Happy Monday and welcome to your regular overview of the best links that came across my timeline in the last week.
How do you make space to wrap your mind around the hyperobject that is AI?
What is a sufficient amount of time, energy or money invested in making sense of it?
Reading this newsletter is a small but recurring habit that helps you stay informed. Similarly, writing these weekly insights has been one of my best tools for understanding the field.
Arguably, the best way to develop a deeper understanding of these tools is by scoping and executing personal AI-oriented projects.
Whether you’re just starting with Claude or ChatGPT or are fully experienced with embedded intelligences, the most reliable way to upskill is by building something with them.
Personally, not a week goes by without attempting to build something using current AI tools. Whether it’s creating a vocabulary of AI-related terms or developing a Tarot experience with AI, I always ground new responsibilities around AI’s capabilities.
I suppose you might be doing something similar in your personal or professional life, and if so, I would love to hear about it. I'm slowly organizing a virtual and physical meetup to share such projects with like-minded readers.
If you'd like to bring something to the meeting, please reply or DM me with what you'd like to present. The meetups will happen in August at the earliest, possibly later. Most importantly, we need interesting projects to showcase, so don't be shy!
Until next week,
MZ
In-depth interview with Anthropic CEO
Dario Amodei is one of my favorite people working in AI, and his interviews are always insightful.
Interpretability might allow you to tell the difference: is the model deducing the answer to the question, or is it memorizing the answer to the question?
Visionaries and Disruptors
Great interview / panel with Sam Altman and Brian Chesky (AirBnb) by the Aspen Institute.
AI is going to be completely embedded in everything that we do...almost every industry can get remade with AI.
What it feels like to receive AI generated email, part 2
The nature of spam will 100% adapt to generative technologies and I bet we’ll see a lot more responses like this blog post.
Lovely and unexpected use of AI, to determine which branches of a tree of preferable to birds.
Mind-reading AI
Top row: original images. Second row: images reconstructed by AI based on brain recordings from a macaque. Bottom row: images reconstructed by the AI system without an attention mechanism.
Practical applications of LLMs
One of my favorite uses for LLMs is proposing passwords that follow specific requirements. Time saver for those rare occasions. In the example below I simply pasted the requirements in the prompt.
Strategies for integrating an AI practice into your organization
Excellent article series by Noelle Saldana.
Integrating an AI practice successfully into your organization is a journey and an investment. The integration is prone to failure as organizations do not hire realistic minimum viable teams and do not often set this new organization up for success. Instead, they often simply install one person, often in a vacuum, and hope for the magic to happen. Don’t let this be you!
Via Guilherme on WhatsApp.
AI Election Forensics
AI Forensics investigated how AI-generated images were used in French political campaigns during the 2024 European Parliament and legislative elections.
Via Gust on WhatsApp.
Motion Copilot
Sooooo, this is new and interesting for designers - a motion copilot that creates keyframes for you based on your prompts. It's the first time, as a motion designer, that I've seen a tool that actually works, albeit still with a lot of constraints.
Via Nicole on WhatsApp.
Via on WhatsApp
Riding the Hype Cycle
Spotted in Amsterdam.
-
🏅 Forwarding this issue to colleagues and friends.
-
🎯 Commenting with your favorite talks and thinkers.
You are receiving this newsletter because you signed up on envisioning.io or Substack.