Envisioning

Happy Monday and welcome to another edition of Artificial Insights, where we explore as many different perspectives about the emerging field of AI as possible. As 2023 starts drawing to a close we may ask ourselves what to expect from the year ahead. This has been a year of (dare I say) unprecedented change when it comes to practical applications of machine learning, and 2024 will be even more so.

I am sure you are also experiencing AI fatigue. The news can be overwhelming, and the conversations about its manifold implications can feel draining. I am not here to fan the flames. Rather I think we should spend more time experimenting with it. The only race worth mentioning is the one towards greater, better use of these tools. Depending on your use, they can be borderline magical.

There is no doubt whether you are going to see more AI in the months and years ahead - I think that’s a certainty. The only question worth asking is what role you want to play in this transition.

PS. I’m in Amsterdam this week for THNK and Lisbon next for WebSummit. DM if you’re around and up for coffee.

From Twitter

Virtual Coworkers

https://twitter.com/Altimor/status/1721250514946732190

Tiny Language Models

Katie Stein from Genpact outlines generative AI's significant potential at Cypher 2023, urging companies to prioritize it for transformative growth. She highlights the importance of considering values beyond efficiency, the need for leadership and societal engagement, and addresses the challenges in AI adoption and the criticality of responsible business practices.

Potential for Transformation: Generative AI is poised to fundamentally change business and society.
Beyond Efficiency: A call for adopting AI strategies valuing more than just cost savings.
AI Adoption Hurdles: Recognizes a high failure rate in AI initiatives, emphasizing the necessity of responsibility and change management.

Mary Mesaglio and Don Scheibenreif at Gartner address AI's transformative role across industries and its implications for human-machine interaction. They underscore AI's potential in enhancing efficiency, customer experience, and innovation, while cautioning about the disruptive challenges it poses to governance and organizational principles.

Transformation Potential: Illustrates AI's capacity to revolutionize industry practices and outcomes.
Ethical Navigation: Emphasizes the need for value-driven principles in AI governance.
Strategic Implementation: Urges clear understanding of AI's organizational benefits and risks.

Lilach and Ethan Mollick from Wharton Interactive present an amazing lecture series about AI's role in education, focusing on democratizing education access and how AI, especially large language models, is used in classrooms. The evolution of AI in language tasks and its unexpected disruptions in high-skill jobs are explored, alongside ethical issues like copyright infringement and biases. The whole series is highly recommended and worth a listen.

Educational Transformation: AI's potential to revolutionize access and learning methods in education.
Ethical Considerations: Highlights the importance of addressing AI-induced ethical dilemmas in education.
AI Advancement: Discusses the unexpected impacts of AI, especially language models, on high-skill professions.

Long Reads from Substack

[

Subconscious

Artificial Intuition, not Artificial Intelligence

LLMs are Artificial Intuition, not Artificial Intelligence. I think this is a meaningful analogy? Let’s piece together a hunch from a few insights… First, this paper suggests LLMs simulate reasoning by reaching for pre-baked reasoning-like behavior in their training data, rather than baking it from scratch…

2 years ago · 3 likes · Gordon Brander

](https://subconscious.substack.com/p/artificial-intuition-not-artificial?utm_source=substack&utm_campaign=post_embed&utm_medium=web)

Emerging Vocabulary

Residual-Aided Intermediate Fusion (RAIF)

An architecture in deep learning focused on multi-modal aesthetic recognition. It employs a strategy that combines feature representations from different stages of a neural network, utilizing both intermediate fusion and residual skip connections to enhance recognition performance. Can be used for multi-modal aesthetic recognition and multimodal matching of vision and language.