In my View, August 13, 2023
In this Sunday’s ‘AI In my View’ piece, I’d like to talk about how AI software and applications have a unique opportunity to empower billions emotionally rather than just rationally as computers have done for decades. New voices on our shoulders at Scale. Let me explain.
Much has been made to date of AI making us smarter, bolstering our IQ as and when needed. Augmenting our capabilities to make us better programmers, office workers, and searchers for knowledge and understanding.
Microsoft brands it cleverly as ‘CoPilot’, for everyone from Coders to their long-lauded ‘knowledge workers’ using Office 365 Productivity apps. Google is enhancing their core Search services with brands like Bard, Gemini, and Generative Search Experience (SGE). OpenAI of course kicked it all off with “ChatGPT”, implying a talkative assistant that’s always there to help with anything on our mind.
We all have a rational and emotional side. And much of the above is to make us better rationally. Assistants that boost our abilities when and where needed, and maybe over time boost our overall IQs during our waking hours.
It’s how we’ve used computers for over 50 years. Making us better able to use our minds to do things. Spreadsheets and word processors that made the PC Era go. Helped Bill Gates really build Microsoft, along with of course Office and Windows 95. What Steve Jobs beautifully called ‘bicycles’ for the mind (still gets me to this day).
But LLM AI in particular has capabilities that can potentially also help us emotionally as well. Provide unconditional emotional support when and where we need it.
Just as pets do for over half the population. Beyond Pokémon and virtual pets that captivated us all for a while on handheld gizmos and consoles.
TikTok has already cracked the code on using AI to entertain us, as I’ve outlined already. That’s the most mainstream way to impact billions emotionally via this technology.
Meta and Mark Zuckerberg are laser focused on this front as well, with their increasingly AI driven Reels ‘short video’ augmentation of Instagram, WhatsApp and other Meta apps that reach billions every day.
And of course Google has already enthralled billions of our young and old worldwide with YouTube for over a decade. And not just with short videos ‘Shorts’. And all these companies are already generating fast growing billions in revenues from these efforts.
But those are mostly ‘one to many’, mostly human created pieces of AI driven content, finding themselves into our personalized ‘For You’ feeds algorithmically. As Goldman Sachs chronicled recently in a report, a $250 billion plus ‘Creator Economy’ growing to almost half a trillion by 2027.
The next holy grail for the industry is ‘one to one’ connections, mostly computer created, affecting each one of us personally. To do everything from entertaining us, to interact with us emotionally, to hopefully make us better over time.
Science Fiction has already taken us there with movies like ‘Her’, the iconic AI-driven ‘Blade Runner’, and so many others.
To be drawn into worlds where our tendency to anthropomorphize digital personas will always be a reality. We’ve been doing it with pets for millennia.
Humans are pre-disposed to a world where our rational brains struggle with our emotional brains. Trained as they are by ‘survival of the fittest’ Darwinian evolution.
Already AI startups from Replika to Character.ai and dozens of others are fast creating ‘smart agents’ that are catering to our every whim and curiosity. Our need for entertainment and companionship. Bill Gates is already focused on these use cases, for both rational and emotional needs.
Mega companies like Meta and billion dollar newly minted unicorns like Inflection and others are racing towards these opportunities. And investing billions.
Indeed, the aspirations and ambitions around the ‘Metaverse’ and AR/VR/MR (artificial, virtual and mixed realities) are but a first step into these worlds. It’s where Apple is leaning in to take us with its Vision Pro next year (a ‘bargain at $3499).
‘Smart Agents’ will be the talk of the town in these worlds to come. They’ll have voices and visual personas. And they will be customized one to one. At least those will be what’s promised.
It is something that I outlined in box number six in the AI Tech wave chart as the holy grail of high growth, high margin opportunity to serve billions very soon.
But the technologies to make all this real good will take longer than we think. It always does.
And despite the rush to fund these ‘smart agent’ companies, we have barely begun here as Microsoft CEO gives his opinion on the state of the art on today’s ‘smart assistants’ like Microsoft Cortana, Amazon Alexa, Apple Siri, Google Assistant and the like, where the big tech companies have already sunk in billions on ‘Voice Assistants’ over the last few years. Now it’s before they’re all revamped with LLM AI technologies in the months to come. He pithily said:
“They’re all dumb as a rock. Whether it's Cortana or Alexa or Google Assistant or Siri, all these just don't work”.
Yet. Lot of technical wood to chop. They’re all going back into the LLM AI workshop.
We have lots of basic technical features to add like user side memory and much much more. AI systems today have the memory of a goldfish as I’ve talked about before. Which is to say, they have barely any memory. (Note for Techies: And this is not just about increased ‘token size’ in LLM AI models like Anthropic’s Claude 2 LLM AI, although that helps. That’s what I call ‘server side’ memory, not personalized, persistent ‘user side memory’ that stays put indefinitely. Lot to be done on memory front broadly. A topic for future posts definitely).
The way we interact with these systems has to move beyond text prompts into a chatbot. Voice and video will play a big role. We will see options in this context before this year is out. But it will take years of experimentation. At scale.
And we’re all going to be eager Guinea pigs for the ride. And likely paying billions for the privilege fairly soon. And of course we will be the product for advertisers, as we’ve always been.
The key here is that the AI software is going to be catering to our emotional side more than our rational brains (and we won’t know how it all really works). That was the basis of the spreadsheets and word processors that were the foundation of the PC and Internet wave empires of the last few decades.
Those applications are being prototyped and experimented with for the AI Age. And more likely than not they’ll be talking to us emotionally more than rationally. Both ‘smart agent software’, and humans leveraging ‘smart agent’ AI software.
Let’s be ready for that. Stay tuned.
I am really enjoying your AI Substack series. Today's issue "Augmenting Us Emotionally" was awesome! Coincidentally, I am an investor in Augment AI which is a personalized AI Assistant. They essentially build a Personal Foundation Model on top of the generic LLMs so the AI can respond in a more personal way...the way you would. I truly believe AI is going to augment everything we do. We are in the first years of a 20 year AI wave. Exciting times!