June 2024
June 18, 2024
Tags with descriptions. Tiny new feature on my blog: I can now add optional descriptions to my tag pages, for example on datasette and sqlite-utils and prompt-injection.
I built this feature on a live call this morning as an unplanned demonstration of GitHub's new Copilot Workspace feature, where you can run a prompt against a repository and have it plan, implement and file a pull request implementing a change to the code.
My prompt was:
Add a feature that lets me add a description to my tag pages, stored in the database table for tags and visible on the /tags/x/ page at the top
It wasn't as compelling a demo as I expected: Copilot Workspace currently has to stream an entire copy of each file it modifies, which can take a long time if your codebase includes several large files that need to be changed.
It did create a working implementation on its first try, though I had given it an extra tip not to forget the database migration. I ended up making a bunch of changes myself before I shipped it, listed in the pull request.
I've been using Copilot Workspace quite a bit recently as a code explanation tool - I'll prompt it to e.g. "add architecture documentation to the README" on a random repository not owned by me, then read its initial plan to see what it's figured out without going all the way through to the implementation and PR phases. Example in this tweet where I figured out the rough design of the Jina AI Reader API for this post.
June 19, 2024
I’ve stopped using box plots. Should you? (via) Nick Desbarats explains box plots (including with this excellent short YouTube video) and then discusses why he thinks "typically less than 20 percent" of participants in his workshops already understand how to read them.
A key problem is that they are unintuitive: a box plot has four sections, two thin lines (the top and bottom whisker segments) and two larger boxes, joined around the median. Each of these elements represents the same number of samples (one quartile each) but the thin lines v.s. thick boxes imply that the whiskers contain less samples than the boxes.
About the Lawrence Times (via) The town of Lawrence, Kansas is where Django was born. I'm delighted to learn that it has a new independent online news publication as-of March 2021 - the Lawrence Times.
It's always exciting to see local media startups like this one, and they've been publishing for three years now supported by both advertiser revenue and optional paid subscriptions.
Weeknotes: Datasette Studio and a whole lot of blogging
I’m still spinning back up after my trip back to the UK, so actual time spent building things has been less than I’d like. I presented an hour long workshop on command-line LLM usage, wrote five full blog entries (since my last weeknotes) and I’ve also been leaning more into short-form link blogging—a lot more prominent on this site now since my homepage redesign last week.
[... 736 words]Civic Band. Exciting new civic tech project from Philip James: 30 (and counting) Datasette instances serving full-text search enabled collections of OCRd meeting minutes for different civic governments. Includes 20,000 pages for Alameda, 17,000 for Pittsburgh, 3,567 for Baltimore and an enormous 117,000 for Maui County.
Philip includes some notes on how they're doing it. They gather PDF minute notes from anywhere that provides API access to them, then run local Tesseract for OCR (the cost of cloud-based OCR proving prohibitive given the volume of data). The collection is then deployed to a single VPS running multiple instances of Datasette via Caddy, one instance for each of the covered regions.
June 20, 2024
State-of-the-art music scanning by Soundslice. It's been a while since I checked in on Soundslice, Adrian Holovaty's beautiful web application focused on music education.
The latest feature is spectacular. The Soundslice music editor - already one of the most impressive web applications I've ever experienced - can now import notation directly from scans or photos of sheet music.
The attention to detail is immaculate. The custom machine learning model can handle a wide variety of notation details, and the system asks the user to verify or correct details that it couldn't perfectly determine using a neatly designed flow.
Free accounts can scan two single page documents a month, and paid plans get a much higher allowance. I tried it out just now on a low resolution image I found on Wikipedia and it did a fantastic job, even allowing me to listen to a simulated piano rendition of the music once it had finished processing.
It's worth spending some time with the release notes for the feature to appreciate how much work they've out into improving it since the initial release.
If you're new to Soundslice, here's an example of their core player interface which syncs the display of music notation to an accompanying video.
Adrian wrote up some detailed notes on the machine learning behind the feature when they first launched it in beta back in November 2022.
