Tag Archives: prediction

Not emergence but convergence

Leave a reply

My post yesterday, ‘Immediate Emergence – Are we ready for this?’ I said, “Think about how fast we are going to see the emergence of intelligent ‘beings’ when we combine the brightest Artificial Intelligence with robotics…” and continued that with, “Within the next decade, you’ll order a robot, have it delivered to you, and out of the box it will be smarter than you, stronger than you, and have more mobility and dexterity than you.”

On the technology front, a new study, ‘Measuring AI Ability to Complete Long Tasks’ proposes: “measuring AI performance in terms of the length of tasks AI agents can complete. We show that this metric has been consistently exponentially increasing over the past 6 years, with a doubling time of around 7 months. Extrapolating this trend predicts that, in under five years, we will see AI agents that can independently complete a large fraction of software tasks that currently take humans days or weeks.”

More from the article:

“…by looking at historical data, we see that the length of tasks that state-of-the-art models can complete (with 50% probability) has increased dramatically over the last 6 years.”

“If we plot this on a logarithmic scale, we can see that the length of tasks models can complete is well predicted by an exponential trend, with a doubling time of around 7 months.”

And in conclusion:

“If the trend of the past 6 years continues to the end of this decade, frontier AI systems will be capable of autonomously carrying out month-long projects. This would come with enormous stakes, both in terms of potential benefits and potential risks.”

When I was reflecting on this yesterday, I was thinking about the emergence of new intelligent ‘beings’, and how quickly they will arrive. With information like this, plus the links to robotics improvements I shared, I’m feeling very confident that my prediction of super intelligent robots within the next decade is well within our reach.

But my focus was on these beings ‘emerging suddenly’. Now I’m realizing that we are already seeing dramatic improvements, but we aren’t suddenly going to see these brilliant robots. It’s going to be a fast but not a sudden transformation. We are going to see dumb-like-Siri models first, where we ask a request and it gives us related but useless follow up. For instance, the first time you say, “Get me a coffee,” to your robot butler Jeeves, you might get a bag of grounds delivered to you rather than a cup of coffee made the way you like it… without Jeeves asking you to clarify the task because you wanting a bag of coffee doesn’t make sense.

These relatively smart, yet still dumb AI robots are going to show up before the super intelligent ones do. So this isn’t really about a fast emergence, but rather it’s about convergence. It’s about robotics, AI intelligence, processing speed, and AI’s EQ (not just IQ) all advancing exponentially at the same time… With ‘benefits and potential risks.’

Questions will start to arise as these technologies converge, “How much power do we want to give these super intelligent ‘beings’? Will they have access to all of our conversations in front of them? Will they have purchasing power, access to our email, the ability to make and change our plans for us without asking? Will they help us raise our kids?

Not easy questions to answer, and with the convergence of all these technologies at once, not a long time to answer these tough questions either.

Immediate Emergence – Are we ready for this?

2 Replies

I have two daughters, both very bright, both with a lot of common sense. They work hard and have demonstrated that when they face a challenge they can both think critically and also be smart enough to ask for advice rather than make poor decisions… and like every other human being, they started out as needy blobs that 100% relied on their parents for everything. They couldn’t feed themselves or take care of themselves in any way, shape, or form. Their development took years.

Think about how fast we are going to see the emergence of intelligent ‘beings’ when we combine the brightest Artificial Intelligence with robotics like this and this. Within the next decade, you’ll order a robot, have it delivered to you, and out of the box it will be smarter than you, stronger than you, and have more mobility and dexterity than you.

Are we ready for this?

We aren’t developing progressively smarter children, we are building machines that can outthink and outperform us in many aspects.

“But they won’t have the wisdom of experience.”

Actually, we are already working on that, “Microsoft and Swiss startup Inait announced a partnership to develop AI models inspired by mammalian brains… The technology promises a key advantage: unlike conventional AI systems, it’s designed to learn from real experiences rather than just existing data.” Add to this the Nvidia Omniverse where robots can do millions of iterations and practice runs in a virtual environment with real world physics, and these mobile, agile, thinking, intelligent robots are going to be immediately out-of-the-box super beings.

I don’t think we are ready for what’s coming. I think the immediate emergence of super intelligent, agile robots, who can learn, adapt, and both mentality and physically outperform us, that we will see in the next decade, will be so transformative that we will need to rethink everything: work, the economy, politics (and war), and even relationships. This will drastically disrupt the way we live our lives, the way we engage and interact with each other and with these new, intelligent beings. We aren’t building children that will need years of training, we are building the smartest, most agile beings the world has ever seen.

