The hottest Reasoning Substack posts right now

And their main takeaways
Category
Top Technology Topics

Olmo 3: America’s truly open reasoning models

Democratizing Automation • 934 implied HN points • 20 Nov 25
🕹 Technology Reasoning
  1. Olmo 3 offers open-source language models that are competitive in performance, allowing the community to explore AI effectively. Both the 7B and 32B models set new standards for open reasoning models.
  2. The project includes a variety of training options to meet different needs, ensuring users can specialize their models for tasks like reasoning and instruction-following. It's all about making AI more accessible and adaptable.
  3. There’s an exciting future for research in reinforcement learning and model development with Olmo 3. The researchers are eager to explore new avenues and improve model capabilities over the coming years.

The Sequence AI of the Week #809: Slow Thinking, Fast Discovery: Inside DeepMind’s Aletheia Architecture

TheSequence • 35 implied HN points • 18 Feb 26
🕹 Technology Reasoning
  1. Aletheia is a DeepMind research agent built on the DeepThink architecture that emphasizes slow, deliberate “System 2” reasoning for autonomous scientific discovery.
  2. It shifts models away from fast next-token prediction toward verification and self-correction, aiming to reduce hallucinations and improve reliability.
  3. By giving the agent tools and the ability to check and admit mistakes, Aletheia enables deeper, more trustworthy exploration and problem solving.

Gain of Function.

News Items • 471 implied HN points • 18 Jan 24
🔬 Science Reasoning
  1. AlphaGeometry AI system solves complex geometry problems as well as a human Olympiad gold-medalist.
  2. AlphaGeometry combines neural language model with a rule-bound deduction engine for reasoning.
  3. Development of AlphaGeometry highlights AI's logic reasoning progress and ability to discover and verify new knowledge.

Are The Ancients Wise?

Bentham's Newsletter • 353 implied HN points • 15 Jan 24
📖 Philosophy Reasoning
  1. Some people believe that the wisdom of the ancients is overrated.
  2. Knowledge and understanding have evolved over time, making modern insights valuable.
  3. Respecting the ancients does not mean we should unquestioningly accept their views.
Get a weekly roundup of the best Substack posts, by hacker news affinity:

Meta-rational practices: Overview 1

Meaningness • 219 implied HN points • 09 Mar 24
📖 Philosophy Reasoning
  1. Meta-rationality is different from traditional rationality and requires an open-ended inquiry and responsiveness to various contexts and purposes.
  2. Meta-rationality involves ongoing consideration of when and how to apply rationality, recognizing situations where rational methods may not be sufficient.
  3. The norms of reasonableness, rationality, and meta-rationality differ in terms of accountability, formal rules adherence, and responsiveness to context and purpose.

How Does "Reasoning" Work?

In My Tribe • 379 implied HN points • 04 Feb 25
🕹 Technology Reasoning
  1. Reasoning in AI often involves finding and using analogies to solve problems. Just like a chess program cuts down on bad moves, AI looks for the best comparisons to answer a question.
  2. Human thought relies heavily on metaphors, which are used to understand new ideas. These metaphors can be good or bad depending on how well they fit the situation.
  3. Both humans and AI have strengths and weaknesses in reasoning. AI can be quicker but may miss the deeper meaning in a question, while humans can make creative leaps but might take longer.

The Latent State Window and the Limits of Reasoning in LLMs

jonstokes.com • 164 implied HN points • 05 Jul 25
🕹 Technology Reasoning
  1. LLMs have limits when it comes to reasoning. If a problem is too complex or involves too many moving parts, the model can struggle to find a solution.
  2. The size of a language model's 'latent state window' matters. This window limits how much information the model can hold while trying to reason, separating it from just the number of tokens it can handle.
  3. To get good results from LLMs, it's best to keep tasks simple and broken down into manageable pieces. If you give the model too much to juggle at once, it won't perform well.

Chain of Thought Prompting for LLMs

Deep (Learning) Focus • 294 implied HN points • 24 Apr 23
🕹 Technology Reasoning
  1. CoT prompting leverages few-shot learning in LLMs to improve their reasoning capabilities, especially for complex tasks like arithmetic, commonsense, and symbolic reasoning.
  2. CoT prompting is most beneficial for larger LLMs (>100B parameters) and does not require fine-tuning or extensive additional data, making it an easy and practical technique.
  3. CoT prompting allows LLMs to generate coherent chains of thought when solving reasoning tasks, providing interpretability, applicability, and computational resource allocation benefits.

The Liar Paradox (and Russell's Paradox) for Non-Logicians (UNLOCKED)

Philosophy for the People w/Ben Burgis • 399 implied HN points • 22 Jan 23
📖 Philosophy Reasoning
  1. The Liar Paradox questions whether statements can be both true and false, challenging fundamental logical principles like Bivalence and the Law of the Excluded Middle.
  2. Russell's Paradox, on the other hand, questions the existence of sets based on self-referential properties, leading to contradictions like a set that contains itself and doesn't.
  3. The debates around these paradoxes highlight the importance of classical logic principles like the Law of Non-Contradiction and Disjunctive Syllogism in everyday reasoning and understanding the world.

3 Tests AGI Must Pass

Insight Axis • 237 implied HN points • 27 Aug 23
🕹 Technology Reasoning
  1. Computers must excel at calculations to form the foundation for any further intelligence programming.
  2. After calculation, computers need to progress to reasoning - the ability to evaluate information and use it to make value-based decisions.
  3. The ultimate test for artificial intelligence is creativity - the capability to acknowledge rules but break them intuitively to create something new.

