weekend ai reads for 2024-06-14

June 14, 2024

📰 ABOVE THE FOLD: DIGITAL TRANSFORMATION

AI Intensifies Digital Transformation — Findings from the State of Entperise AI and the Enterprise Cloud Index reveal why IT leaders are enabling AI applications and leveraging AI to manage their complex data center operations. / Nutanix (14 minute read)

In this year’s report we discuss the latest developments on 3 values shifts and 1 technological shift, including the potential upsides and downsides, the innovative companies spurring the shift, and white space opportunity.

  • PDF at link; requires email

The State of AI at Work [PDF] / Asana with Anthropic

 

📻 QUOTE OF THE WEEK

I’m the CEO of a two-month-old company and have been able to get meetings with CEOs of 160-year-old publishing companies

James Smith, CEO of Human Native AI (source)

 

🏗️ FOUNDATIONS & CULTURE

WWDC, Apple Intelligence, Apple Aggregates AI / Stratechery by Ben Thompson (15 minute read)

In other words, to the extent that Musk hates OpenAI, he should be happy about this partnership: Apple is clearly not sharing private data with OpenAI, and honestly the warnings it throws up every time you access the service are probably going to get pretty annoying pretty quickly; what the company is doing is providing a standardized interface for OpenAI to get access to potential customers for impressive yet commoditized use cases that Apple doesn’t need to spend resources on, because OpenAI and any of its would be competitors will be compelled to make the investment and accept Apple’s terms in an attempt to find some sort of sustainable advantage.

on-edge LLM inference: a small, low latency AI model (3bn params) will be included in future iOS versions, and it will be able to understand user commands, the current screen, and take actions on apps.

private cloud compute: The on-device LLM may decide to offload certain complex tasks to more powerful models hosted in Apple’s data centers (called “private cloud compute”).

3rd party model inference: Users will also have the option to use OpenAI’s ChatGPT directly from Siri or certain iOS apps. Note, this is not the same as using ChatGPT as a replacement for Siri - which is what many thought the OpenAI partnership meant. Rather, ChatGPT is provided as an alternative to Apple’s models in certain situations (e.g. user is about to perform email revision, and ChatGPT’s response is offered as a choice).

Situational Awareness — The Decade Ahead [PDF] / Leopold Aschenbrenner

Developing an LLM: Building, Training, Finetuning / Ahead of AI, Substack (sorry) (58:46 video at link)

If your weekend plans include catching up on AI developments and understanding Large Language Models (LLMs), I've prepared a 1-hour presentation on the development cycle of LLMs, covering everything from architectural implementation to the finetuning stages.

  • related (1), Generative AI for Beginners — Learn the fundamentals of building Generative AI applications with our 18-lesson comprehensive course by Microsoft Cloud Advocates. / Microsoft, Github

    • good content, lot of reading and decent videos

    • obviously Azure-centric

  • related (2), Say What You See — Learn the art of image prompting with the help of Google Al. / Google Arts & Culture

    • a proper fun way to learn about prompting

 

🎓 EDUCATION

How Students are Actually Using Generative AI as Early Adopters — Discover how students are leveraging AI to transform learning / AI Supremacy, Substack (sorry) (22 minute read)

Predictive Analytics & the Business of College Admissions / AI Tool Report (51:06 Spotify podcast at link)

Three features of AI Tutor Pro are designed to support a variety of learner characteristics. One feature allows learners to select the language complexity of a session by choosing whether they want to interact using language on the elementary school, high school, undergraduate, or professional level.

Inside Barnard’s pyramid approach to AI literacy — The New York institution’s unusual take on artificial intelligence could serve as a blueprint for others grappling with implementation. / Inside Higher Ed (6 minute read)

 

📊 DATA & TECHNOLOGY

We achieved the highest impact with UX that naturally blends into users’ workflows. In all the above examples, a suggestion is presented to the user, taking them to the next step in their workflow with one tab or click. Experiments requiring the user to remember to trigger the feature have failed to scale.

Pricing: Costs vary depending on business size, model capabilities, usage and other factors. IBM also offers a free trial of a “lite” version of its watsonx Code Assistant for Red Hat Ansible Light Speed product.

The Rise of AI Agent Infrastructure / Madrona (11 minute read)

Today, many agents are almost entirely vertically integrated, without much managed infrastructure. That means: self-managed cloud hosts for the agents, databases for memory and state, connectors to ingest context from external sources, and something called Function Calling, Tool Use, or Tool Calling to use external APIs.

For some companies/approaches, the LLM or a series of LLMs handles the conversational flow and emotionality. In other cases, there are unique engines to add emotion, manage interruptions, etc. “Full stack” voice providers offer this all in one place.

How Bias Hides in ‘Kitchen Sink’ Approaches to Data — In risk modeling, AI researchers take a more-is-better approach to training data, but a new study argues that a less-is-more approach may be preferable. / Human-Centered Artificial Intelligence, Stanford University (5 minute read)

 

🎉 FUN and/or PRACTICAL THINGS

Hadana: Your AI travel planner

  • requires registration

Transferscope – Synthesized Reality

Created by Christopher Pietsch, Transferscope is a working prototype that merges a handheld apparatus with artificial intelligence, allowing the user to view our physical world through an AI interpreter.

  • fascinating use of AI in a real-world device

GigaBrain - Search Reddit and Other Communities for Answers from Real People

AI Text to Sound Effects Generator — Generate any sound imaginable from a text prompt / ElevenLabs

Advancing personal health and wellness insights with AI / Google Research (12 minute read)

The Personal Health Large Language Model (PH-LLM) is a fine-tuned version of Gemini, designed to generate insights and recommendations to improve personal health behaviors related to sleep and fitness patterns. By using a multimodal encoder, PH-LLM is optimized for both textual understanding and reasoning as well as interpretation of raw time-series sensor data such as heart rate variability and respiratory rate from wearables.

  • amazing generative video

  • free to try; requires sign-in

 

🧿 AI-ADJACENT

NotebookLM — Note Taking & Research Assistant Powered by AI

 

⋄