The Curious Engineer

All blogs / AI / ML

Max Woolf's Blog

minimaxir.com

A blog written by data scientist Max Woolf out of San Francisco.

AI / MLPythonRust

’15’20’25
Posts
97 from 1 May 2013 to 25 Oct 2026
Posting
monthly (8 in the last year)
Kind
personal blog
Feed
https://minimaxir.com/index.xml

2026 6

  1. OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?
  2. Writing Rust code that's faster than state-of-the-art libraries by asking agents to make the code faster
  3. LLMs break down in funny ways when told the Jacobian Conjecture counterargument
  4. What's the deal with all the random weekly quota resets for agents lately?
  5. The mysterious Hy3 LLM is topping OpenRouter Model Rankings by a large margin
  6. An AI agent coding skeptic tries AI agent coding, in excessive detail

2025 9

  1. Nano Banana Pro is the best AI image generator, with caveats
  2. Nano Banana can be prompt engineered for extremely nuanced AI image generation
  3. Claude Haiku 4.5 does not appreciate my attempts to jailbreak it
  4. Can modern LLMs actually count the number of b's in "blueberry"?
  5. LLMs can now identify public figures in images
  6. Predicting Average IMDb Movie Ratings Using Text Embeddings of Movie Metadata
  7. As an Experienced LLM User, I Actually Don't Use Generative LLMs Often
  8. The Best Way to Use Text Embeddings Portably is With Parquet and Polars
  9. Can LLMs write better code if you keep asking them to “write better code”?

2024 4

  1. Generating Distinct AI Voice Performances By Prompt Engineering GPT-4o
  2. AI Seinfeld was the peak of AI-generated content. It will never happen again.
  3. The Super Effectiveness of Pokémon Embeddings Using Only Raw JSON and Images
  4. Does Offering ChatGPT a Tip Cause it to Generate Better Text? An Analysis

2023 6

  1. Pushing ChatGPT's Structured Data Support To Its Limits
  2. Please Don't Ask if an Open Source Project is Dead
  3. The Greatest Threat to Generative AI is Humans Being Bad at Using it
  4. I Made Stable Diffusion XL Smarter by Finetuning it on Bad AI-Generated Images
  5. The Problem With LangChain
  6. ChatGPT's API is So Good and Cheap, It Makes Most Text Generating AI Obsolete

2022 4

  1. Stable Diffusion 2.0 and the Importance of Negative Prompts for Good Results
  2. I Resurrected "Ugly Sonic" with Stable Diffusion Textual Inversion
  3. How to Create a Blog Post Title Optimizer with GPT-3 and Hacker News Data
  4. Absurd AI-Generated Professional Food Photography with DALL-E 2

2021 3

  1. How to Generate Customized AI Art Using VQGAN and CLIP
  2. Fun and Dystopia With AI-Based Code Generation Using GPT-J-6B
  3. Easily Transform Portraits of People into AI Aberrations Using StyleCLIP

2020 2

  1. Tempering Expectations for GPT-3 and OpenAI’s API
  2. How to Build a Twitter Text-Generating AI Bot With GPT-2

2019 3

  1. Visualizing Airline Flight Characteristics Between SFO and JFK
  2. Experiments with Making Convincing AI-Generated Fake News
  3. How To Make Custom AI-Generated Text With GPT-2

2018 7

  1. Run Any Scheduled Task/Cron Super-Cheap on Google Cloud Platform
  2. Things About Real-World Data Science Not Discussed In MOOCs and Thought Pieces
  3. Problems with Predicting Post Performance on Reddit and Other Link Aggregators
  4. Analyzing IMDb Data The Intended Way, with R and ggplot2
  5. How to Quickly Train a Text-Generating Neural Network for Free
  6. Visualizing One Million NCAA Basketball Shots
  7. A Visual Overview of Stack Overflow's Question Tags

