Jon Calder

Today we present new solutions to the Liner Shipping Network Design and Scheduling Problem, released as part of our new Shipping Network Design API, with the goal of maximizing the efficiency of container shipping networks at world-wide scale. Read more → goo.gle/3X61X1Z

Motorists involved in the injuring and killing of cyclists are often not prosecuted, or allowed to get back behind the wheel in a matter of months. This has led to cyclists simply not reporting accidents. #CarteBlanche @ermbates

Meet Exercism's #48in24! 👏 This year, we're challenging you to solve a different programming challenge each week. It's 100% free, gonna be tons of fun, and will help you level up your skills. Signup here: exercism.org/challenges/48i… Intro here: youtube.com/watch?v=v6fMaO…

# On the "hallucination problem" I always struggle a bit with I'm asked about the "hallucination problem" in LLMs. Because, in some sense, hallucination is all LLMs do. They are dream machines. We direct their dreams with prompts. The prompts start the dream, and based on the LLM's hazy recollection of its training documents, most of the time the result goes someplace useful. It's only when the dreams go into deemed factually incorrect territory that we label it a "hallucination". It looks like a bug, but it's just the LLM doing what it always does. At the other end of the extreme consider a search engine. It takes the prompt and just returns one of the most similar "training documents" it has in its database, verbatim. You could say that this search engine has a "creativity problem" - it will never respond with something new. An LLM is 100% dreaming and has the hallucination problem. A search engine is 0% dreaming and has the creativity problem. All that said, I realize that what people *actually* mean is they don't want an LLM Assistant (a product like ChatGPT etc.) to hallucinate. An LLM Assistant is a lot more complex system than just the LLM itself, even if one is at the heart of it. There are many ways to mitigate hallcuinations in these systems - using Retrieval Augmented Generation (RAG) to more strongly anchor the dreams in real data through in-context learning is maybe the most common one. Disagreements between multiple samples, reflection, verification chains. Decoding uncertainty from activations. Tool use. All an active and very interesting areas of research. TLDR I know I'm being super pedantic but the LLM has no "hallucination problem". Hallucination is not a bug, it is LLM's greatest feature. The LLM Assistant has a hallucination problem, and we should fix it. Okay I feel much better now :)

I remember my amazement when I first learned of all the things Ken Thompson has invented/created, things we still use today: UNIX, utf8, regex, and many more.

🎉I've been working on this for > 1 year and I could not be more excited to launch it (soon)❗️ Shinymetrics lets you monitor #RStats Shiny applications. Before I do that I need some more testing; if you are interesting join the waiting list! (see shinymetrics.com)

Amazing: Google partnered with American Airlines, used AI to tweak flight paths to reduce contrails. The results are in, and if scaled up this could immediately reduce aviation's global warming impact by almost 20% Technology just keeps beating degrowth! blog.google/technology/ai/…

@mike_ketchbrook @EtienneBacher @theRcast @rweekly_org @RLadiesGlobal @ivelasq3 @Damien_Dotta It was included under New Packages as part of this week's edition - presumably just didn't come up as part of the highlights discussion (apologies).

Episode 132 of #rstats @rweekly_org Highlights is out! podverse.fm/episode/EO0Wl1… 📺 Automated dashboard of @RLadiesGlobal YouTube feeds @ivelasq3 🚀 Serverless Shiny apps guide (Veerle van Leemput) ❄️ Polar data library cookbook @Damien_Dotta h/t @mike_ketchbrook @jonmcalder

Recycling of solar panels is going to become increasingly important as the vast number of panels being manufactured begin to reach end-of-life: "Over 60% of the value is contained in 3% of the weight of the solar panels" bbc.com/news/science-e…

The new issue of @rweekly_org is out! rweekly.org/2023-W20.html #RStats

The new issue of @rweekly_org is out! rweekly.org/2023-W18.html #RStats

The new issue of @rweekly_org is out! rweekly.org/2023-W17.html #RStats

The new issue of @rweekly_org is out! rweekly.org/2023-W16.html #RStats

The new issue of @rweekly_org is out! rweekly.org/2023-W15.html #RStats

The new issue of @rweekly_org is out! rweekly.org/2023-W14.html #RStats

The new issue of @rweekly_org is out! rweekly.org/2023-W13.html #RStats

The new issue of @rweekly_org is out! rweekly.org/2023-W12.html #RStats

The new issue of @rweekly_org is out! rweekly.org/2023-W11.html #RStats

The new issue of @rweekly_org is out! rweekly.org/2023-W10.html #RStats

The new issue of @rweekly_org is out! rweekly.org/2023-W09.html #RStats