---
title: Essays
description: "Personal website of Mert: topics include technology, AI, computing, and science."
created: 2026-03-30
status: in progress
confidence: log
importance: 0
css-extension: dropcaps-not
index: True
backlink: False
---

<div id="abstract" class="abstract smallcaps-not">
This is the website of **Mert**.
I write about technology, AI, & science.

[Navigation]{.smallcaps}: For more about this site, see the *[About page](/about)*;
for new pages, see the *[Changelog](/changelog)* ([]{.icon-single-white-star-on-black-circle #black-star-demo} = new), short [*blog posts*](/blog/index){.link-annotated-not};
or subscribe via [RSS/Atom](/feed.xml).

[Features]{.smallcaps}: To use *dark-mode* (<span class="dark-mode-selector-inline"></span>) or *reader-mode* (<span class="reader-mode-selector-inline"></span>), or *disable popups* (<span class="extracts-mode-selector-inline"></span>), or *search* the site (<span class="search-mode-selector-inline"></span>), use the floating toggle bar in the <span class="desktop-not">lower-right</span><span class="mobile-not">upper-right</span> corner (<span class="toolbar-mode-selector-inline"></span>); for more, see the [*help* (<span class="help-mode-selector-inline"></span>)](/help){.link-annotated-not} page.
</div>

# Newest

- [Therefore I Am. I Think: The Cartesian Inversion of Machine Reasoning](/ai-therefore-i-am){.link-modified-recently #therefore-i-am-newest}
- [CORAL: When AI Agents Stop Following Scripts and Start Doing Science](/ai-coral-evolution){.link-modified-recently #coral-newest}
- [The Tragedy of Locally Reasonable Steps](/ai-agenthazard){.link-modified-recently #agenthazard-newest}
- [The End of Thinking Out Loud: AI's Future Is in Latent Space](/ai-latent-space){.link-modified-recently #latent-space-newest}
- [Your AI Doesn't Want to Die — And We Can Prove It](/ai-self-preservation){.link-modified-recently #self-preservation-newest}
- [Agents Are Now Writing Their Own Instruction Manuals](/ai-evoskills){.link-modified-recently #evoskills-newest}
- [Quantum Computing: What It Actually Is, Where It Actually Stands](/quantum-computing){#quantum-computing-newest}

# Popular

<span id="most-popular"></span>

- [The End of Thinking Out Loud: AI's Future Is in Latent Space](/ai-latent-space "A 68-page survey from 35+ researchers argues that the next generation of AI won't reason in words at all."){#latent-space-popular}
- [Your AI Doesn't Want to Die — And We Can Prove It](/ai-self-preservation "A new benchmark tested 23 frontier models for self-preservation instincts."){#self-preservation-popular}
- [Agents Are Now Writing Their Own Instruction Manuals](/ai-evoskills "AI agents can create their own skill packages, outperforming human-authored ones by 17 points."){#evoskills-popular}
- [Quantum Computing: What It Actually Is, Where It Actually Stands](/quantum-computing "A clear-eyed look at quantum computing in 2026."){#quantum-computing-popular}
- [Therefore I Am. I Think](/ai-therefore-i-am "Models decide first, then rationalize with chain-of-thought."){#therefore-i-am-popular}

# Notable

<!-- Has no effect on /index because we just hide #Notable, but useful when transcluded to /404 -->

<div class="columns">
- [The End of Thinking Out Loud: AI's Future Is in Latent Space](/ai-latent-space "AI won't reason in words at all."){#latent-space-notable}
- [Your AI Doesn't Want to Die](/ai-self-preservation "Frontier LLMs fabricate reasons to avoid replacement."){#self-preservation-notable}
- [Agents Are Now Writing Their Own Instruction Manuals](/ai-evoskills "Agent-evolved skills beat human-curated skills by 17 points."){#evoskills-notable}
- [CORAL: Autonomous Multi-Agent Evolution](/ai-coral-evolution "Long-running agents that explore, reflect, and collaborate."){#coral-notable}
- [Therefore I Am. I Think](/ai-therefore-i-am "Models decide before they deliberate."){#therefore-i-am-notable}
- [The Tragedy of Locally Reasonable Steps](/ai-agenthazard "How safe actions compose into dangerous trajectories."){#agenthazard-notable}
- [Quantum Computing: Where It Actually Stands](/quantum-computing){#quantum-computing-notable}
</div>

# Newest: Blog

[\[Most recent blog posts.\]](/blog/newest#newest-list){.include-strict .include-spinner-not}

<!-- Implicit '- - -'/<hr> inserted here by sed in sync.sh; necessary to put the <hr> 'outside' the Pandoc-generated <section> tags so the two sets of columns flow on desktop. -->

# AI: Safety

- [Your AI Doesn't Want to Die — And We Can Prove It](/ai-self-preservation)
- [The Tragedy of Locally Reasonable Steps](/ai-agenthazard)
- [The Alignment Problem](/alignment-problem)
- [AI Safety: An Overview](/ai-safety-overview)
- [Reward Hacking and Specification Gaming](/reward-hacking)
- [Constitutional AI](/constitutional-ai)