The false AGI threshold

Leave a reply

I think a very conservative prediction would be that we will see Artificial General Intelligence (AGI) in the next 5 years. In other words, there will be AI agents, working with recursive self-improvement, that will learn how to do new tasks outside the realm of its training, faster than a human could. But when this actually happens will be open for debate.

The reason for this is that there isn’t going to be some magical threshold that AGI suddenly passes, but rather a continuous moving of the bar that defines what AGI is. There will be a working definition of AGI that puts up an artificial threshold, then an AGI model will achieve that definition and experts will admit that this model surpasses that definition, but will still think the model lacks some sufficient skills or expected outputs to truly call it AGI.

And so over the next 5 years or so we will have far more sophisticated AI models, all doing things that today we would define as AGI but will not meet the newest definition. The thing is that these moving goal posts will not be adjusted incrementally but rather exponentially. What that means is that the newer definition of AGI is going to include significantly greater output expectations. Then looking back, we will be hard pressed to say “this model’ was the one or ‘this day’ was the day that we achieved AGI.

Sometime soon there will be an AI model put out into the world that will build its own AI agent that starts a billion dollar company without the aid of a human… but that might happen even before consensus that AGI has been achieved. There will be an AI agent that costs lives or endangers lives with its decisions made in the real world, but that too might happen before consensus that AGI has been achieved.

Because the goal posts will keep moving while the technology is on an exponential curve, we are not going to have a magic threshold day when AGI occurred. Instead, in less than 5 years, well before 2030, we are going to be looking back in amazement wondering when we passed the threshold? But make no mistake, that’s going to happen and we don’t have an on/off switch to stop it when it does. This is both exciting and scary.

A prediction nears reality

Leave a reply

Andrew Wilkinson said on X:

“Just watched my 5-year-old son chat with ChatGPT advanced voice mode for over 45 minutes.

It started with a question about how cars were made.

It explained it in a way that he could understand.

He started peppering it with questions.

Then he told it about his teacher, and that he was learning to count.

ChatGPT started quizzing him on counting, and egging him on, making it into a game.

He was laughing and having a blast, and it (obviously) never lost patience with him.

I think this is going to be revolutionary. The essentially free, infinitely patient, super genius teacher that calibrates itself perfectly to your kid’s learning style and pace.

Excited about the future.”

– – –

I remember visiting my uncle back when I was in university. The year was somewhere around 1988-90. So, at least 34 years ago. We were talking about the future and Joe Truss explained to me what learning would be like in the coming age of computers.

He said, (loosely paraphrased, this was a long time ago):

‘In the future we will have virtual teachers that will be able to teach us exactly what we want to know in exactly the format we need to learn best. You want to learn about relativity? How about learning from Einstein himself? You’ll see him in front of you like he is real. And he will not just lecture you, he will react to your questions and even bio-feedback. You look puzzled, he might ask a question. He sees you looking up and to the left, which he knows means you are trying to visualize something, and so he changes his lesson to provide an image. He will be a teacher personalized to any and all of your learning needs.’

We aren’t quite there yet, but the exchange Andrew Wilkinson’s son had with ChatGPT, and the work being done in virtual and augmented reality, suggest that Joe’s prediction is finally coming into being.

I too am excited about the future, and more specifically, the future of learning.

It’s already here!

Just yesterday morning I wrote:

Robots will be smarter, stronger, and faster than humans not after years of programming, but simply after the suggestion that the robot try something new. Where do I think this is going, and how soon will we see it? I think Arther C. Clarke was right… the most daring prophecies seem laughably conservative.

Then last night I found this post by Zain Khan on LinkedIn:

🚨 BREAKING: OpenAI just made intelligent robots a reality

It’s called Figure 01 and it’s built by OpenAI and robotics company Figure:

  • It’s powered by an AI model built by OpenAI
  • It can hear and speak naturally
  • It can understand commands, plan, and carry out physical actions

Watch the video below to see how realistic it’s speech and movement abilities are. The ability to handle objects so delicately is stunning.

Intelligent robots aren’t a decade away. They’re going to be here any day now.

This video, shared in the post, is mind-blowingly impressive!

