Ephes Blog

I use keming as an adjective to describe bad kerning --Xe

Worked a little bit on the upcoming pytest-course. My wife made a website for Thomy Saurk which was going live last week. Then there was a new django-cast release. It provides some fixes for the podcast_audio field (if it's not set there's no twitter-card and the episode is not showing up in the feed) and it's now possible to add custom templates for error views (for 404, 500, 403 etc HTTP status codes) if you create a theme for django-cast.

And after two or three years of using PyCharm or vs code for writing Python almost exclusively, I started to use vim and neovim again and spent already a significant amount of time on their configuration 😅. But I missed using vim for sure.

Articles

• Set Studio: respecting the fabric of the internet | Case study
• Top 10 Django Third-Party Packages | No surprises here...
• PostgreSQL® JSONB Cheatsheet: Complete & Fast Lookup Guide
• An example of LLM prompting for programming
• Science and Technology links (April 11 2023)
• Deploying key transparency at WhatsApp
• We must slow down the race to God-like AI | Despite the clickbait title, this article is actually good
• Running Tasks Concurrently in Django Asynchronous Views
• Reduce - The Power of a Single Python Function
• Prompt injection: what’s the worst that can happen?
• Color Theme Switcher

Video

• Sparks of AGI: early experiments with GPT-4 | I wouldn't call it AI leave alone AGI, but there's something, yep
• Modern Web Development: Centering DIVs in new exciting wrong ways with AI!

Books

• Natural Language Processing with Transformers Book

Weeknotes

• Weeklog for Week 14: April 03 to April 09 | Johannes
• April Week 14 | Bithive
• Week in Review 2023-04-10 | Luis

Mastodon / Twitter

• Gene Hackman would be the perfect name for a CRISPR engineer.
• I remember Max Keiser calling this the most revolutionary graph he'd ever seen
• Fantastic #TomLehrer content in today's FT

Software

• htmx release 1.9.0
• httpx release 0.24.0
• Neapolitan - easy crud views for Django
• List of LLaMA models | Lots of interesting links...
• Python 3.11.3
• fedifeed - Display Mastodon Posts in a curated feed with an user-customisable algorithm
• Paperless-ngx is a document management system that transforms your physical documents into a searchable online archive so you can keep, well, less paper
• The RWKV Language Model (and my LM tricks)
• DeepSpeed Chat: Easy, Fast and Affordable RLHF Training of ChatGPT-like Models at All Scales
• Unpoly - The unobtrusive JavaScript framework for server-side web applications
• Enhance - The HTML first full stack web framework
• Image BPP Calculator
• zola (né Gutenberg) - A fast static site generator in a single binary with everything built-in
• Running Jupyter notebooks online | Cool list of sites where you can run your notebooks

Podcasts

• LWiAI Podcast #117 - Google’s Bard Rush, BloombergGPT, ChatGPT King, Balenciaga Harry Potter (Last Week in AI) | Trying to keep up with all this buzz
• #370 – Edward Frenkel: Reality is a Paradox – Mathematics, Physics, Truth & Love (Lex Fridman Podcast)
• The End of Mediocrity (part 2) (Akimbo - A podcast from Seth Godin) | Yup
• ChatGPT Is Not Intelligent w/ Emily M. Bender (Tech Won't Save Us) | Well, I agree with the analysis of why the letter is more of a marketing move and misleading. And I also think there're some valid critical points considering the social implications etc (I'm not an expert on this stuff, but it sounds plausible to me). But most of the technical assumptions and arguments are just wrong. And I have a problem with the structure of the presented arguments (yes, it's that bad). I use this LLM stuff for some time now and it has been really useful for stuff I'm an expert on. Dismissing this utility based on theoretical grounds gives me some strong "Pope vs Galileo - I don't have to look" vibes. TLDR: for this to be considered the strongest criticism, it is pretty weak.
• FS264 Die Große Elchwanderung (Freak Show) | After 1.5 years, finally! 🎉

"for legal reasons this is a joke" --julia, serial project starter

Lots of colored eggs. Short work week where I was collecting some material for an upcoming pytest-course and did some research about this whole shiny new LLM field which is completely wild atm. Recorded a podcast episode about LLMs. Released a new django-cast version with some minor improvements. Fixed a 500 error on this site caused by the csrf-middleware raising an error and context-processors weren't run so the custom error template didn't get its base template (this needs more fixing).

