • 6 implied HN points • 18 Dec 25
đź“– Philosophy Decision Theory
  1. If you want to hit your target, take more shots — more attempts raise your chance of success.
  2. Trying lots of ideas across different areas (projects, posts, dating, work) leads to more wins because each attempt gives feedback you can learn from and improve.
  3. Unlikely successes become likely with enough trials, so don’t be discouraged by early failures — persistence and volume pay off.

Finding the third way

The Uncertainty Mindset (soon to become tbd) • 59 implied HN points • 31 May 23
đź“– Philosophy Decision Theory
  1. There are two common reactions to uncertainty: one is to act like everything is knowable and try to control it, which can lead to poor decisions. The other is to give up and think that nothing can be done about the unknown, which doesn't help either.
  2. Instead of sticking to those two extremes, there's a better approach. It's important to recognize that not-knowing can lead to new ideas and actions.
  3. We can break down uncertainties into different types. Understanding these helps us figure out how to deal with situations where we don't have all the answers.

A brief note on Newcomb's problem

Philosophy bear • 28 implied HN points • 02 Mar 25
đź“– Philosophy Decision Theory
  1. Newcomb's problem involves two boxes: one potentially holds more money than the other based on what an oracle predicts you will choose. Choosing both boxes might seem smart since it guarantees some amount, but there's a deeper philosophical debate about the best strategy.
  2. Some people strongly believe that just picking the blue box (the one with the most potential money) is obviously the best choice, but it's unfair to call those who pick both boxes 'stupid' since their reasoning also makes sense.
  3. Ultimately, this problem raises important questions about decision-making and what it means to want something, suggesting that understanding our desires and decision strategies is complex and varies from person to person.

My Talk About the Sleeping Beauty Paradox

By Reason Alone • 16 implied HN points • 07 Nov 24
đź“– Philosophy Decision Theory
  1. The Sleeping Beauty paradox involves a coin flip that affects how often she wakes up, which raises questions about probability. People have different opinions on how she should assess the chance of heads when she wakes up.
  2. One group, called 'halfers', believes the chance of heads remains 50/50 since she doesn't gain new information about the coin when waking up.
  3. Another group, 'thirders', argues she should think there's a one in three chance it's heads because of how many times she might wake up, depending on the coin flip.

#9: Tradeoffs

The Uncertainty Mindset (soon to become tbd) • 79 implied HN points • 01 Jan 20
đź“– Philosophy Decision Theory
  1. Life is full of tradeoffs. When you want something, like a stable job or a big house, you may have to give up other freedoms, like moving easily or having more free time.
  2. It’s important to clearly define what tradeoffs you are okay with. This helps you make better decisions, whether it’s about what to eat or where to live.
  3. Instead of just asking what you want, think about what you are willing to sacrifice. This question helps you understand your priorities and can lead to clearer choices.

Total and Partial Ordering

Kartick’s Blog • 0 implied HN points • 13 Feb 25
🚌 Education Decision Theory
  1. Total ordering means you can rank all your options from best to worst. For example, you can see which forex provider offers the best deal for your money.
  2. Partial ordering is when you can compare some options, but not others. Like, you might know one car is better than another, but you can't say how a car compares to a scooter.
  3. In total ordering, you end up with one best choice, while in partial ordering, there can be multiple good choices without a clear best.