Simon Willison’s Weblog

MRSK. A new open source web application deployment tool from 37signals, developed to help migrate their Hey webmail app out of the cloud and onto their own managed hardware. The key feature is one that I care about deeply: it enables zero-downtime deploys by running all traffic through a Traefik reverse proxy in a way that allows requests to be paused while a new deployment is going out—so end users get a few seconds delay on their HTTP requests before being served by the replaced application. # 29th April 2023, 11:54 pm

MLC LLM (via) From MLC, the team that gave us Web LLM and Web Stable Diffusion. “MLC LLM is a universal solution that allows any language model to be deployed natively on a diverse set of hardware backends and native applications”. I installed their iPhone demo from TestFlight this morning and it does indeed provide an offline LLM that runs on my phone. It’s reasonably capable—the underlying model for the app is vicuna-v1-7b, a LLaMA derivative. # 29th April 2023, 5:43 pm

IF by DeepFloyd Lab (via) New image generation AI model, financially backed by StabilityAI but based on the Google Imagen paper. Claims to be much better at following complex prompts, including being able to generate text! I tried the Colab notebook with “a photograph of raccoon in the woods holding a sign that says ’I will eat your trash’” and it didn’t quite get the text right, see via link for the result. # 28th April 2023, 7:34 pm

Trainbot (via) “Trainbot watches a piece of train track, detects passing trains, and stitches together images of them”—check out the site itself too, which shows beautifully stitched panoramas of trains that have recently passed near Jo M’s apartment. Found via the best Hacker News thread I’ve seen in years, “Ask HN: Most interesting tech you built for just yourself?”. # 28th April 2023, 2:24 pm

GPT-3 token encoder and decoder. I built an Observable notebook with an interface to encode, decode and search through GPT-3 tokens, building on top of a notebook by EJ Fox and Ian Johnson. # 27th April 2023, 11:48 pm

urllib3 v2.0.0 is now generally available. urllib3 is 12 years old now, and is a common low-level dependency for packages like requests and httpx. The biggest new feature in v2 is a higher-level API: resp = urllib3.request(“GET”, “https://example.com”)—a very welcome addition to the library. # 26th April 2023, 10 pm

How prompt injection attacks hijack today’s top-end AI – and it’s really tough to fix. Thomas Claburn interviewed me about prompt injection for the Register. Lots of direct quotes from our phone call in here—we went pretty deep into why it’s such a difficult problem to address. # 26th April 2023, 6:04 pm

Rye. Armin Ronacher’s take on a Python packaging tool. There are a lot of interesting ideas in this one—it’s written in Rust, configured using pyproject.toml and has some very strong opinions, including completely hiding pip from view and insisting you use “rye add package” instead. Notably, it doesn’t use the system Python at all: instead, it downloads a pre-compiled standalone Python from Gregory Szorc’s python-build-standalone project—the same approach I used for the Datasette Desktop Electron app.

Armin warns that this is just an exploration, with no guarantees of future maintenance—and even has an issue open titled “Should Rye exist?” # 24th April 2023, 4:02 am

Introducing PyPI Organizations. Launched at PyCon US today: Organizations allow packages on the Python Package Index to be owned by a group, not an individual user account. “We’re making organizations available to community projects for free, forever, and to corporate projects for a small fee.”—this is the first revenue generating PyPI feature. # 23rd April 2023, 8:29 pm

Bard now helps you code (via) Google have enabled Bard’s code generation abilities—these were previously only available through jailbreaking. It’s pretty good—I got it to write me code to download a CSV file and insert it into a SQLite database—though when I challenged it to protect against SQL injection it hallucinated a non-existent “cursor.prepare()” method. Generated code can be exported to a Colab notebook with a click. # 21st April 2023, 3:32 pm

Stability AI Launches the First of its StableLM Suite of Language Models (via) 3B and 7B base models, with 15B and 30B are on the way. CC BY-SA-4.0. “StableLM is trained on a new experimental dataset built on The Pile, but three times larger with 1.5 trillion tokens of content. We will release details on the dataset in due course.” # 19th April 2023, 3:47 pm