This is just the beginning… we are moving exponentially fast into a future that is hard to imagine. Last week I would have guessed we were 5-10 years away from this, and it’s already here! Where will we really be with AI robotics 5 years from now?

(Whatever you just guessed is probably laughably conservative.)

The most daring prophecies

In the early 1950’s Arthur C. Clarke said,

“If we have learned one thing from the history of invention and discovery, it is that, in the long run — and often in the short one — the most daring prophecies seem laughably conservative.”

As humans we don’t understand exponential growth. The well known wheat or rice on a chessboard problem is a perfect example:

If a chessboard were to have wheat placed upon each square such that one grain were placed on the first square, two on the second, four on the third, and so on (doubling the number of grains on each subsequent square), how many grains of wheat would be on the chessboard at the finish?

The answer: 264−1 or 18,446,744,073,709,551,615… which is over 2,000 times the annual world production of wheat.

All this to say that we are ill-prepared to understand how quickly AI and robotics are going to change our world.

1. Robots are being trained to interact with the world through verbal commands. They used to be trained to do specific tasks like ‘find one of a set of items in a bin and pick it up’. While the robot was sorting, it was only sorting specific items it was trained to do. Now, there are robots that sense and interpret the world around them.

“The chatbot can discuss the items it sees—but also manipulate them. When WIRED suggests Chen ask it to grab a piece of fruit, the arm reaches down, gently grasps the apple, and then moves it to another bin nearby.

This hands-on chatbot is a step toward giving robots the kind of general and flexible capabilities exhibited by programs like ChatGPT. There is hope that AI could finally fix the long-standing difficulty of programming robots and having them do more than a narrow set of chores.”

The article goes on to say,

“The model has also shown it can learn to control similar hardware not in its training data. With further training, this might even mean that the same general model could operate a humanoid robot.”

2. Robot learning is becoming more generalized: ‘Eureka! NVIDIA Research Breakthrough Puts New Spin on Robot Learning’.

“A new AI agent developed by NVIDIA Researchthat can teach robots complex skills has trained a robotic hand to perform rapid pen-spinning tricks — for the first time as well as a human can…

Eureka has also taught robots to open drawers and cabinets, toss and catch balls, and manipulate scissors, among other tasks.

The Eureka research, published today, includes a paper and the project’s AI algorithms, which developers can experiment with using NVIDIA Isaac Gym, a physics simulation reference application for reinforcement learning research. Isaac Gym is built on NVIDIA Omniverse, a development platform for building 3D tools and applications based on the OpenUSD framework. Eureka itself is powered by the GPT-4 large language model.”

3. Put these ideas together then fast forward the training exponentially. We have robots that understand what we are asking them, which are trained and positively reinforced in a virtual physics lab. These robots are practicing how to do a new task before actually doing it… not practicing a few times, or even a few thousand times, but actually doing millions of practice simulations in seconds. Just like the chess bots that learned to play chess by playing itself millions of times, we will have robots where we ask them to do a task and they ‘rehearse’ it over and over again in a simulator, then do the task for the first time as if it had already done it perfectly thousands of times.

In our brains, we think about learning a new task as a clunky, slow experience. Learning takes time. When a robot can think and interact in our world while simultaneously rehearsing new tasks millions of times virtually in the blink of an eye, we will see them leap forward in capabilities at a rate that will be hard to comprehend.

Robots will be smarter, stronger, and faster than humans not after years of programming, but simply after the suggestion that the robot try something new. Where do I think this is going, and how soon will we see it? I think Arther C. Clarke was right…

…the most daring prophecies seem laughably conservative.

The future is now

I’ve shared Chat GPT a couple times (1, 2), and I really think that tools like this are going to create a massive shift in jobs, education, and creativity. It can been seen as both scary and exciting.

On another front, scientists have achieved ‘ignition’ in a nuclear fusion test. This is the creation of a fusion reaction where the energy output is greater than the energy input. For over a decade this was an unachievable goal, any reaction created required so much energy to produce that the costs were greater than the returns.

If you showed someone from 1995 the technology we had 25 years later in 2020, they would be impressed and amazed. I’m reminded of Arthur C. Clarke’s quote, “Any sufficiently advanced technology is indistinguishable from magic.” But what I’m seeing right now seems like magic.

I think the leaps in technology and ingenuity that will happen in the next 25 years will so far exceed what we saw happen in the last 25 years that it will feel more like we are 50 years in the future rather than just 25. So, buckle up and get ready for a wild ride into the future… it’s just getting started now!