Articles

• Django: Avoid database queries in template context processors | Probably I should get rid of the db query in my theme context processor for django-cast
• Ban 1+N in Django
• Thoughts on a Crazy Week in AI News
• Slack | Martin Fowler on including slack in your iterations
• CAN Injection: keyless car theft | Ahahaha...
• Exclusive: OpenAI Used Kenyan Workers on Less Than $2 Per Hour to Make ChatGPT Less Toxic | Well, just to be able to find the link again. This article does not look credible at all (as well as the whole time.com site, meh).
• Thinking About Code Review | Yep, PRs make sense in an open-source context. Otherwise, not so much...
• What are we going to do about ChatGPT? | More reasonable than the letter
• Programming-language popularity by GitHub pull requests

Mastodon / Twitter

• Everyone needs a break | Webcomic about the need to get away from everything
• Brain vs Sleep webcomic - "pacific ocean"

Freelance

• Double Your Freelancing | Ok, maybe I have to do some marketing...
• 030 | The Side Hustle | The Unspoken Lever of FI | Fixed income would be nice, too

Papers

• Eight Things to Know about Large Language Models | Fascinating

Weeknotes

• Weeknotes: A new llm CLI tool, plus automating my weeknotes and newsletter | Simon Willison
• Weeklog for Week 13: March 27 to April 02 | Johannes

Software

• Django Playwright and Pytest example
• unittest-parametrize | Works like the @pytest.mark.parametrize decorator from pytest
• Django 4.2 released
• Let it be free! Open-sourcing “Baby AGI”, a paired down version of the “Task-Driven Autonomous Agent” at 105 lines of code. Three task agents (execution, creation, prioritization) work in harmony… forever.
• Just been informed that #AutoGPT is ranked #1 trending in the entire world on @Github !🤯
• SentencePiece is an unsupervised text tokenizer and detokenizer mainly for Neural Network-based text generation systems
• Slint 1.0: The Next-Generation Native GUI Toolkit Matures | Native UIs in rust
• pinecone.ai - Long-term Memory for AI
• babyagi - This Python script is an example of an AI-powered task management system | Twitter thread
• Libraries.io monitors 7,144,509 open source packages across 32 different package managers, so you don't have to | Used this to track how many packages depend on packages I wrote...
• Tabby - Self-hosted AI coding assistant. An opensource / on-prem alternative to GitHub Copilot
• opentalk.eu - video conference software which allows self-hosting

Podcasts

• What AI can do for you on Theory of Change (no link to original podcast, since the feed is broken) | Really good episode!
• Soumith Chintala: PyTorch On the past and present of machine learning frameworks, and the story of PyTorch and its creator (The Gradient Podcast) | Interesting
• The Story of Visual Studio Code with Erich Gamma and Kai Maetzel (The Hanselminutes Podcast) | Surprising how early and small vscode started...
• Aidan Gomez, co-founder and CEO of Cohere joins Host Pieter Abbeel to discuss transformers, large language models, command and instruct models, prompt engineering, LLMs competitive landscape (The robot brains)

if my editor reads this: everything’s totally fine --Ethan Marcotte

Not much work, but a fair share of paperwork, and finally traveled to Erlangen. New django-cast-release and applied with django-cast to the prototype fund🤞. Downloaded the LLaMA weights, and hope to find some time to play around with LLMs locally next week.

Articles

• AI-enhanced development makes me more ambitious with my projects | Example is a proxy fetch function to submit past chatGPT conversations to datasette, cool!
• Cheating is All You Need | First Steve Yegge post I've seen in a long time, welcome back 😁
• Learn Images | I guess I have to
• Rust's Golden Rule
• My Favorite CSS Snippets
• Defaulting on single page applications (SPA)
• Pause Giant AI Experiments: An Open Letter | I think this is dumb, but some smart people did sign it? wtf...
• Unravelling del
• Illustrating Reinforcement Learning from Human Feedback (RLHF) | How to get from GPT3 to chatGPT
• Aligning language models to follow instructions | InstructGPT
• Learning from human preferences | RLHF explanation from openAI
• Think of language models like ChatGPT as a “calculator for words”
• The age of average | Hmm, for me it's more like regardless of how hard I try to zig, I still zag.

Weeknotes

• Weeklog for Week 12: March 20 to March 26 | Johannes
• Weeknote for Week 13 | Bithive

