weekend ai reads for 2024-10-04

This book is a short introduction to deep learning for readers with a STEM background, originally designed to be read on a phone screen.

If it’s a 2 minute task, do it yourself

If it’s a 20 minute task, use an LLM

If it’s a 200 minute task, use an LLM as a co-pilot

There are a lot of details we still have to figure out, but it’s a mistake to get distracted by any particular challenge. Deep learning works, and we will solve the remaining problems. We can say a lot of things about what may happen next, but the main one is that AI is going to get better with scale, and that will lead to meaningful improvements to the lives of people around the world.

The answer is that AI use that boosts individual performance does not always translate to boosting organizational performance for a variety of reasons. To get organizational gains requires R&D into AI use and you are largely going to have to do the R&D yourself. I want to repeat that: you are largely going to have to do the R&D yourself.

41% of CIOs said net new spend. 35% said existing software spend. Only 6% said it would come from professional services.

AI consulting retainers offer a middle ground between fully outsourcing your AI needs and building an in-house team from scratch. Under this model, you engage an AI consulting firm on an ongoing basis, typically for a fixed monthly fee. This fee, often ranging from $15,000 to $30,000 per month, grants you access to a team of AI experts, cutting-edge technologies and industry best practices.

The key advantage of this approach is its ability to dramatically reduce both upfront and ongoing costs. By leveraging a consulting firm’s existing infrastructure and expertise, you sidestep the need for massive initial investments in hardware, software and personnel. You're essentially renting not just technology but also the brainpower needed to implement it effectively.

This paper introduces the concept of an education tool that utilizes Generative Artificial Intelligence (GenAI) to enhance storytelling for children. The system combines GenAI-driven narrative co-creation, text-to-speech conversion, and text-to-video generation to produce an engaging experience for learners. We describe the co-creation process, the adaptation of narratives into spoken words using text-to-speech models, and the transformation of these narratives into contextually relevant visuals through text-to-video technology. Our evaluation covers the linguistics of the generated stories, the text-to-speech conversion quality, and the accuracy of the generated visuals.

In Round 1, we compare the models within each family—that is, GPT-4o-mini vs. GPT-4o (OpenAI, 2024); Llama-3.1-8b vs. Llama-3.1-70b (Meta, 2024); and Gemma-2-9b vs. Gemma-2-27b (Google, 2024). Results from Round 1 indicate that within each model family, the smaller models generally outperformed their larger counterparts.

AI tools can increasingly help with other mundane tasks (“toil” in coder-speak), such as writing notes about what the code does or designing tests to make sure code won’t malfunction. Writing code is only a part of the job of a software engineer, accounting for about 40% of their time, according to Bain, a consultancy.

This is a domain-expert finetune based on the US Army field manuals (the ones that are published and available for civvies like me). It’s focused on factual question answer only, but seems to be able to answer slightly deeper questions in a pinch.