6 weeks later

I said this in a post, ‘If or When?‘ six weeks and 2 days ago, with respect to Omicron and coronavirus:

I’m afraid my crystal ball gazing conjures no clear images of what’s to come in the short term. What’s my best guess? Businesses and maybe even schools closing from staffing shortages; Tougher restrictions in public (indoor) spaces; and a lot of people self-or-family isolating.

What’s my hope? Less and less hospitalization with more and more mild cases that look like a ‘regular’ flu.

We didn’t have any full school closures in our district but 2 weeks after this post we had a 3-day period where many schools were short staffed and support teachers, and vice principals, were reallocated to help fill the needs. Tighter restrictions didn’t happen. And many people did get mild versions of Omicron although so few of them were tested, because of shortages of tests and long lineups, that I think the number of cases reported was probably underreported by a factor greater than 10.

Now, there are no issues with staffing, and there is an abundance of rapid tests available, and I think things are much better. The only big news here in Canada is a protest against mandates and vaccines that is being held in Ottawa, with smaller versions showing up in other major cities. I’m not going to link to any of the reports because I don’t believe they deserve the attention.

I find it sad how many people are misguided by these protests, and think they are somehow fighting for freedom by blockading people and making others less free? It’s sad how messages of hate and ignorance are part of this convoy. It’s ironic that when we are on the brink of restrictions being lifted, thanks to a vast majority of Canadians doing their part in slowing the pandemic by following the mandates and getting vaccinated, these unvaccinated, anti-mask protesters feel like they are somehow doing something to better the world they live in.

6 weeks after my original post I still haven’t had covid, but I’d say this has been as much about luck as any other reason. Many people around me have been equally as cautious as me and have contracted it. All these people have had versions that are mild, or at least not bad enough to go to the hospital. And I still wonder if for me it’s a question of if or when? But in this 6 weeks I’ve seen things getting much better. I wonder if mask restrictions will continue in schools after March Break? I wonder if there will be crowd restrictions at theatres and sports events, or if restaurants will still require vaccine passports?

Unless another contagious and more deadly strain start spreading, I think we might see a lot less restrictions in the coming months… But it won’t be because of this ‘freedom rally’ that’s happening. On the contrary, it will literally be despite them and their anti-community stance, not doing their part to help reduce the concerns and spread of coronavirus.

The real numbers

In the last week and a half I’m aware of 18 people whom I personally know have caught covid-19, likely the Omicron variant. But of these 18, (half of which are 2 families), I think only one of them is reported. Others sat on 811 for more than 2 hours before giving up on the call. Others have mild symptoms, and here in BC they are asking people with mild symptoms not to get tested, because testing locations are being overwhelmed.

Extrapolating on these numbers, even if I overestimated the amount of cases in BC to be 1/3 reported (that would be 6/18 rather than 1/18), then the real number of new covid cases is a massive number.

This is surprisingly a good thing. First of all, our hospitals are not overwhelmed with serious cases. Secondly, at this rate of infection, we are moving quickly to heard immunity.

Well, that’s my hope anyway. I am only working with my own anecdotal evidence, and coronavirus has kicked my predictions in the butt more than once. Still, it’s good to look forward and see a little light at the end of the tunnel.

Maybe by spring everything will be open and masks will be optional. Maybe. Still, it feels good to have a little optimism right now. Even if we really don’t know the real numbers, I think they are looking good… Time will tell.

Still a hopeful timeline

I’ve been saying since before Christmas 2020 that, “Things will start to return to normal by January 2022.”

With the Delta variant being very contagious, and new variants like B.1.621.1 showing up on the US east coast, I don’t think school is going to start in September anywhere near as normal as we might have thought it would in June. Going into a busy coffee shop this morning, I was the only one wearing a mask, including the employees. The only covid precautions I saw were hand sanitizer at the entrance and the barista using tongs to put a straw into a cold drink.

I think new variants are going to give us a bit of a spanking, then we’ll smarten up and re-tighten up our precautions enough to slowly move in the right direction.

Although it may have seemed a bit pessimistic at the time, I am now hopefully optimistic that my prediction can be met. Despite the evidence of a 4th wave, it could be smaller, less alarming, less dangerous (thanks to the vaccine), and better managed. The ‘better managed’ is the key. We move forward with caution, or we move backwards.

Even if you are double vaccinated, wear a mask in public indoor spaces. Let’s keep these variants away.