Sitemap - 2023 - Matt Rickard
Pseudonyms in American History
Standard Causes of Human Misjudgment (Munger)
The Unreasonable Effectiveness of Monte Carlo
Drawbacks of Moving to the Edge
Duties of the Board of Directors
Take Your Time Making Decisions
The Encyclopedia of Integer Sequences
What if Google Wasn’t The Default?
Copilot is an Incumbent Business Model
The Context Length Observation
To be, or not to be; ay, there’s the point.
Static Sites Aren't Simple Anymore
Why Model Evaluation is Difficult
Regulatory Capture in the Railroad Industry
Infrastructure as Code Will be Written by AI
Positioning Yourself Near The Opportunity
Horizontal Tuning: Instruction, Chat, and What Else?
Retrieval Augmented Generation
Pitfalls of File-based Routing
Can OpenAI Win Consumer and Enterprise?
Revision: Generative Text-to-UI
Why Is the Frontend Stack So Complicated?
Incentives Behind Programming Languages
The Benefit of a Personal Notation
Pitfalls of Database Migrations
The Compression / Learning Duality
Passkeys, Crypto, and Signing AI Content
Multi-Modal AI is a UX Problem
The Fundamental Attribution Error
Customized End-User Software (with AI)
Five Principles from Renaissance Technologies
The Age-old Resistance to Generated Code
The New Economics of Generating Code
The Inevitability of Scope Creep
Fine-tuning Stable Diffusion XL with Personal Photos
Raft: The Distributed Systems Algorithm
Type Constraints for LLM Output
Technical Metrics to Track in Engineering Orgs
The Contrarian Strategy of OpenAI
The Free Lunch of Model Distillation
A Model API Gateway for 20+ LLMs
The Truth About GPU Utilization
Directly Responsible Individuals
Akin's Laws of Spacecraft Design
Dark Mode and Marginal Benefit
Deterministic, Structured LLM Completions
Prescriptive vs. Descriptive Feedback
Focus On What Could Go Right (Not Wrong)
Automatic and Universal Adversarial Prompts
Minimum Effective Dose and Discrete Outcomes
Git Merge Strategies and Algorithms
Virtual Machines and Path Dependence
Why Did Meta Open-Source Llama 2?
Literate Programming with LLMs
Schopenhauer's 36 Stratagems for Winning Arguments
Scale to Zero for AI Workloads
Hard Launches vs. Soft Launches
Virtualized Linux on macOS Internals
Categorization and Classification with LLMs
ActivityPub, Decentralization - Do Users Care?
Customer Obsession is the Only Winning Strategy
Where AI Fits in Engineering Organizations
Google Search's Death by a Thousand Cuts
Monopolizing Useless Resources
Platform Engineering vs. DevOps
Model Evaluation is (Still) An Art
No Feature is Free (Especially AI Ones)
Mixture of Experts: Is GPT-4 Just Eight Smaller Models?
The Low-Background Steel of AI
Why Does Every AI Cycle Start With Chat?
A Token Efficient Language for LLMs
Keep The Main Thing The Main Thing
It’s Too Early To Call Winners in AI
The Myth of The AI Infrastructure Phase
ChatGPT Plugins Don't Have PMF
Levels of Autonomy in AI Agents
The Problems with Tokenization in LLMs
What Diffusion Models Can Teach Us About LLMs
Sequence and Version Control Models
Practical Dependency Management for Developers
Faster Horses: AI Products That Companies Think They Want
Prompt Engineering is Configuration Engineering
A List of Things I Was Wrong About
The ChatGPT Plugin Specification
The New Tax on Engineering Against the Grain
A List of Leaked System Prompts
React LLM: Easily Run Models in the Browser with Headless Components
Context-Free Grammar Parsing With LLMs
Taylor Swift and Launch Cadence
Self-hosted Compilers and Bootstrapped AI
Why Declarative Build Systems Aren’t Popular
ReLLM: Exact Structure for Large Language Model Completions
The Python Family of Languages
Applications of Category Theory
Implementing LLMs in the Browser
Principle of Least Astonishment
Probabilistic Data Structures and LLMs
16 Lessons from the Tanenbaum–Torvalds Debates
Tanenbaum–Torvalds Debates, Part 1
Building A Kernel From Scratch
Autonomous LLM Agents Are At Least 10 Years Out
The Peanut Butter Manifesto (2006)
Foundational Models Are Not Enough
Imperative, Declarative, Interrogative, and Exclamative Interfaces
A List of 1 Billion+ Parameter LLMs
No GPUs Before Product Market Fit
The Path Dependence of YAML Templates
Buyers in the Foundational Model Stack
A High-level LLMOps Architecture
Consumption Pricing Units in AI
Reflections on 10,000 Hours of DevOps
A Hacker's Guide to LLM Optimization
Code, not Chat, in Generative AI
Necessary Conditions for an App Store Monopoly
Are Incumbents Accruing All The AI Value?
Modeling Context Length vs. Information Retrieval Cost in LLMs
Framework-Defined Infrastructure
The Value of Software Generalists
Foundational Models Are Commodities
On OpenAI's Kubernetes Cluster
Automatic1111 and AI Aggregators
Chain of Thought Paradigms in LLMs
Applications, Services, and Workflows Overload
Will The AI Stack Be Open Or Closed?
The Bundling of the Software Developer
Commoditization of Large Language Models: Part 3
Why Python Won't Be the Language of LLMs
Should OSS Projects Have Telemetry?
If It Doesn’t Ship, It Doesn’t Exist
Future is Not Evenly Distributed
Zero Interest Rate Policy Phenomena
Type-Safe API Calls: tRPC vs. gRPC
Cloud Services Ranked: Build vs. Buy
Prompt Engineering Shouldn't Exist
All Models Are Wrong (But Some Are Useful)
The Problem with Drag-n-Drop Infrastructure
Lessons from the Last AI Cycle