The Future, In Five Prophecies | Issue #6
My topmost predictions for AI in the coming decade
7 MINS READ
Hello,
With this newsletter, we have crossed five issues. We’ll celebrate 500+ together. Thank you for every read.
We changed our title style from uppercase to title case.
If you haven’t subscribed yet, please do. And share.
Happy reading!
Today, I’ll put on the garb of a prophet and pronounce a slew of predictions for AI in the next few years.
Why should you listen to me?
There’s no reason to. Lol.
I got you there. You thought I was going to lay out an array of credentials?
Maybe, investor in several AI companies. Featured on the TIME list of most influential people in AI, etc.
Well, I am none of those (yet) but here’s surely what I am.
I have been reading about AI for ~7 years (books, blogs, articles, essays), have read AI news every day since January, tinkered with over 50 AI tools, have worked on some AI projects myself, done a few courses, I’m a master-level prompt engineer (no fluff. I can prove it at the drop of a hat), and I have listened to the predictions of people with the credentials.
In essence, I am not saying I have my finger on the pulse of AI. I am saying that, far more than the average person, I am very conversant with what AI has been and what it is now. So, I can make intelligent conjectures about where it is headed (or is going to land. On its head.).
The following predictions are neither purposely brazen nor deliberately shy (conservative) of some of the (attention-grabbing) predictions you may have seen or heard in the past months or years.
They are just mine.
So here goes nothing:
AGI will be created this decade.
Am I starting with the biggest, boldest statement? Or do I truly believe this? The latter.
AGI stands for Artificial General Intelligence. It has been the vague and distant AI dream; the AI community’s “elixir of life”. AGI is to the AI community, what Heaven or paradise is to the faithful. It is indeed the holy grail.
It refers to an artificial intelligence that is just as good as a human; an artificial superintelligence. AGI would be able to think and reason and communicate and create, with its cognitive functions objectively at the level of an average human. Although, it’s the kind of average that has all of Wikipedia in its head.
AGI will take the wheel to help us in our most cherished position: thinking.
As a thinking machine armed with absurdly vast amounts of data, AGI could pull together millions of data on compounds, doctor’s reports, patients’ records to present the cure for cancer.
Or, analyze crop yield, theoretically restructure government budgets, simulate labor and other relevant parameters to solve world hunger.
The applications are potentially limitless. Our biggest problems would be brought to this demigod of our own creation. Poverty and wealth inequality, and climate change, and war and intergovernmental avarice (which I think can only resolve ourselves, when we eliminate greed and selfishness.)1
I believe AGI will be the last in our long, long line of tools. It will be the tool that makes other tools.
AI and data scientists are pursuing this goal with renewed vigor; recent breakthroughs for example, in Large Language Models (LLMs), have shattered the ceiling of what was thought possible, salaries are more buoyant as venture capital and Big Tech have quit goofing around and are pouring dollars into AI projects, betting on a public that is insatiable for AI innovation and that they hope will reward capital with free-flowing profits.
For context on funding, recently, Visa pledged to invest $100M in companies using generative AI for fintech, Amazon invested $4 billion in Anthropic and OpenAI is still receiving funding, making it to an ~$80 billion-dollar-valuation.
And you know how humans seem to have a knack for achieving impossible things with sheer willpower and taxpayer’s money. Or, should I reference the moon landing?2
Every government will have an AI ministry, every big business will have an AI division
Many predictions for AI that people make, seem to flow from “we’ll create AGI in ten years” and I hinted some of them while I was writing about that. However, here, I am predicting future response to AI and the level of adoption.
Intergovernmental bodies like the United Nations have started talks on the promise and precautions of AI. It was even one of the agenda on the floor of the last General Assembly.
In the U.S., the White House got big companies with interests in AI, like Google and Meta to sign a document promising to develop the tech responsibly. U.S. senators also held a closed-door meeting with AI bigwigs like the CEOs of Big Tech, and newer AI startups like OpenAI and Anthropic.
In the next decade, most governments will either treat AI separately from overall technology adoption or it will be primal focus3. Most governments will allocate significant budgetary spend to AI strategies. The current governmental AI turf between countries like the US and China, will brew hotter.
Also, corporate bodies will continue using AI to seek new fronts in customer acquisition and retention, leading to big divisions for managing how it is incorporated into business systems, like sales and marketing, as well as the development of AI, standalone or as product features. (Aside: The corporate AI wars will potentially plateau if open-source models outcompete proprietary software.)
It is here that a burgeoning field for AI consultants emerges plainly. These consultants will help governments and corporate bodies to chart courses to incorporate AI.
AI Agents will be mainstream in two years
Now, what are these?
When you ask ChatGPT a question, it promptly answers and waits for the next question. This is a linear approach.
AI agents, however, are goal oriented. That is, they plan and execute tasks.
When you ask an AI agent a question, it takes the prompt as a goal, writes a list of steps needed to achieve that goal, and starts running them one-by-one, marking them off as it achieves them, like a to-do list.
AI agents are built to connect to other services (via APIs) so they can execute a wide array of tasks, like linking to Maps for directions, Notion for doc management, or SpeechRecognition, as the name suggests for spicing up the user interface.
Think, a lite version of Tony Stark’s J.A.R.V.I.S, stripped of AR, holograms and AGI-level intelligence.
AI agents will be useful from anything as a sales call rep to a personal virtual assistant to sift through your mails and they mark the perceptible next step in AI development - the era of autonomous or interactive AI.
AI agent projects on GitHub, the world’s most popular collaborative coding site, are mushrooming right now. Examples include, Auto-GPT and BabyAGI.
At the risk of offending you, I’d have to stop here for now.
I’m sorry.
This seems to be longer than I had previously thought. I'll send the last two predictions to you tomorrow.
You like this issue or newsletter? Please, please, share.
If everyone here added only one person, we’d double the readers here.
Our next Sunday issue will be a deep dive. I’ll explore an incredibly useful way you can use AI to 10x your productivity.
Enjoy the rest of your day.
With love & ink,
Emmanuel
Still, who will stop us from asking The Superintelligence? Don’t people Google their names?
I will discuss the rapidly shifting AI-scape in a soon-to-come issue and finish a more elaborate issue on AGI later.
The UAE has a ‘minister for artificial intelligence’ and many governments will follow in the coming years.