Not AI
World Smallest Office Suite: Obviously impractical, but the way he implements the apps is very interesting.
I found the best anagram in English: Perl veteran finds an interesting way to rank anagrams.
AI
Bing’s A.I. Chat: ‘I Want to Be Alive. 😈’: Microsoft’s new chatbot goes crazy after a journalist uses psychology to manipulate it. The article contains the full transcript and nothing else. It’s a fascinating read.
LessWrong: Sydney and Bing: Some interesting insights into how LLMs work and why they can produce such crazy behavior.
Creating desktop backgrounds using Midjourney: Blog author creates a desktop background using Midjourney. Not bad!
Building LLM Applications for Production: Detailed guide for enterprising people who want to integrate LLMs into applications.
Unspeakable Tokens: A deep dive into the inner workings of ChatGPT, and why it stops responding or replies weird or creepy things to seemingly simple requests.
Ethan Mollick’s tweet: Awesome things ChatGPT’s experimental code interpreter can do.
It is starting to get strange: Longer article about the code interpreter. Worth reading, very impressive.
I have similarly uploaded a 60MB US Census dataset and asked the AI to explore the data, generate its own hypotheses based on the data, conduct hypotheses tests, and write a paper based on its results. It tested three different hypotheses with regression analysis, found one that was supported, and proceeded to check it by conducting quantile and polynomial regressions, and followed up by running diagnostics like Q-Q plots of the residuals. Then it wrote an academic paper about it.
Thinking companion, companion for thinking: Using AI to get constructive criticism.
Sam Altman: OpenAI CEO on GPT-4, ChatGPT, amd the Future of AI: A long interview filled with interesting insights. Sam strikes me as a very honest, careful, and humble person. I may be wrong, but I’m glad he’s at the helm of OpenAI.
Let’s be bear or bunny: Funny hallucination by otherwise very inpressive LLM running locally on an iPhone.
Replacing my best friends with an LLM trained on 500,000 group chat messages: Guy trains an LLM on his group chat messages with his best friends with predictable but nevertheless very funny results.
An example of LLM prompting for programming: Martin Fowler shows how his colleague uses ChatGPT for coding.
Scientists use GPT LLM to passively decode human thoughts: Big if true.
IBM plans to replace 7,800 human jobs with AI, report says: So it begins.
How to use AI to do practical stuff: A new guide: A nice guide you can send to your friends who want to try this new AI thing.
replit-code-v1-b3: Repl.it’s new open source code LLM - surprisingly good. Demo: https://huggingface.co/spaces/replit/replit-code-v1-3b-demo