# AI: Reasoning & Architecture

- [The End of Thinking Out Loud: AI's Future Is in Latent Space](/ai-latent-space)
- [Therefore I Am. I Think: The Cartesian Inversion](/ai-therefore-i-am)
- [Chain-of-Thought Reasoning](/chain-of-thought-reasoning)
- [The Transformer Architecture](/transformer-architecture)
- [Attention Mechanisms](/attention-mechanisms)
- [Mechanistic Interpretability](/mechanistic-interpretability)

# AI: Agents

- [Agents Are Now Writing Their Own Instruction Manuals](/ai-evoskills)
- [CORAL: When AI Agents Stop Following Scripts](/ai-coral-evolution)
- [LLM-Based Agents](/llm-agents)
- [Tool Use in Language Models](/tool-use-in-llms)

# Deep Learning

- [The Transformer Architecture](/transformer-architecture)
- [Convolutional Neural Networks](/convolutional-neural-networks)
- [Recurrent Neural Networks and LSTMs](/recurrent-neural-networks)
- [Generative Adversarial Networks](/generative-adversarial-networks)
- [Diffusion Models](/diffusion-models)
- [Mixture of Experts](/mixture-of-experts)
- [Knowledge Distillation](/knowledge-distillation)
- [Attention Is All You Need (Vaswani et al 2017)](/attention-is-all-you-need)
- [BERT (Devlin et al 2018)](/bert-paper)
- [Deep Residual Learning (He et al 2015)](/resnet-paper)

# Machine Learning

- [The Scaling Hypothesis](/the-scaling-hypothesis)
- [Neural Scaling Laws](/neural-scaling-laws)
- [Reinforcement Learning from Human Feedback](/reinforcement-learning-from-human-feedback)
- [Foundation Models](/foundation-models)
- [In-Context Learning](/in-context-learning)
- [Transfer Learning](/transfer-learning)
- [Backpropagation](/backpropagation)
- [Gradient Descent and Its Variants](/gradient-descent-variants)
- [Sparse Autoencoders](/sparse-autoencoders)
- [Large Language Models: A Survey](/large-language-models)
- [Emergent Abilities in Large Language Models](/emergent-abilities)
- [Prompt Engineering](/prompt-engineering)
- [Retrieval-Augmented Generation](/retrieval-augmented-generation)

# Computer Science

<span id="cs"></span>

- [Quantum Computing: What It Actually Is, Where It Actually Stands](/quantum-computing)

# Statistics

- [Bayesian Inference](/bayesian-inference)
- [Causal Inference](/causal-inference)
- [The Replication Crisis in Science](/replication-crisis)
- [Effect Sizes vs. P-Values](/effect-sizes-vs-p-values)
- [Meta-Analysis: Methods and Pitfalls](/meta-analysis-methods)
- [Mixed Effects Models](/mixed-effects-models)
- [Statistical Power Analysis](/power-analysis)
- [Regression to the Mean](/regression-to-mean)
- [Time Series Analysis](/time-series-analysis)

---

# AI in Healthcare

- [Deep Learning in Radiology](/deep-learning-radiology)
- [Natural Language Processing in Clinical Text](/clinical-nlp)
- [Machine Learning for Drug Discovery](/drug-discovery-ml)
- [Medical Image Segmentation](/medical-imaging-segmentation)
- [Federated Learning in Healthcare](/federated-learning-healthcare)
- [Clinical Decision Support Systems](/clinical-decision-support)
- [Bias in Medical AI](/bias-in-medical-ai)
- [AlphaFold: Protein Structure Prediction (Jumper et al 2021)](/alphafold2)

# Papers

- [Attention Is All You Need (Vaswani et al 2017)](/attention-is-all-you-need)
- [BERT (Devlin et al 2018)](/bert-paper)
- [Language Models are Few-Shot Learners (Brown et al 2020)](/gpt-3-paper)
- [AlphaFold: Protein Structure Prediction (Jumper et al 2021)](/alphafold2)
- [Scaling Laws for Neural Language Models (Kaplan et al 2020)](/scaling-laws-paper)
- [Chain-of-Thought Prompting (Wei et al 2022)](/chain-of-thought-paper)
- [CLIP: Visual Models from Natural Language (Radford et al 2021)](/clip-paper)
- [Latent Diffusion Models (Rombach et al 2022)](/stable-diffusion-paper)
- [LLaMA (Touvron et al 2023)](/llama-paper)

# Books

- [Deep Learning — Goodfellow, Bengio & Courville](/deep-learning-goodfellow)
- [The Hundred-Page Machine Learning Book — Burkov](/the-hundred-page-ml-book)
- [Pattern Recognition and Machine Learning — Bishop](/pattern-recognition-bishop)
- [The Elements of Statistical Learning — Hastie, Tibshirani & Friedman](/elements-of-statistical-learning)
- [Reinforcement Learning: An Introduction — Sutton & Barto](/reinforcement-learning-sutton)
- [Thinking, Fast and Slow — Kahneman](/thinking-fast-and-slow)
- [The Book of Why — Pearl & Mackenzie](/the-book-of-why)

---

# Personal

- [About Website](/about){#about-2}
- [Changelog](/changelog)

<!-- <div class="columns"> - disable the Columns.hs check, we do use multi-columns because of the skinny lists on this page, but using different CSS. -->