Newsletters: 2024
-
· Substack
Things we learned about LLMs in 2024
Key themes and pivotal moments from the last 12 months in Large Language Models
-
· Substack
QvQ—Qwen’s new visual reasoning model
Plus DeepSeek v3, an open weights model that rivals Claude 3.5 Sonnet
-
· Substack
December in LLMs has been a lot
Plus notes on OpenAI's o3, and my approach to running a link blog
-
· Substack
Gemini 2.0 Flash “Thinking Mode”
Plus building Python tools with a one-shot prompt using uv run and Claude Projects
-
· Substack
Gemini 2.0 Flash: An outstanding multi-modal LLM with a sci-fi streaming mode
Plus ChatGPT Canvas can make API requests now, but it's complicated
-
· Substack
I can now run a GPT-4 class model on my laptop with Llama 3.3 70B
And an async browser library for alert(), confirm() and prompt() built using OpenAI o1
-
-
· Substack
Ask questions of SQLite databases and CSV/JSON files in your terminal
Plus Storing times for human events and more
-
· Substack
Civic Band—scraping and searching PDF meeting minutes from hundreds of municipalities
Plus notes from Bing Chat—Our First Encounter With Manipulative AI
-
· Substack
Qwen2.5-Coder-32B is an LLM that can code well that runs on my Mac
Plus Visualizing local election results with Datasette, Observable and MapLibre GL
-
· Substack
VERDAD—tracking misinformation in radio broadcasts using Gemini 1.5
Plus generating documentation from unit tests using LLM and files-to-prompt
-
-
· Substack
Run prompts against images, audio and video in your terminal using LLM
Plus generate and execute jq programs from a prompt with llm-jq
-
· Substack
The new Claude analysis JavaScript code execution tool
Plus Claude 3.5 Opus has been delayed or cancelled, and some uv tips
-
· Substack
Anthropic’s new Computer Use capability
And a new paper about obfuscated prompt injection attacks
-
· Substack
Everything I built with Claude Artifacts this week
And unning Llama 3.2 Vision and Phi-3.5 Vision on a Mac with mistral.rs
-
· Substack
Video scraping using Google Gemini
Plus new OpenAI Audio APIs, and ChatGPT will happily write you a thinly disguised horoscope
-
· Substack
Gemini 1.5 Flash-8B, FLUX 1.1 Python 3.13...
... and Deno 2, and 34 links, 6 quotations and 4 TILs
-
· Substack
OpenAI DevDay: Let’s build developer tools, not digital God
News from OpenAI’s developer platform event
-
· Substack
NotebookLM’s automatically generated podcasts are surprisingly effective
Plus DjangoCon US 2024, and how to run Qwen-2 VL on a Mac
-
· Substack
Llama 3.2 and plugins for Django
Plus things I’ve learned on the board of the Python Software Foundation
-
· Substack
OpenAI’s new o1 chain-of-thought models
Plus Teresa T the whale, Pixtral from Mistral, podcast notes and more
-
· Substack
Building a tool showing how Gemini Pro can return bounding boxes for objects in images
Plus creating alt text for a bot using GPT-4o, and converting a PDF paper to semantic HTML with Gemini 1.5 Pro
-
· Substack
Claude’s API now supports CORS requests, enabling client-side applications
Plus Optimizing Datasette and two weeks worth of new links and quotes
-
· Substack
django-http-debug, mostly written by Claude
Plus the latest in the ongoing LLM pricing war
-
· Substack
Datasette 1.0a14: The annotated release notes
Plus notes on how Apple and NVIDIA have been training their latest models
-
· Substack
Llama 3.1, now available in LLM
The best available openly licensed model is now competitive with GPT-4o and Claude 3.5 Sonnet
-
· Substack
GPT-4o mini, LLM 0.15, sqlite-utils 3.37 and building a staging environment
Plus Mistral NeMo, Codestral Mamba and Llama-3-Groq-Tool-Use Models
-
· Substack
Imitation Intelligence keynote at PyCon 2024
And: Give people something to link to so they can talk about your features and ideas
-
· Substack
Open challenges for AI engineering
Plus a flurry of tiny tools built using Claude 3.5 Sonnet
-
· Substack
Building search-based RAG using Claude 3.5 Sonnet, Datasette and Val Town
The new Claude 3.5 Sonnet is now the best available LLM
-
· Substack
Language models on the command-line
Plus Datasette Studio, prompt injection against GitHub Copilot Chat and more
-
· Substack
Thoughts on the WWDC 2024 keynote on Apple Intelligence
Plus Claude's personality, Qwen 2 model censorship, OpenAI Voice Engine and more
-
· Substack
Accidental prompt injection against RAG applications
Plus notes on Microsoft Recall and more
-
· Substack
Training is not the same as chatting: ChatGPT and other LLMs don’t remember everything you say
Plus notes from PyCon and LLM support for GPT-4o and Gemini Flash
-
· Substack
ChatGPT in 4o mode doesn’t have the new voice and image features yet
Plus Gemini Flash and PaliGemma from Google I/O
-
· Substack
GPT-4o, a new version of LLM and more thoughts on slop
Notes on GPT-4o, OpenAI's latest model release
-
· Substack
LLM slop, datasette-secrets, llm-evals, gpt2-chatbot and a whole lot more
Plus 28 links and 19 quotes from the past two weeks
-
· Substack
Options for accessing Llama 3 from the terminal using LLM
Run Llama 3 on your laptop or access it using a number of different API providers
-
· Substack
AI for Data Journalism: demonstrating what we can do with this stuff right now
Plus news on Mistral, Reka, Claude 3 and more
-
· Substack
Three major LLM releases in 24 hours
Google Gemini Pro 1.5 is free, GPT-4 Turbo has Vision, Mixtral 8x22B released in a tweet
-
· Substack
Building files-to-prompt entirely using Claude 3 Opus
Plus "llm cmd" and running OCR against PDFs and images directly in your browser
-
· Substack
Building and testing C extensions for SQLite with ChatGPT Code Interpreter
A much more advanced Code Interpreter exercise
-
-
· Substack
The GPT-4 barrier has finally been smashed
Prompt injection and jailbreaking are not the same thing
-
-
· Substack
The killer app of Gemini Pro 1.5 is video
Plus weeknotes and a whole bunch of links, quotations and TILs
-
· Substack
Datasette 1.0a8: JavaScript plugins, new plugin hooks and plugin configuration in datasette.yaml
Plus 26 links, 7 quotations and 1 TIL
-
· Substack
LLM 0.13: The annotated release notes
And was that George Carlin "AI special" actually written by AI?
-
· Substack
Talking about Open Source LLMs on Oxide and Friends
Plus Tom Scott, and the formidable power of escalating streaks
-
· Substack
Stuff we figured out about AI in 2023
Plus recommendations to limit the blast radius for prompt injection
