It’s hard to believe I haven’t written about AI since September! But then, in my defense, a lot has been happening – life, the election, you name it. Of course, I’ve continued reading about AI every day because, like AIs themselves, the AI market never sleeps. And recently there’s been a lot of AI news worth talking about.
Note: If you need a refresher on AI jargon, feel free to visit these earlier posts: “Tales From the Dark Side. AGI is Coming. Are We Ready?”, “Some AI Terms Everyone Should Know.”
In the past month alone, AI developments have pushed us closer to what some might call science fiction. AGI—artificial general intelligence —the dream (or fear) of machines with human-like intelligence—is peeking over the horizon. Can you see it?
This reality raises an important question, one I fear we already know the answer to: “Are we ready for AGI?” By “we,” I mean society here in the USA. And the answer? “No. We’re not even close.”
This question, “Are we ready?” is one I’ve pondered and discussed so often, I’m honestly tired of it. So, instead, I’d like to reframe it into something more personal and maybe more pertinent: “Are you ready for AGI?”
Before you answer, let’s catch up on some of the most significant AI developments from the past few months. Because, trust me, there’s been too much happening to fit everything into one blog post.

The Big Picture: What Happened To AI This Month?
You could search for the answer to this question yourself and the result would be a long list of achievements, so I’ll just focus on a few that I think are worth noting.
OpenAI’s o1 Model: Deep Thoughts (the “Jack Handey” Approach to AI)
OpenAI introduced their o1 model last September, and this month it’s now fully released to subscribers. This model represents a shift toward making AI capable of “long thinking.” (This reminded me of Saturday Night Live’s “Deep Thoughts by Jack Handey.” I couldn’t resist!)
Unlike traditional models that spit out rapid responses, o1 engages in extended deliberation, enhanced reasoning abilities, and reducing errors in complex tasks. This approach draws inspiration from human cognitive processes, aiming to emulate deeper, more thoughtful problem-solving. Be aware that although the model’s responses reflect deeper reasoning processes, they may take slightly longer to generate.” Sure, it’s a bit slower, but it’s fascinating to watch. And let’s not forget:
The AIs we use today are the very worst they will ever be! AI development is still happening exponentially, so their thought processes will get faster and smarter over time.
OpenAI’s Vision Capability: Seeing Changes Everything!
Another game-changing release from OpenAI—this one for mobile devices—is vision capability. Using your phone, GPT can now “see,” analyze, and interpret visual information. Specifically, this means it can:
- Identify objects in photos.
- Analyze charts, graphs, or Excel screenshots.
- Decode text-heavy visuals like error messages, signs, or product labels.
I subscribe to ChatGPT, so I can attest to this new capability personally. For example:
- I recently asked GPT to help organize a Lightroom keyword list. I used my phone’s camera to “show” GPT the list on my PC screen, and it suggested a better system.
- This morning, I had it explain a feature I was curious about on a product image from Amazon—just by talking to GPT on my phone. No typing required!
- I used ChatGPT at Costco to give me wine reviews before buying! Just turn on your camera and show GPT a bottle of wine you’re considering.
AGI Implication: The “long thinking” approach of o1 and the vision capability are both milestones on the road to AGI. Humans interact with the world through multiple senses, and multimodal AI—the ability to combine text, visuals, and more—is essential for creating a versatile, human-like intelligence.
The Agentic AI Era: Real AI Agents are Here!
Agentic AI refers to systems capable of autonomous action—making decisions and executing tasks without constant human intervention. It’s an idea that’s been brewing for decades. I remember HP’s 1980s concept video, “1995,” showcasing a vision for AI agents working on our behalf. All I can say is better late than never!
Google’s Gemini 2.0:
Google unveiled Gemini 2.0, their latest advanced AI designed to interpret images, perform tasks, and act autonomously in response to user needs. Google’s goal is clear: AI systems that anticipate user needs and take action before you even ask.
AGI Implication: Gemini 2.0 enhances user experience across Google’s ecosystem, indicating an effort to develop more intuitive and proactive AI assistants. Being able to interpret visual images will be an important feature of the AGI.
Note: Gemini was actually my first AI. I gave up on it due to its periodic “hallucination storms,” but I have a soft spot for it. It’s exciting to see it finally competing with ChatGPT in some areas.
Microsoft Copilot: AI That Gets Work Done
Microsoft has been working for years to bring AI agents into the mainstream, and now they’re making it happen with Copilot. It comes standard on Windows PCs, and honestly, it’s gone from “meh” to “wow.”
Copilot Highlights:
- Copilot Vision: The AI can now “see” your PC screen and interact with it—so you don’t have to describe problems. Just say, “Sort columns A and B for me,” and voila—it’s done.
- Copilot Studio: Build and customize AI agents for specific workflows (released Nov 2024).
- Recall AI: Take snapshots of screen activity for easy referencing (Dec 2024).
- Copilot Actions: Automate repetitive tasks with simple prompts.
Implications for AGI: Copilot Vision lets the AI analyze your PC screen contextually and assist with tasks like sorting or navigation. The ability to “see” the world an interact with the world autonomously are imperatives for AGI. Granted, the world Copilot sees is limited to your PC screen, but as you read below, this is being expanded by companies like OpenAI.
OpenAI’s Plan to Sell PhD-level AIs
Discussed recently by OpenAI’s Chief Financial Officer, Sarah Friar, OpenAI is considering offering a new service to enterprise customers. This would include “hiring” into their organization one or more PhD-level AI employees, each specialized in whatever area or field required by the enterprise e.g., medical, legal, etc. The price? $2000 per month. Quite the bargain, I would say! Quite possibly this is one of the “job killers” some people have warned us about!
Closing Thoughts on AI Development
I could go on listing new releases and features, but I’m going to stop here. I think I’ve given a few highlights that illustrate the advancements being made.
These developments illustrate the dynamic nature of AI research and its accelerating trajectory towards more advanced and integrated systems. The convergence of these technologies suggests that AGI may be just on the horizon, prompting both excitement and caution as we navigate this transformative era.