The latest open artifacts (#6): Reasoning models, China's lead in open-source, and a growing multimodal space

Democratizing Automation • 261 implied HN points • 27 Jan 25
🕹 Technology Reasoning
  1. Chinese AI labs are now leading the way in open-source models, surpassing their American counterparts. This shift could have significant impacts on global technology and geopolitics.
  2. A variety of new AI models and datasets are emerging, particularly focused on reasoning and long-context capabilities. These innovations are making it easier to tackle complex tasks in coding and math.
  3. Companies like IBM and Microsoft are quietly making strides with their AI models, showing that many players in the market are developing competitive technology that might not get as much attention.

Program-Aided Language Models

Deep (Learning) Focus • 196 implied HN points • 22 May 23
🕹 Technology Reasoning
  1. LLMs can struggle with tasks like arithmetic and complex reasoning, but using an external code interpreter can help them compute solutions more accurately.
  2. Program-Aided Language Models (PaL) and Program of Thoughts (PoT) techniques leverage both natural language and code components to enhance reasoning capabilities of LLMs.
  3. Decoupling reasoning from computation within LLMs through techniques like PaL and PoT can significantly improve performance on complex numerical tasks.

The Sequence Opinion #672: Mind Over Model: Chain-of-Thought vs. System 1/System 2

TheSequence • 105 implied HN points • 26 Jun 25
🕹 Technology Reasoning
  1. Chain-of-thought reasoning in AI helps it to process and structure information more clearly. This is similar to how humans take time to think through problems rather than jumping to conclusions.
  2. Human thought has two systems: System 1, which is quick and instinctive, and System 2, which is slower and more deliberate. This comparison helps us understand AI reasoning better.
  3. Understanding the similarities and differences between AI reasoning and human cognition can give us insights into how to improve AI systems in the future. It's important to keep exploring these connections.

Salience Landscaping

the shimmering void • 93 implied HN points • 08 Jun 25
📖 Philosophy Reasoning
  1. Our brains deal with a lot of information, but we need to filter and prioritize what's important. This filtering helps us focus on what's relevant in the moment.
  2. Curiosity is a natural response to uncertainty. It's like a feeling that nudges us to explore new ideas or solutions when we're unsure about something.
  3. Improving our awareness of what we care about can help us make better decisions and avoid self-deception, especially in a world filled with distractions.

Do OpenAI's New Reasoning Models (o1 Series) Differ Politically from Their Predecessors?

Rozado’s Visual Analytics • 150 implied HN points • 28 Jan 25
🕹 Technology Reasoning
  1. OpenAI's new o1 models are designed to solve problems better by thinking through their answers first. However, they are much slower and cost more to run than previous models.
  2. The political preferences of these new models are similar to earlier versions, despite the new reasoning abilities. This means they still lean left when answering political questions.
  3. Even with their advanced reasoning, these models didn't change their political views, which leads to questions about how reasoning and political bias work together in AI.

Your Emotions Aren't Your Heart of Hearts

apxhard • 51 implied HN points • 31 Jul 25
📖 Philosophy Reasoning
  1. Your emotions are not the same as your true self. It's important to understand that feelings are just a part of you, not the whole you.
  2. Finding a balance between emotion and reason is key. When we connect our thinking and feeling, we can make better choices and understand ourselves more clearly.
  3. Family should be our main focus for values, not just individual desires. Caring for our family helps guide our emotions and decisions in a way that benefits everyone.

#108 It's a process

Design Lobster • 299 implied HN points • 02 May 22
🪑 Design Reasoning
  1. The design process can sometimes feel like magic when a solution comes together, often due to abductive reasoning that brings out novel ideas.
  2. Creativity thrives in spaces outside of formal work processes, like in unscheduled moments or unconventional events like 'unconferences'.
  3. Design work is a continuous journey of developing new understandings and appreciations as you navigate through the stages, emphasizing the importance of flexibility in thinking.

What Should Count As Nutty?

David Friedman’s Substack • 242 implied HN points • 20 Jan 24
🎭️ Culture Reasoning
  1. It's not enough to have mistaken beliefs to be considered nutty.
  2. Beliefs that no reasonable person with your intellectual background could hold may qualify as nutty.
  3. Defending beliefs in a consistent, intelligent manner doesn't make someone a nut, but ignoring known facts to maintain beliefs may suggest otherwise.

Reasonableness: An Introduction

The End(s) of Argument • 39 implied HN points • 10 Jun 23
🚌 Education Reasoning
  1. Two primary accounts of the relation between evidence and belief in misinformation research are naive and non-naive models, both with limitations.
  2. People's pursuit of reasonableness influences how they collect and share evidence to support their beliefs, aiming to seem rational to others.
  3. Beliefs are often maintained through a balance of evidence and perceived reasonableness, impacting how individuals process and adopt new information.

"Mid" is unfalsifiable

Philosophy bear • 64 implied HN points • 18 May 23
📖 Philosophy Reasoning
  1. Calling someone 'mid' is unfalsifiable because it's hard to prove you're not just average.
  2. Labeling someone as 'mid' can lead to implications similar to being called 'ugly'.
  3. The concept of 'mid' has become popular due to unrealistic expectations and fear of commitment in relationships.

Understanding human stupidity in a post-truth era

Klement on Investing • 9 implied HN points • 05 Jan 24
📖 Philosophy Reasoning
  1. Human stupidity involves a temporary inability to properly reason, plan, or learn.
  2. Stubbornness often accompanies stupidity, making people hold on to disproven beliefs.
  3. In a post-truth era, combating human stupidity requires strong institutions, satire, education, and sometimes allowing people to face the consequences of their beliefs.