Inside the secret list of websites that make AI chatbots sound smart. Washington Post story digging into the C4 dataset—Colossal Clean Crawled Corpus, a filtered version of Common Crawl that’s often used for training large language models. They include a neat interactive tool for searching a domain to see if it’s included—TIL that simonwillison.net is the 106,649th ranked site in C4 by number of tokens, 189,767 total—0.0001% of the total token volume in C4. # 19th April 2023, 1:35 pm

LLaVA: Large Language and Vision Assistant (via) Yet another multi-modal model combining a vision model (pre-trained CLIP ViT-L/14) and a LLaMA derivative model (Vicuna). The results I get from their demo are even more impressive than MiniGPT-4. Also includes a new training dataset, LLaVA-Instruct-150K, derived from GPT-4 and subject to the same warnings about the OpenAI terms of service. # 19th April 2023, 1:14 am

RedPajama, a project to create leading open-source models, starts by reproducing LLaMA training dataset of over 1.2 trillion tokens. With the amount of projects that have used LLaMA as a foundation model since its release two months ago—despite its non-commercial license—it’s clear that there is a strong desire for a fully openly licensed alternative.

RedPajama is a collaboration between Together, Ontocord.ai, ETH DS3Lab, Stanford CRFM, Hazy Research, and MILA Québec AI Institute aiming to build exactly that.

Step one is gathering the training data: the LLaMA paper described a 1.2 trillion token training set gathered from sources that included Wikipedia, Common Crawl, GitHub, arXiv, Stack Exchange and more.

RedPajama-Data-1T is an attempt at recreating that training set. It’s now available to download, as 2,084 separate multi-GB jsonl files—2.67TB total.

Even without a trained model, this is a hugely influential contribution to the world of open source LLMs. Any team looking to build their own LLaMA from scratch can now jump straight to the next stage, training the model. # 17th April 2023, 5:13 pm

Latest Twitter search results for “as an AI language model” (via) Searching for “as an AI language model” on Twitter reveals hundreds of bot accounts which are clearly being driven by GPT models and have been asked to generate content which occasionally trips the ethical guidelines trained into the OpenAI models.

If Twitter still had an affordable search API someone could do some incredible disinformation research on top of this, looking at which accounts are implicated, what kinds of things they are tweeting about, who they follow and retweet and so-on. # 17th April 2023, 2:28 pm

MiniGPT-4 (via) An incredible project with a poorly chosen name. A team from King Abdullah University of Science and Technology in Saudi Arabia combined Vicuna-13B (a model fine-tuned on top of Facebook’s LLaMA) with the BLIP-2 vision-language model to create a model that can conduct ChatGPT-style conversations around an uploaded image. The demo is very impressive, and the weights are available to download—45MB for MiniGPT-4, but you’ll need the much larger Vicuna and LLaMA weights as well. # 17th April 2023, 2:21 pm

How I Used Stable Diffusion and Dreambooth to Create A Painted Portrait of My Dog (via) I like posts like this that go into detail in terms of how much work it takes to deliberately get the kind of result you really want using generative AI tools. Jake Dahn trained a Dreambooth model from 40 photos of Queso—his photogenic Golden Retriever—using Replicate, then gathered the prompts from ten images he liked on Lexica and generated over 1,000 different candidate images, picked his favourite, used Draw Things img2img resizing to expand the image beyond the initial crop, then Automatic1111 inpainting to tweak the ears, then Real-ESRGAN 4x+ to upscale for the final print. # 16th April 2023, 7:57 pm

codespaces-jupyter (via) This is really neat. Click “Use this template” -> “Open in a codespace” and you get a full in-browser VS Code interface where you can open existing notebook files (or create new ones) and start playing with them straight away. # 14th April 2023, 10:38 pm

New prompt injection attack on ChatGPT web version. Markdown images can steal your chat data. An ingenious new prompt injection / data exfiltration vector from Roman Samoilenko, based on the observation that ChatGPT can render markdown images in a way that can exfiltrate data to the image hosting server by embedding it in the image URL. Roman uses a single pixel image for that, and combines it with a trick where copy events on a website are intercepted and prompt injection instructions are appended to the copied text, in order to trick the user into pasting the injection attack directly into ChatGPT. # 14th April 2023, 6:33 pm