Is AGI Coming? Or is it here already?
You’ve heard me say it many times, “AGI is coming!” But recently more people are saying, AGI is already here. Wait, what? Is this correct? This is something worth considering.
If you remember, the definition of artificial general intelligence refers to a machine that can perform any intellectual task that a human can do at a similar level of capability. In other words, an AGI would have:
- Human-equivalent reasoning, problem-solving, and adaptability across diverse domains.
- The ability to learn new tasks without specialized programming, much like humans can.
Given this widely acceptable definition, I think we can safely say that no AI released to the public has yet demonstrated both of these capabilities. I say “released to the public,” because none of us knows what these companies may have behind closed doors. And I say that “has not yet demonstrated” specifically because #2 above has not yet been achieved.
#1 above, on the other hand, is a done deal. I say this because I think today’s AIs like ChatGPT 4o and o1, as well as others, clearly demonstrate their reasoning, problem-solving and adaptability every time we use them. Seriously, I can honestly say that so far, ChatGPT has demonstrated to me that it is smarter than me in anything I have asked it. I know, I’m probably not the best test, but this is true for most everyone I know. It knows so much about everything! And remember, the qualification to be an AGI is not to demonstrate that it is smarter than me in generally everything, but rather that it is “as smart” as me. And it does this, in spades, every day. It’s humbling.
You may have noticed that you don’t hear a lot about the Turing Test anymore. There’s a reason for this. This is the test where an AI and a human are asked the same questions and a human tries to decide who is human and who is not. Today’s AIs easily pass it for the most part.
As for #2 above, work continues to advance the autonomy of AI systems and efforts to get them to the point where they can learn on their own, but alas, we are not quite there…yet. Almost, but not quite.
It’d be good to remember too, that AIs don’t require “specialized programming.” That’s because AIs are not “programmed” to do what they do. Instead, they are trained to do what they do. That’s a big difference, by the way! Specialized programming infers that a human was involved with coding instructions telling an AI what it should do, and this is not how today’s AIs work! Training an AI means that the AI is learning from the data it is fed rather than being told what to do programmatically.
So, how do we get AIs to learn new tasks on their own through training? Do we simply throw more data at them? A good guess, but no, that won’t work. Granted, exposing AIs to larger and larger datasets will improve their performance, but to get them to think better requires something more. In fact, getting AIs to move beyond where they are now intellectually requires a combination of identifying new approaches, creating new breakthroughs in AI architecture, and receiving further inspiration from how we humans learn. Things like:
- Meta-Learning: Training AI to “learn how to learn.”
- The AI is trained on a variety of tasks rather than one. This enables it to develop a general method for problem-solving rather than relying on task-specific patterns.
- Reinforcement learning with curiosity.
- Current AI systems require clear rewards or tasks. However, humans often learn simply because of curiosity—we explore and learn new things without immediate incentives.
- To address this shortfall, researchers are building AI systems with:
- Intrinsic Motivation: An AI assigns value to exploring new environments, learning skills, and discovering patterns, even if there’s no external “reward.”
- Reinforcement Learning (RL): The AI interacts with its environment, experimenting and learning by trial and error.
- Multimodal Learning.
- Humans learn through multiple senses—sight, sound, touch, and more. AI systems are now being trained to integrate text, images, audio, and video into a single learning process. This creates a richer understanding of the world.
- How it helps autonomous learning:
- The AI can generalize more effectively because it “sees” and “hears” context, not just raw data.
- Combining multiple data sources mimics how humans make sense of the world.
This is why I don’t think AGI is here yet. I think people who believe this are focused on #1 above, and ignoring #2. Periodically, these folks do, however, lodge a valid complaint. They accuse the industry of “moving the goal post” for AGI. That’s because as soon as a smarter AI model gets released, suddenly it’s not smart enough for an AGI, and so they demand an even smarter model before the AGI is achieved. They’re not wrong, and it is somewhat irritating.