OMR [Optical Music Recognition] is an inherently hard problem, significantly more difficult than text OCR. For one, music symbols have complex spatial relationships, and mistakes have a tendency to cascade. A single misdetected key signature might result in multiple incorrect note pitches. And there’s a wide diversity of symbols, each with its own behavior and semantics — meaning the problems and subproblems aren’t just hard, there are many of them.
[...] And then some absolute son of a bitch created ChatGPT, and now look at us. Look at us, resplendent in our pauper's robes, stitched from corpulent greed and breathless credulity, spending half of the planet's engineering efforts to add chatbot support to every application under the sun when half of the industry hasn't worked out how to test database backups regularly.
Claude 3.5 Sonnet. Anthropic released a new model this morning, and I think it's likely now the single best available LLM. Claude 3 Opus was already mostly on-par with GPT-4o, and the new 3.5 Sonnet scores higher than Opus on almost all of Anthropic's internal evals.
It's also twice the speed and one fifth of the price of Opus (it's the same price as the previous Claude 3 Sonnet). To compare:
- gpt-4o: $5/million input tokens and $15/million output
- Claude 3.5 Sonnet: $3/million input, $15/million output
- Claude 3 Opus: $15/million input, $75/million output
Similar to Claude 3 Haiku then, which both under-cuts and out-performs OpenAI's GPT-3.5 model.
In addition to the new model, Anthropic also added a "artifacts" feature to their Claude web interface. The most exciting part of this is that any of the Claude models can now build and then render web pages and SPAs, directly in the Claude interface.
This means you can prompt them to e.g. "Build me a web app that teaches me about mandelbrot fractals, with interactive widgets" and they'll do exactly that - I tried that prompt on Claude 3.5 Sonnet earlier and the results were spectacular (video demo).
An unsurprising note at the end of the post:
To complete the Claude 3.5 model family, we’ll be releasing Claude 3.5 Haiku and Claude 3.5 Opus later this year.
If the pricing stays consistent with Claude 3, Claude 3.5 Haiku is going to be a very exciting model indeed.
One of the core constitutional principles that guides our AI model development is privacy. We do not train our generative models on user-submitted data unless a user gives us explicit permission to do so. To date we have not used any customer or user-submitted data to train our generative models.
llm-claude-3 0.4. LLM plugin release adding support for the new Claude 3.5 Sonnet model:
pipx install llm
llm install -U llm-claude-3
llm keys set claude
# paste AP| key here
llm -m claude-3.5-sonnet \
'a joke about a pelican and a walrus having lunch'
June 21, 2024
It is in the public good to have AI produce quality and credible (if ‘hallucinations’ can be overcome) output. It is in the public good that there be the creation of original quality, credible, and artistic content. It is not in the public good if quality, credible content is excluded from AI training and output OR if quality, credible content is not created.
Val Vibes: Semantic search in Val Town. A neat case-study by JP Posma on how Val Town's developers can use Val Town Vals to build prototypes of new features that later make it into Val Town core.
This one explores building out semantic search against Vals using OpenAI embeddings and the PostgreSQL pgvector extension.
OpenAI was founded to build artificial general intelligence safely, free of outside commercial pressures. And now every once in a while it shoots out a new AI firm whose mission is to build artificial general intelligence safely, free of the commercial pressures at OpenAI.
Building search-based RAG using Claude, Datasette and Val Town
Retrieval Augmented Generation (RAG) is a technique for adding extra “knowledge” to systems built on LLMs, allowing them to answer questions against custom information not included in their training data. A common way to implement this is to take a question from a user, translate that into a set of search queries, run those against a search engine and then feed the results back into the LLM to generate an answer.
[... 3,372 words]Datasette 0.64.8. A very small Datasette release, fixing a minor potential security issue where the name of missing databases or tables was reflected on the 404 page in a way that could allow an attacker to present arbitrary text to a user who followed a link. Not an XSS attack (no code could be executed) but still a potential vector for confusing messages.