Mastodon / Twitter

• I wish elon had bought LinkedIn
• 1/The call for a 6 month moratorium on making AI progress beyond GPT-4 is a terrible idea.
• Last weekend's GPT-4 Hackathon at the @agihouse_ai had: ✅ 170 AI founders and researchers ✅ Talks by @eladgil , @ClementDelangue ( @huggingface ), @8enmann ( @AnthropicAI ), @robertnishihara ( @anyscalecompute ), & @BorisMPower ( @OpenAI ) ✅ 50 projects Projects that stood out👇
• llama.cpp just got access to the new Copilot for Pull Request technical preview by @github
• music is mostly just really loud math
• a big deal: @elonmusk , Y. Bengio, S. Russell, ⁦⁦ @tegmark ⁩, V. Kraknova, P. Maes, ⁦ @Grady_Booch , ⁦ @AndrewYang ⁩, ⁦ @tristanharris ⁩ & over 1,000 others, including me, have called for a temporary pause on training systems exceeding GPT-4 | Uhm, no.
• To run the released model with latest llama.cpp, use the "convert-unversioned-ggml-to-ggml" python script and apply the following patch to llama.cpp The latest llama.cpp offers significant performance and accuracy improvements in the inference computation
• Made some quick notes on how to use the OpenAI Python library to make ChatGPT API calls and stream out the response tokens as they arrive
• Announcing the Local LLaMA podcast 🎙️🦙 In today's episode we have LLaMA, GGaMA, SSaMA and RRaMA joining us to discuss the future of AI
• Self-attention is the primary building block of large language models (LLMs) and transformers in general. But, how exactly does it work? 🧵 [1/8]

Open Source Funding

• Zuschüsse: Direkte finanzielle Förderung | NRW
• Time to give back to the community | Open Source Fund

Software

• BlenderGPT - This addon allows you to use Blender with natural language commands using OpenAI's GPT-3.5/GPT-4 | Andrew Ng doesn't like it and I agree
• perplexity.ai | Question -> answer chatbot
• Today we're releasing GPT4All, an assistant-style chatbot distilled from 430k GPT-3.5-Turbo outputs that you can run on your laptop.
• OpenAI Cookbook | Examples of how to use the openAI APIs
• GenAI: generative AI tooling for IPython | Use openAI api directly in notebooks...

Podcasts

• #367 – Sam Altman: OpenAI CEO on GPT-4, ChatGPT, and the Future of AI (Lex Fridman Podcast) | Pretty solid
• pytest tips and tricks for better testing (Talk Python To Me) | Useful tips
• Episode 83: Influencers (Hotel Bar Sessions) | Fun episode, but influencers are maybe a too easy target
• Episode 88: ChatGPT (Hotel Bar Sessions) | Broke my rule not to listen to non-expert people talking about a topic I know something about -> instant regret.
• ChatGPT, GPT4 hype, and Building LLM-native products — with Logan Kilpatrick of OpenAI (Latent Space Podcast)
• Chelsea Finn joins Host Pieter Abbeel to discuss distribution shift, meta-learning, editing LLMs, single-life RL, and what can AI not (yet) do today? (The Robot Brains)
• Alexandr Wang joins Host Pieter Abbeel to discuss data, labeling, foundation models, LLMs, truthfulness, reinforcement learning for human feedback (RLHF), and AI for defense (The Robot Brains) | Nice episode slug btw 😂
• How Django Really Works - Chaim Kirby (Django Chat) | Good! How do we spread the word about Django?
• Prompt Engineering and AI Constitutions with Stephen Wolfram | Nice one

we saved daylight but at what cost --K

Much less work last week, but I still had a lot of paperwork to do. Worked on htmx integration for django-cast during this year's Python Spring Sprint & Hackathon in Düsseldorf. Htmx is now working for pagination in both built-in themes. See the latest release notes for all django-cast-related updates. Had a lot of fun attending the latest Django Meetup Cologne.

Despite being self-employed for over ten years now, I never publicly announced being open to new projects. Until now 😁!

Articles

• How virtual environments work | Didn't know there are so few lines needed to create a virtualenv
• Marketing for Developers | Hmm, didn't know about the Django subreddit
• Add Vue To Your Django Templates With Vite | Second part of this tutorial, got a working Django-Vite-Vue combination running using it 👍
• Sentry - Development Philosophy | I'm a little bit skeptical about claims like "we are not artists" or "Our work is there to solve customer problems". Writing computer programs is one of the purest creative activities in human history (pure as in "only limited by our imagination, not practical skills or laws of physics"). Claiming it isn't smells like overcompensation (bad) or ignorance (even worse).
• Lessons from a Pessimist: Make Your Pessimism Productive | Good point - I tame my pessimism by framing it as "optimistic realism"
• Chris’ Corner: Websites, Highlights, and Guesses
• OpenAI is testing the ability to run code and use third-party plugins in ChatGPT | Interesting
• The End of Front-End Development | Holy shit, 65k? I mean, ok it looks great and stuff but...