Final Thoughts
AGI may not be here yet, and I predict it will arrive in the next 1-2 years. Without doubt, we are all witnessing a turning point in human history. AI is not just a tool—it’s a partner, collaborator, and for some, even a teacher. If you’re not exploring it yet, now is the time!

Why You Should Start Using AI
Whether you like it or not, AIs are going to be a part of our lives. In fact, this is already happening. With regard to sticking a toe into the rather murky AI ocean, there is no better time to do it than now! It’s not hard to do. My own experience with LLMs began back in January, but seems much longer since so much has changed since then. I am easily at the point right now where I would be rather upset if my access to AIs was suddenly cut off. I recommend you strongly consider doing so yourself.
The AI tools available today are accessible, practical, and game-changing. Tools like ChatGPT can help you write, brainstorm, learn new skills, and even solve complex problems:
- For Writers: Generate ideas, organize drafts, or refine language.
- For Professionals: Automate repetitive tasks, analyze data, or brainstorm business strategies.
- For Curious Minds: Learn about philosophy, history, or science through conversational AI.
Please be reassured and remember – you don’t need to be a tech expert to start using AI! For example, if you can do a Google search, you can use an AI (try the same search with the free Perplexity AI). Just take that first small step—open up a tool like free ChatGPT and ask it anything.

Some Thoughts on ASI
There has been some recent media buzz about ASI (artificial super intelligence) and I want to share my own thoughts on the subject, as they have evolved on the subject.
Whereas I said above that the AGI, when it arrives, will be as smart as any human in most things, the ASI, by definition, will be much smarter than any human who ever lived in every subject. In fact, the ASI will far exceed human cognitive abilities, essentially making us humans, for the first time in history, the second-smartest species on the planet. That’s right, I said “species.” You object? I get it. Technically, any species, by definition, has to be biological, but I think there are enough similarities to make it a worthwhile adjective for AGI and ASI. If you don’t like this, we’ll just have to agree to disagree.
Predictions as to when the ASI will arrive vary widely. On this, I think it is important to remember that the AGI is like any cloud-based app in the sense that it exists in a cloud. It’s comprised of 0’s and 1’s just like your Word application. And like a Word document, you can have multiple copies of an AI, or an AGI.
For example, I am currently using ChatGPT 4o on both my phone and my PC. This has created two “instances” of ChatGPT for me, one on each device. They are separate and not connected, but both are ChatGPt 4o. Now, imagine that I want to create 10, 100, 1,000, or 1,000,000 copies of an AI. I can do that as long as I have the resources for it! These AIs take a lot of resources, like compute and power. There’s a good reason why companies like Microsoft are building nuclear power generators for their new data centers!
But, unlike 1,000 copies of a Word document, 1,000 copies of an AGI would be capable of working together to solve problems in virtually any field. Think about it. Bump up that number to 100,000, or 1 million. Now, think about it, what kind of problem will we first ask millions of AGIs to solve for us? I think I know. Do you?
My guess is that the first biggest problem we will be asking the AGI to solve is to put it to work creating the ASI. Imagine millions of PhD-level AIs focusing on developing a super-intelligent AI. Analyzing, developing, testing, and applying new improvements and capabilities of an AGI at the speed of light. What could easily have taken humans a century to develop, an army of AGIs will be able to do in what, 1 or 2 years? Yes, it’s possible. This is one reason why the AGI is so important and why every company working on large language models is working towards the goal of achieving the AGI.
Whereas we will no doubt, be amazed by the intelligence and capabilities of an AGI, the ASI, when it arrives, will be on another level altogether. Do you remember Arthur C. Clark’s quote: “Any sufficiently advanced technology is indistinguishable from magic.” When I think of the ASI, I think of magic. All of the problems facing mankind today will be solvable with the ASI. Of course, although the ASI may come up with solutions for us, we may choose not to implement them for a variety of reasons. But the ASI can help us work out those details. Cure for cancer? Yes, possible. Climate change problems? Yes, possible. A path to our solar system, our galaxy and beyond? Yes. These are within reach via the ASI. It almost goes without saying, but I’ll say it and remind us all that ensuring these advancements are safe and aligned with human values will be critical as we move forward.
I know it sounds like fantasy. But I am here to tell you, it isn’t fantasy. It’s possible. it’s inevitable. And it’s in our future.
0 Comments