Editor’s Note: In every boom, there are two waves of profits…
The first wave is usually centered on a single leader in the market. By getting in early in these first-wave companies, we can see massive gains.
But then, there’s a second wave centered around smaller, lesser-known stocks that piggyback off the market leader… And thanks to their smaller size, they’re often able to deliver even bigger gains.
Jeff has used this two-wave blueprint many times over his more than 30 years working in the tech industry… and now, as artificial intelligence continues to grow more advanced, he’s using it again in this AI boom.
To give you all the details, Jeff has put together a special presentation on how you can use it to take advantage of the second wave of the AI boom.
You can go here to access Jeff’s presentation.
In November 2022, OpenAI released its ChatGPT artificial intelligence (AI). It changed the trajectory of the industry.
Around the same time, another AI with an entirely different skill set was announced.
Cicero – an AI agent developed to play the game Diplomacy – came out of Meta’s Fundamental AI Research Diplomacy team.
Gameplay is often dismissed as being a limited experiment… like playing around in a walled garden. But that’s not the case, as complex game environments are simply stepping stones to real-world applications.
Diplomacy is an interesting challenge for an AI because it cannot be solved through self-play. The game involves seven players who compete to control supply centers on a map. Players engage in private negotiations, often collaborating to mutually beneficial ends.
The implications are what make this challenge interesting.
An AI agent has to be able to assess and understand other players’ motivations and strategies and negotiate agreements with multiple parties.
The goal is to control the majority of the supply centers on the map.
Below is a high-level architecture of how Cicero – the AI agent – was designed to achieve this task.
Source: Human-level play in the game of Diplomacy by combining language models with strategic reasoning
While the research published on Cicero was drowned out by all the excitement around ChatGPT in late 2022, it shouldn’t have been.
What made the research so exciting was that it demonstrated a solution to achieving complex reasoning, one of the biggest challenges on the road to creating an artificial general intelligence (AGI).
The team at Meta entered Cicero into 40 online games of Diplomacy with unknowing humans.
The results were incredible.
Cicero ranked in the top 10% of all human players who had played more than one game.
In other words, human-level performance in complex reasoning.
So it wasn’t a surprise when, last summer, I saw that one of the lead authors of the research, Noam Brown (no relation), was recruited to join OpenAI.
After all, complex reasoning capabilities are like The Fifth Element for artificial general intelligence.
And things are about to get very real.
It appears that OpenAI is getting close to releasing an updated version of its AI, known as Strawberry.
OpenAI CEO Sam Altman went so far as to tease the industry with a post on X that generated quite a bit of excitement.
Strawberry is significant because it is the new code name for what was formally known as Q* (pronounced Q star). Q* was the subject of my November 27, 2023, Outer Limits – The Drivers Behind Sam Altman’s Ousting.
The day before he was fired, Altman made a passing comment that not many people paid attention to. But I did. He signaled that OpenAI’s AI model release in 2024, “will have taken such a leap forward that no one expected.”
Q* – and its rumored ability to reason and solve math problems it had not specifically been trained on – is what caused me to re-evaluate my prediction on when we would see an artificial general intelligence.
Back in 2019, I had predicted that it would happen by 2028, which was seen at the time as a ridiculously aggressive prediction.
Considering the developments of the last six months, I revised my prediction to be no later than 2026.
Just imagine… OpenAI has had the last nine months – and basically unlimited resources – to improve Q*, i.e. Strawberry.
Back then, some executives at OpenAI felt the company was on the cusp of achieving AGI… so much so that they wanted to slow down development for safety concerns. (I covered this incredible power struggle right here in Outer Limits.)
OpenAI also reportedly met with national security officials to demonstrate the latest capabilities of its artificial intelligence.
Sure, there might be a business opportunity there with the U.S. government, but the more important purpose is to get ahead of what’s coming and proactively communicate the capabilities of its AI through open dialog, rather than risk a sharp regulatory attack… or worse, an appropriation of some kind in the name of a “national security risk.”
But Strawberry is just a piece of the puzzle.
Strawberry is part of a large framework – a major enhancement to OpenAI’s large language models (LLMs).
OpenAI released its latest version, GPT-4o, this May with impressive multi-modal capabilities. The industry is now anxiously awaiting GPT-5 – now rumored to be named Orion – which we may see this fall.
Yes, developments in AI really are moving this quickly.
Alphabet just released Gemini 1.5 Pro this May. And to the industry’s surprise, xAI (Musk’s AI startup) released Grok 2 a couple of weeks ago… and it is performing at the level of Gemini and OpenAI’s GPT-4o.
This pace of development and advancement every single month allows us to sense how quickly AGI is coming.
If we just take OpenAI’s developments this year as an example:
Those in the industry are acutely aware of the speed, and how much money is at stake for the winner(s).
The below headline from the Financial Times this morning encapsulates the significance of this megatrend that we’re in the midst of.
Right around the time that this issue of The Bleeding Edge will be published, so too will we be reading about NVIDIA’s earnings.
The company has practically become a household name, and economists now see its earnings results as important as the US jobs report
Its growth represents the growth of artificial intelligence as an industry. The expectation is around $29 billion in revenues for its fiscal second quarter (ending July 31).
It’s a big number, but more relevant is that the number is more than 110% compared to its fiscal second quarter of 2023.
Its growth, and that of many others, is unstoppable right now.
Microsoft (MSFT), Apple (AAPL), Alphabet (GOOGL), Tesla (TSLA), Meta (META), OpenAI, and many others are investing hundreds of billions to reach AGI.
They can see it, feel it, and touch it. They’re so close. And the competition is so intense.
Which is exactly why they won’t stop.
The Bleeding Edge is the only free newsletter that delivers daily insights and information from the high-tech world as well as topics and trends relevant to investments.
The Bleeding Edge is the only free newsletter that delivers daily insights and information from the high-tech world as well as topics and trends relevant to investments.