Weeknotes

• Weeknotes: AI won’t slow down, a new newsletter and a huge Datasette refactor | Simon Willison
• Weeknotes March 19 - 26 | Bithive
• Weeklog for Week 11: March 13 to March 19 | Johannes

Books

• Every Layout | Has a cool website, too

Software

• django-hatchway - Hatchway is an API framework inspired by the likes of FastAPI, but while trying to keep API views as much like standard Django views as possible
• Django Ninja - Fast Django REST Framework
• Orval - Generate client with appropriate type-signatures | OpenAPI-Schema to typescript code generation tool
• Codon is a high-performance Python compiler that compiles Python code to native machine code | Interesting
• Dalai - Run LLaMA and Alpaca on your computer
• Podsqueeze - Generate shownotes, timestamps, newsletters and more for your podcast with one click! | Bootstrapped SaaS from an ex-colleague

Mastodon / Twitter

• Carbonara mit Bildern
• Someone referred to Podcasting 2.0 the other day. Oh my god. There is only one podcasting. I can tell this is going to be really tiresome. You know like Web 3 and all that bullshit. The other day someone told me that the real Web 3 is AI. Sigh.
• This blog post seemed very normal until I hit this bit GITHUB ACCIDENTALLY POSTED THEIR PRIVATE KEYS TO GITHUB THERE IS LITERALLY NO ONE ON EARTH ABLE TO USE THIS PROGRAM SAFELY

Podcasts

• Die Geschichte hinter PHPUnit – mit Sebastian Bergmann (Super Duper Developers Club) | Good episode
• DjangoCon US 2023 - Drew Winstel (Django Chat) | Good as always
• Revision 560: TypeScript 5.0 (Working Draft) | We sometimes joke about including the pre and post-show chatter because it's way more entertaining 😁 than the "real episode". Here they did exactly that and it worked out great!
• WR1467 Pazifistische pflegebedürftige Pensionäre (WRINT: Realitätsabgleich)
• A.I. Is About to Get Much Weirder. Here’s What to Watch For. (The Ezra Klein Show) | Quite solid for a journalistic publication
• Accessibility, Frontend und der ganze Rest mit Manuel Matuzović (Wo wir sind ist vorne.) | Wow, they talked about htmx and alpine in a frontend podcast 😄

sorry about that, i wasn’t making a strange face because of what you said on the team sync meeting, but because i’ve been watching my own camera the whole time, and at all times i am nothing but a horse discovering his own reflection --henry ✷

Covid is gone 😷, but I had to deal with a lot of sickness-related stuff last week, therefore not a lot happened except the usual work stuff. I updated my local Mastodon instance to version 4.1.1 and it went surprisingly smooth. Just like all the others I played around with GPT-4 and found it to be a bit better (more knowledge about details, working better as a world model).

For django-cast I tried to add htmx support but got stuck fixing bugs introduced by the week before last week's theme-switching feature - details are in the release notes for django-cast 0.2.9. Django-cast was also featured in last week's Django News newsletter 😎. This comes a little bit early but at least it's pushing me to create some proper installation documentation 😬.

Articles

• The EU's new Cyber Resilience Act is about to tell us how to code
• Docker is deleting Open Source organisations - what you need to know | I switched away from docker for my own projects in mid-2021 and never looked back, seems this bet is paying off now...
• Science and Technology links (March 11 2023)

Weeknotes

• Weeklog for Week 10: March 06 to March 12 | Johannes
• Weeknotes March 12 - 19 | Bithive
• Week 11: GPT 4 & new Python tooling | Oliver

Mastodon / Twitter

• Having a single user (or in my case three-user) #mastodon instance has some annoyances, but a couple of clever hacks have really improved the experience. Quick thread | Howto follow hashtags if you are on a small instance like me - found it useful!
• English Nursery Rhyme (ca 1764) | Spot on

Software

• descript - a new way to make video and podcasts | Hmm, editing using the transcript is a nice idea, but the audio enhancement is not good (compared to auphonic for example) maybe I look again in a year 🙂
• Learn How to Support Apple Devices | Lots of nice tutorials for the Apple ecosystem

Podcasts

• Riskante Zeiten – Sicherheit weiter denken (Sein und Streit) | Interesting
• Welche Einsichten verdanken Sie der Psychoanalyse? (Das philosophische Radio)
• Rick Rubin (Design Matters) | Great episode
• Yoshua Bengio: large language models, higher cognition, causality, working memory, responsible AI (The Robot Brains Podcast) | Fantastic episode! Is reducing the world to a few discrete parameters (like words in a sentence) a way of regularization? Maybe that's how the brain subsystems manage to avoid a combinatorial explosion while communicating. Reminds me of software: Splitting up the functionality in modules so that there's not a combinatorial explosion of code paths is the hard problem here, too (ok, this idea needs some more thought but I just wanted to get it out of my head...).

Out of Context Images

Couldn't go outside much due to covid. Sometimes I think some higher being put a sticker like that below (reading "please ignore the dog") on my back.