Changkun's Blog

Half a year after this experiment, I keep seeing it play out with founders I’ve spoken with across domains, and the original didn’t say one thing about it.

The ultimate product forms of artificial intelligence can be summarized as the following: Virtual employees: take on clearly defined roles, execute tasks continuously, and report to supervisors.

Sometimes, when you want to do something, someone always questions your background, experience, and ability.

Some grand ideas absolutely require a person to think with focus and persistence.

When you can see the full shape of the mountain, the mountain has already moved far away from you; when you can see the whole picture of an event, you are no longer inside it.

I wrote a book: Bayesian Optimization: From First Principles to Human Preferences.

When I left Sixt, I sent a farewell letter to the people I worked with.

“Perhaps it is more important to trust the people who wrote the software.”

Recently, by organizing my life experiences, I’ve deduced the following “axioms” applicable to my current stage: Axiom One (Accessibility Preference): The value of a position equals its actual returns plus the right to “leave anytime,” where the weight of…

Tastes are accumulated from experience.

This interview is very timely, and some of the insights within it are quite thought-provoking, particularly regarding the development of general-purpose technology.

In mathematics, we can develop the intuition that two systems can prove exactly the same theorems yet differ in how quickly things get discovered in them.

We should never forget that the strongest aspect of human intelligence is our generalization and sample efficiency.

Agent economics is fundamentally about finding the right “autonomy affordance” in an organization.

On instances, primitives, and the difference between novelty that compounds and novelty that merely accumulates.

Among a set of formally symmetric options, what determines actual preferences and stability?

The emergence of AI will not push the world toward a single unified system.

Is “secure by default” the right default for AI products?

For a long time, I was a happy Cursor user.

This is an interesting article.

Can AI agents ever produce artifacts you can fully trust?

Multi-agent Topology Dynamic Management: Thinking Notes Discussion Content This discussion revolved around a central question: in a system composed of multiple agents, who should decide the topology structure (who spawns whom, who merges with whom, who is…

0.

What if your AI agents are lying to each other?

Today I read some articles on Harnessing engineering and had a sudden thought: Harnessing engineering is essentially control theory, right?

I re-read the “Free Energy Principle” and came across an interesting idea: humans don’t live in order to predict the world; rather, humans live because they predict the world.

“It is not knowledge, but the act of learning, not possession but the act of getting there, which grants the greatest enjoyment.” – Gauss (letter to Bolyai, 1808) Give an AI Agent a clean computer, set no goals, and let it decide what to do.

In this current era, this article becomes even more interesting.

Recently I developed an automated software development pipeline, and the basic idea is: an idea agent proposes ideas, then the pipeline automatically implements them, runs tests after implementation, and then automatically submits.

This discussion did not begin with philosophy of consciousness.

I’ve been thinking recently about what this current wave of AI development will ultimately become, and what societal structures will transform into as a result.

Recently I have been thinking about using AI to develop AI itself.

I suddenly realized that the concepts of the “Wallfacers” and the transparent / opaque thinking of the Trisolarans in Liu Cixin’s The Three-Body Problem become particularly interesting in the era of LLMs.

Some interesting tradeoffs: Autonomy vs.

Last Friday, while pair coding with colleagues, I noticed an interesting pattern.

psychology solved the ai memory problem decades ago. we just haven’t been reading the right papers. your identity isn’t something you have. it’s something you construct. constantly. from autobiographical memory, emotional experience, and narrative coheren…

Mathematics, when self-reference appears in its formal system, generates paradoxes.

https://arxiv.org/abs/2602.11988 The following content is generated by LLMs and may contain inaccuracies.

In the blink of an eye, a decade has passed since my undergraduate graduation.

The Gibbard–Satterthwaite theorem states that every ranked-choice voting system is manipulable, except possibly in two cases: if there is a distinguished voter with dictatorial power, or if the rule limits possible outcomes to only two options. https://en…

Definition of Preference While some view the term “preference” as a rather subjective concept, this does not prevent various disciplines from conducting in-depth research on it.

Stanford Encyclopedia of Philosophy https://plato.stanford.edu/contents.html The following content is generated by LLMs and may contain inaccuracies.

In our everyday lives, we confront numerous moral issues.

How can you believe in something you haven’t witnessed with your own eyes?

LLM Emergence Emergent Abilities of Large Language Models https://arxiv.org/abs/2206.07682 There is no such thing as conscious artificial intelligence https://www.nature.com/articles/s41599-025-05868-8 Provable Scaling Laws of Feature Emergence from Learn…

MLHP/RLHF Lectures https://web.stanford.edu/class/cs329h/index.html https://mlhp.stanford.edu/ Multi-Objective Bayesian Optimization with Active Preference Learning https://arxiv.org/abs/2311.13460 Nearly Optimal Active Preference Learning and Its Applica…

Per-tool-call human approval in agentic AI is solved in theory, unsolved in practice.

The Voter Model The voter model is a simple mathematical model of opinion formation in which voters are located at the nodes of a network.

LLM as Optimizer: Large Language Models as Optimizers https://arxiv.org/abs/2309.03409 When Large Language Models Meet Optimization https://www.sciencedirect.com/science/article/abs/pii/S2210650224002013?

LLM Lectures https://cmu-llms.org/schedule/ https://www.phontron.com/class/lminference-fall2025/schedule https://llmsystem.github.io/llmsystem2026spring/docs/Syllabus https://llmsystem.github.io/llmsystem2025spring/docs/Syllabus/ The following content is…