From 'Her' to Present-Day AI

Every now and then a movie comes out that completely blows me away. Denis Villeneuve’s 2021 'Dune', the Wachowskis’ 1999 classic 'The Matrix', and 2004’s 'Eternal Sunshine of the Spotless Mind' all fall into this category for me. 'Her', the 2013 Spike Jonze's thought-provoking film film, also is included in this short list of films in my humble opinion.

“Her” is a thought-provoking and moving film that explores the nature of love, relationships, and the future of artificial intelligence. It is a film that will stay with you long after you have seen it. I rewatched over the past weekend and it holds up extremely well.? I remember reading a great piece in WIRED by Vlad Sejnoha titled “Can We Build ‘Her’?: What Samantha Tells Us About the Future of AI” back in 2014 and I have to admit, what I find most interesting is how much has actually changed in the AI space in the past ten years since the movie was released.

Central to the movie’s storyline is the love affair between Theodore, played by Joaquin Phoenix, and the operating system (AI) Samantha, voiced by Scarlett Johansson. This is a relationship that grows (almost) entirely through conversations between the two characters. In "Her," we witnessed Theodore conversing naturally with Samantha, blurring the lines between human and AI interaction.?Back in 2013 the concept of an AI with this capability was clearly science fiction.

Since then, Natural Language Processing (NLP) has evolved in leaps and bounds. For example, in 2018, Google introduced BERT (Bidirectional Encoder Representations from Transformers), which significantly advanced language representation learning. BERT's pre-training and fine-tuning approach provided contextual understanding of words in sentences and led to breakthroughs in various NLP tasks, including sentiment analysis, named entity recognition, and text classification.?

In 2020, OpenAI introduced GPT-3 (Generative Pre-trained Transformer 3), a language model with a staggering 175 billion parameters. GPT-3 revolutionized NLP by demonstrating its ability to generate coherent, contextually relevant, and human-like text. It showcased advancements in tasks such as language translation, question-answering, and text completion. Conversations with AI have become more natural, dynamic, and engaging than ever before.

Another huge technological advancement since the release of ‘Her’ is LaMDA (Language Model for Dialogue Applications). Google introduced LaMDA back in 2021, a language model designed specifically for dialogue applications. LaMDA focuses on improving conversational understanding, allowing users to engage in more natural and context-rich conversations with AI systems. It aims to provide more meaningful and coherent responses, enhancing the interactive capabilities of AI-powered dialogue systems. In February 2023, Google announced Bard, a conversational artificial intelligence chatbot powered by LaMDA. Bard (ok, and ChatGPT) are helping me research and write this article, while other AI tools such as Notion AI are fine tuning this post’s end product.?

The interaction between the OS Samantha and Theodore IRL is also one of my favorite parts of the movie ‘Her’. Theodore uses a foldable phone looking device to interact with the OS via a camera and audibly through an earpiece. The scene they share on the beach is one of my personal favorites and is remarkable storying telling. Funny enough, this is another area where significant advancement has taken place since the release of the film.?

Google has been exploring the fusion of text with other modalities such as images and videos to enhance NLP capabilities. We call this Multimodal NLP. By combining visual and textual information, multimodal NLP enables more comprehensive and nuanced understanding of content. Google's research in this area aims to create models that can analyze, interpret, and generate text based on multimodal inputs.

AI can now analyze and interpret images and videos with astonishing accuracy (And yes, we still have a long way to go. The 'hand' thing is annoying). Object recognition, facial recognition, and scene understanding have become commonplace, powering applications ranging from autonomous vehicles to medical imaging. AI's ability to "see" and understand visual data has opened up a world of possibilities, impacting various industries and research domains.

Samantha's emotional depth in "Her" fascinated me, igniting curiosity about AI's ability to comprehend human emotions. Again, a shout out to Scarlett Johansson who we never actually see in the movie. Talk about conveying emotion only through voice. Fast forward to the present, and emotional AI has become a thriving field. Researchers and developers have made strides in emotion recognition and sentiment analysis, allowing AI to better understand and respond to human emotions. From chatbots offering empathetic support to emotion-sensing technologies, AI is increasingly becoming attuned to our emotional needs.

From the imaginative world of 'Her' to the reality of today, AI has leaped forward, transforming the way we interact, understand, and perceive artificial intelligence. We've witnessed groundbreaking advancements in Natural Language Processing, the rise of voice assistants, the emergence of emotional AI, the development of computer vision. Just imagine where we will be ten years from now.