Custom GPTs for information scanning, Eliot Peper on writing sci-fi,

“The empires of the future are the empires of the mind.”

― Winston Churchill

For the last couple of decades I have expected AI to give us better ways to find the information that we want. Despite advances and a few bright spots here and there, I remain deeply disappointed with the available information scanning products.

Generative AI offers a promise of giving us the power to create our own tools rather than rely on Big Tech or startups to come up with products that suit us.

In the ‘In practice’ section below I share how I built a simple custom GPT, AI Information Scanner. It gives full configuration details so you can adapt it to create your own. It is just a prototype, but points to the potential for far better personal tools soon.

Ross.

📖In this issue

  • How to build a custom GPT for information scanning

  • AI creates a massive leap forward in materials science

  • Sophisticated prompting strategies

  • Sci-fi novelist Eliot Peper on how to write science fiction, information feeds, and cognitive health

  • Redefining distinctive human capabilities with the advent of generative AI

🤸🏽In practice

How to create a custom GPT for information scanning

Generative AI offers real potential to give us more control over our information sources and scanning.

It turns out that OpenAI’s custom GPTs offer a promising avenue. Here is a guide with full working example on how to create your own from this template.

👩🤖What you need to know in Humans + AI

AI massively augments new materials research

Google DeepMind has announced it has created an AI that has generated over 2 million new materials form inorganic crystals and made this openly available to researchers. There is still plenty of work from researchers to bring the most promising to fruition, but there are potential applications to improving batteries, solar panels, semiconductors, and more.

Governance shifts at OpenAI will shape the future of AI

OpenAI Sam Altman did an in-depth interview on his firing and re-hiring, though he was not fully transparent, including refusing to say why he had been fired. While he emphasized that AI safety comes first, the governance shifts will certainly give more power to him and co-founder Greg Brockman. Humans and human emotions and power battles are shaping AI.

Sophisticated prompting strategy for deep domain expertise

One of the debates in AI is whether we will get more domain-specific fine-tuned models trained on defined bodies of expertise, for more efficient, superior results. A new prompting approach in which GPT-4 generates step-by-step thinking and excellent results has been used in medicine, beating medical models, and can be applied in other domains.

🎙️Latest podcast episode

Eliot Peper on writing science fiction, information feeds, inhabiting the edge, and habits for better cognition (AC Ep21)

🧑‍🎓What you will learn

  • Insights on the interplay between technology, literature, and society (03:26)

  • Future of digital services and the integration of programming ideas in writing (11:40)

  • Exploring the feed as a meta-system that encompasses the entire internet (13:50)

  • Emphasizing the importance of simplicity (17:20)

  • How ‘caring’ can be a differentiator in business and personal life (20:56)

  • Exploring fiction through detailed observation (24:22)

  • Drawing inspiration from personal anomalies (27:00)

  • The importance of good sleep (30:28)

  • Boosting mental capacity by spending less than you earn (31:37)

  • The role of having a community in enhancing mental health and cognitive function (33:05)

  • Enhancing life’s journey with consistent habits and patience (34:43)

Redefining distinctive human capabilities with the advent of generative AI

Defining our distinctive human capabilities, the ones that distinguish us from machines for the longest, is perhaps the single most important issue we face.

If we understand that we can redesign work, focus on developing our unique capabilities, and best complement ourselves with machines.

Thank you for reading!

Please do let us know any feedback and what you’d like to see in the newsletter. We’ll be continuing to develop the content to be better and more useful for you in coming weeks and months.