June 22, 2024
Wikipedia Manual of Style: Linking (via) I started a conversation on Mastodon about the grammar of linking: how to decide where in a phrase an inline link should be placed.
Lots of great (and varied) replies there. The most comprehensive style guide I've seen so far is this one from Wikipedia, via Tom Morris.
In our “who validates the validators” user studies, we found that people expected—and also desired—for the LLM to learn from any human interaction. That too, “as efficiently as possible” (ie after 1-2 demonstrations, the LLM should “get it”)
June 23, 2024
The people who are most confident AI can replace writers are the ones who think writing is typing.
llama.ttf (via) llama.ttf is "a font file which is also a large language model and an inference engine for that model".
You can see it kick into action at 8m28s in this video, where creator Søren Fuglede Jørgensen types "Once upon a time" followed by dozens of exclamation marks, and those exclamation marks then switch out to render a continuation of the story. But... when they paste the code out of the editor again it shows as the original exclamation marks were preserved - the LLM output was presented only in the way they were rendered.
The key trick here is that the font renderer library HarfBuzz (used by Firefox, Chrome, Android, GNOME and more) added a new WebAssembly extension in version 8.0 last year, which is powerful enough to run a full LLM based on the tinyllama-15M model - which fits in a 60MB font file.
(Here's a related demo from Valdemar Erk showing Tetris running in a WASM font, at 22m56s in this video.)
The source code for llama.ttf is available on GitHub.
For some reason, many people still believe that browsers need to include non-standard hacks in HTML parsing to display the web correctly.
In reality, the HTML parsing spec is exhaustively detailed. If you implement it as described, you will have a web-compatible parser.
June 24, 2024
Microfeatures I Love in Blogs and Personal Websites (via) This post by Daniel Fedorin (and the accompanying Hacker News thread) is a nice reminder of one of the most fun things about building your own personal website: it gives you a low-risk place to experiment with details like footnotes, tables of contents, linkable headings, code blocks, RSS feeds, link previews and more.
New blog feature: Support for markdown in quotations. Another incremental improvement to my blog. I've been collecting quotations here since 2006 - I now render them using Markdown (previously they were just plain text). Here's one example. The full set of 920 (and counting) quotations can be explored using this search filter.
What Apple unveiled last week with Apple Intelligence wasn't so much new products, but new features—a slew of them—for existing products, powered by generative AI.
[...] These aren't new apps or new products. They're the most used, most important apps Apple makes, the core apps that define the Apple platforms ecosystem, and Apple is using generative AI to make them better and more useful—without, in any way, rendering them unfamiliar.
Listen to the AI-generated ripoff songs that got Udio and Suno sued. Jason Koebler reports on the lawsuit filed today by the RIAA against Udio and Suno, the two leading generative music startups.
The lawsuit includes examples of prompts that the record labels used to recreate famous songs that were almost certainly included in the (undisclosed) training data. Jason collected some of these together into a three minute video, and the result in pretty damning. Arguing "fair use" isn't going to be easy here.
June 25, 2024
Claude Projects. New Claude feature, quietly launched this morning for Claude Pro users. Looks like their version of OpenAI's GPTs, designed to take advantage of Claude's 200,000 token context limit:
You can upload relevant documents, text, code, or other files to a project’s knowledge base, which Claude will use to better understand the context and background for your individual chats within that project. Each project includes a 200K context window, the equivalent of a 500-page book, so users can add all of the insights needed to enhance Claude’s effectiveness.
You can also set custom instructions, which presumably get added to the system prompt.
I tried dropping in all of Datasette's existing documentation - 693KB of .rst files (which I had to rename to .rst.txt for it to let me upload them) - and it worked and showed "63% of knowledge size used".
This is a slightly different approach from OpenAI, where the GPT knowledge feature supports attaching up to 20 files each with up to 2 million tokens, which get ingested into a vector database (likely Qdrant) and used for RAG.