2017 11

  1. Benchmarking Modern GPUs for Maximum Cloud Cost Efficiency in Deep Learning
  2. Making Magic: the GIFening
  3. How to Make High Quality Data Visualizations for Websites With R and ggplot2
  4. Benchmarking TensorFlow on Cloud CPUs: Cheaper Deep Learning than Cloud GPUs
  5. Predicting the Success of a Reddit Submission with Deep Learning and Keras
  6. The Decline of Imgur on Reddit and the Rise of Reddit's Native Image Hosting
  7. Advantages of Using R Notebooks For Data Analysis Instead of Jupyter Notebooks
  8. Leaving Apple Inc.
  9. Pretrained Character Embeddings for Deep Learning and Automatic Text Generation
  10. Predicting And Mapping Arrest Types in San Francisco with LightGBM, R, ggplot2
  11. Playing with 80 Million Amazon Product Review Ratings Using Apache Spark

2016 10

  1. What Percent of the Top-Voted Comments in Reddit Threads Were Also 1st Comment?
  2. Visualizing How Developers Rate Their Own Programming Skills
  3. Methods for Finding Related Reddit Subreddits with Simple Set Theory
  4. How to Create a Network Graph Visualization of Reddit Subreddits
  5. Creating Stylish, High-Quality Word Clouds Using Python and Font Awesome Icons
  6. Blockbuster Movies with Male Leads Earn More Than Those with Female Leads
  7. The Importance of Sanity-Checking Datasets Before Analysis
  8. Facebook Reactions and the Problems With Quantifying Likes Differently
  9. Video Games and Charity: Analyzing Awesome Games Done Quick 2016 Donations
  10. Movie Review Aggregator Ratings Have No Relationship with Box Office Success

2015 14

  1. Let's Code an Analysis and Visualizations of Yelp Data using R and ggplot2
  2. Mapping Where Arrests Frequently Occur in San Francisco Using Crime Data
  3. Analyzing San Francisco Crime Data to Determine When Arrests Frequently Occur
  4. How to Visualize New York City Using Taxi Location Data and ggplot2
  5. Quantifying and Visualizing the Reddit Hivemind
  6. How to Analyze Every Reddit Submission and Comment, in Seconds, for Free
  7. Coding, Visualizing, and Animating Bootstrap Resampling
  8. Plotting a Map of New York City Using Only Taxi Location Data
  9. How to Scrape Data From Facebook Page Posts for Statistical Analysis
  10. Why is the Most-Viewed Gaming Video on YouTube About Cars 2?
  11. Analyzing the Patterns of Numbers in 10 Million Passwords
  12. An Introduction on How to Make Beautiful Charts With R and ggplot2
  13. Quantifying the Clickbait and Linkbait in BuzzFeed Article Titles
  14. Locating All the Christmas Trees on Instagram

2014 13

  1. A Statistical Analysis of 142 Million Reddit Submissions
  2. The Quality, Popularity, and Negativity of 5.6 Million Hacker News Comments
  3. The Statistical Difference Between 1-Star and 5-Star Reviews on Yelp
  4. The Data From Our Comments to the FCC About Net Neutrality
  5. The Wikipedia Entries Which Are Most-Edited by Members of the U.S. Congress
  6. Impact of the New Show HN Section on Show HN Submissions
  7. Who Performs the Best in Online Classes?
  8. The Interesting Percentages of Female Students in MIT and Harvard Online Courses
  9. A Statistical Analysis of 1.2 Million Amazon Reviews
  10. Mapping San Francisco Locations Using Facebook Data
  11. Does Adding Many Tags to an Instagram Photo Maximize the Number of Likes?
  12. A Statistical Analysis of All Hacker News Submissions
  13. How Many #nofilter Instagram Photos Actually Have No Filter?

2013 5

  1. Probabilistically Generating GitHub Projects
  2. Reddit is Growing Slowly, but Surely
  3. When Uninstalling A PC Game Erases the Entire Hard Drive
  4. A Role-Playing Game With an Easily Exploitable Random Number Generator
  5. A Profanity-Laced Video Game Password That Breaks Everything