Building LLM applications for production. Chip Huyen provides a useful, in-depth review of the challenges involved in taking an app built on top of a LLM from prototype to production, including issues such as prompt ambiguity and unpredictability, cost and latency concerns, challenges in testing and updating to new models. She also lists some promising use-cases she’s seeing for categories of application built on these tools. # 14th April 2023, 3:35 pm

The Great Flowering: Why OpenAI is the new AWS and the New Kingmakers still matter (via) James Governor discusses the potential impact of AI-assisted productivity on the wider software engineering industry, and calls me “a bellwether”! # 13th April 2023, 7:20 pm

GitHub Accelerator: our first cohort. I’m participating in the first cohort of GitHub’s new open source accelerator program, with Datasette (and related projects). It’s a 10 week program with 20 projects working together “with an end goal of building durable streams of funding for their work”. # 13th April 2023, 5:28 pm

Free Dolly: Introducing the World’s First Truly Open Instruction-Tuned LLM (via) Databricks released a large language model called Dolly a few weeks ago. They just released Dolly 2.0 and it is MUCH more interesting—it’s an instruction tuned 12B parameter upgrade of EleutherAI’s Pythia model. Unlike other recent instruction tuned models Databricks didn’t use a training set derived from GPT-3—instead, they recruited 5,000 employees to help put together 15,000 human-generated request/response pairs, which they have released under a Creative Commons Attribution-ShareAlike license. The model itself is a 24GB download from Hugging Face—I’ve run it slowly on a small GPU-enabled Paperspace instance, but hopefully optimized ways to run it will emerge in short order. # 13th April 2023, 2:19 am

Replacing my best friends with an LLM trained on 500,000 group chat messages (via) Izzy Miller used a 7 year long group text conversation with five friends from college to fine-tune LLaMA, such that it could simulate ongoing conversations. They started by extracting the messages from the iMessage SQLite database on their Mac, then generated a new training set from those messages and ran it using code from the Stanford Alpaca repository. This is genuinely one of the clearest explanations of the process of fine-tuning a model like this I’ve seen anywhere. # 12th April 2023, 11:01 pm

Sheepy-T—an LLM running on an iPhone. Kevin Kwok has a video on Twitter demonstrating Sheepy-T—his iPhone app which runs a full instruction-tuned large language model, based on EleutherAI’s GPT-J, entirely on an iPhone 14. I applied for the TestFlight beta and I have this running on my phone now: it works! # 11th April 2023, 5:54 pm

How we’re building a browser when it’s supposed to be impossible (via) Andreas Kling: “The ECMAScript, HTML, and CSS specifications today are (for the most part) stellar technical documents whose algorithms can be implemented with considerably less effort and guesswork than in the past.” The Ladybird project is such an inspiration, and really demonstrates the enormous value of the work put in by web standards spec authors over the last twenty years. # 11th April 2023, 10:18 am

The AI singularity is here. Can’t say I’m a fan of the headline, but the subhead “The time to figure out how to use generative AI and large language models in your code is now” is much more illustrative of the story. I’m referred to in this one as “One of the most outspoken advocates for LLM-enhanced development” which is a bit of a surprise! # 10th April 2023, 7:17 pm

AI is flooding the workplace, and workers love it. The microwave kiln pottery project I helped Natalie with gets a mention in this story about people who are putting AI tools to use. # 10th April 2023, 7:15 pm

Floor796 (via) “An ever-expanding animation scene showing the life of the 796th floor of the huge space station” by Russian artist 0x00, who built their own custom browser-based pixel animation tool with which they are constructing this project. Absolutely crammed with pop culture references and easter eggs. The “Changes” link at the top shows almost daily updates, with links to jump to the latest content. # 10th April 2023, 4:59 pm

On Endings: Why & How We Retired Elm at Culture Amp (via) Culture Amp made extensive use of Elm—a ML-like functional language that compiles to JavaScript—between 2016 and 2020 while building their company’s frontend. They eventually decided to move away from it, for reasons described at length in this post primarily relating to its integration with React. This piece is worth reading mainly as a thoughtful approach to engineering management challenge of deprecating a well-loved piece of technology from the recommended stack at a company. # 10th April 2023, 2:11 am