It looks like Claude instead handle a smaller amount of extra knowledge but paste the whole thing into the context window, which avoids some of the weirdness around semantic search chunking but greatly limits the size of the data.
My big frustration with the knowledge feature in GPTs remains the lack of documentation on what it's actually doing under the hood. Without that it's difficult to make informed decisions about how to use it - with Claude Projects I can at least develop a robust understanding of what the tool is doing for me and how best to put it to work.
No equivalent (yet) for the GPT actions feature where you can grant GPTs the ability to make API calls out to external systems.
Polyfill supply chain attack hits 100K+ sites
(via)
Short version: if you are loading assets from the polyfill.io domain you need to remove that right now: the new owners of the domain (as of a few months ago) appear to be using it to serve malicious JavaScript.
polyfill.io was a fascinating service. It was originally developed and supported by the Financial Times, but span off as a separate project several years ago.
The key idea was to serve up a set of JavaScript polyfills - pieces of code that implemented missing web platform features for older browsers - dynamically, based on the incoming user-agent. This required a CDN that varied its output dynamically based on the user-agent, hence the popularity of the single hosted service.
Andrew Betts, the original author of the service, has been warning people to move off it since February 2024:
If your website uses
polyfill.io, remove it IMMEDIATELY.I created the polyfill service project but I have never owned the domain name and I have had no influence over its sale.
He now works for Fastly, which started offering a free polyfill-fastly.io alternative in February. Andrew says you probably don't need that either, given that modern browsers have much better compatibility than when the service was first introduced over a decade ago.
There's some interesting additional context in a now-deleted GitHub issue, preserved here by the Internet Archive.
Usually one answer to protecting against this style of CDN supply chain attack would be to use SRI hashes to ensure only the expected script can be served from the site. That doesn't work here because the whole point of the service is to serve different scripts to different browsers.
June 26, 2024
picopilot (via) Kyle Carberry's "GitHub Copilot in 70 lines of JavaScript". The title is a little hyperbolic, but the code itself really does implement an OpenAI powered Visual Studio Code text completion extension in 71 lines of code. This is an excellent example for learning what a minimal VS Code extension looks like.
Here's the system prompt it uses:
You provide code completion results given a prefix and suffix. Respond with a JSON object with the key 'completion' containing a suggestion to place between the prefix and suffix. Follow existing code styles. Listen to comments at the end of the prefix. The language is "{language}".
Then it passes the prefix and suffix as two user messages, and uses the "response_format": {"type": "json_object"} option to enforce JSON output from the GPT-4o API.
The feature this is missing is the thing that makes GitHub Copilot so impressive: Copilot does a whole bunch of clever tricks to find snippets of relevant code from the current and other nearby files and includes them with the prompt, resulting in much higher quality completions.
June 27, 2024
Open challenges for AI engineering
I gave the opening keynote at the AI Engineer World’s Fair yesterday. I was a late addition to the schedule: OpenAI pulled out of their slot at the last minute, and I was invited to put together a 20 minute talk with just under 24 hours notice!
[... 5,640 words]Python 3.12 change results in Apple App Store rejection
(via)
Such a frustrating demonstration of the very worst of Apple's opaque App Store review process. The Python 3.12 standard library urllib package includes the string itms-services, and after much investigation Eric Froemling managed to determine that Apple use a scanner and reject any app that has that string mentioned anywhere within their bundle.
Russell Keith-Magee has a thread on the Python forum discussing solutions. He doesn't think attempts to collaborate with Apple are likely to help:
That definitely sounds appealing as an approach - but in this case, it’s going to be screaming into the void. There’s barely even an appeals process for app rejection on Apple’s App Store. We definitely don’t have any sort of channel to raise a complaint that we could reasonably believe would result in a change of policy.
June 28, 2024
Django: Test for pending migrations
(via)
Neat recipe from Adam Johnson for adding an automated test to your Django test suite that runs manage.py makemigrations --check to ensure you don't accidentally land code that deploys with a missing migration and crashes your site. I've made this mistake before myself so I'll be adding this to my projects.