Learning Has Changed Forever
Ignore it and get left behind.
Many years ago Google began to kill learning. It wasn’t intentional and I’m sure most would argue against this, but their entire design is meant to find answers to singular specific questions before making a hand off. It was never meant to teach or provide a complete guide to learning about a topic.
There is an unspoken belief that google provides truth. We search for anything we wonder and expect the results to appear.
But, it’s a myth born from convenience. If anything Google should be feared?—?after all, it knows everything we care about and only has to show a few websites in return. They also just make it up sometimes.
Google is great for the following:
But, they’re horrible at the inevitable next questions. If we learn about something new, there will be an almost universal need to understand related areas. Google fakes this by providing summary level websites or opinion pieces.
It wasn’t until I began using AI regularly that I really began to see it. Almost like believing Twitter provides accurate news and nothing else. Yes, you can learn by using Google, but you’re much more likely to get distracted.
The point here is that the technology we use affects how we try to solve problems. It literally affects how we think. Since Google has dominated for a decade plus, most people struggle to learn new topics because they don’t know where to start. An interest arises organically, then we ask google, and then we either go down a rabbit hole of subjective nonsense or we move on?—?curiosity answered.
Let’s talk software development.
For many years, if a developer was asked to do something new, they would likely go to Google and begin searching goals. Very quickly, they’d be forced to choose a path. Do they navigate to some school resource to learn the entirety of something before even beginning, or do they start the task and scour the realm of stack overflow and various forums to fill in the specific blanks they’re wondering about. Perhaps if they’re lucky they’ll have access to code examples from their company’s source, or maybe a similar GitHub repository.
In the era of Build Fast And Break Things (Facebook), I’d wager most developers take the second approach. It’s understandable because it forces you into a situation where results can quickly be seen and an illusion of learning is created. If you do it enough, real learning, simply from repetition, could even begin to unfold.
But, the end result is vague. Yes, things get built, but there are holes as to why stuff works. Progression falls apart and innovation stalls.
Enter AI
Most articles that surround AI are about its ability to write well or about how some company now offers AI for a specific purpose. There’s a creative aspect that gives incredible speed to content creators, and I’m sure 95%? of what’s been written online today came from a bot somewhere. As an aside, this article doesn’t use AI if anyone was wondering.
What I very rarely hear about though is how people use AI as an extension of themselves to build and learn things they never thought possible. In truth, it’s probably not marketed because it’s difficult to explain. “Hey, I’m now 10x the developer that I was last week because of a robot machine.” I don’t think anyone would be lining up to 10x their employee’s salary.
But, I don’t know why not. Developers that use AI well are easily this much better, if not 100x.
For early developers, it can be very difficult to understand how things work, especially the syntax. They might be able to design a system in their head, but when it comes to creating the actual code and to understand how pieces work, many walls appear.
These are the people who need to use AI the most.
Imagine going through school with a person standing next to you who knows everything, who has infinite patience, and who can teach you at your own pace and in a way that makes sense specifically to you.
Take any project you may be considering. Think it through until you get to the first hurdle. Hell, even start from the very first idea, and ask AI what the ideal approach to doing XYZ is. At first it will provide an overview, or maybe several options. But, if you keep talking, you can delve into long term strategy and what makes sense based on a wide range of other factors.
A Quick Comparison Between Learning With Google Vs ChatGPT
My son has been learning with Microsoft’s MakeCode Arcade for awhile now and I was curious what should come next.
So, I searched Google and ChatGPT to answer the following question:
“what is the next step to take after using makecode for a while?”
领英推荐
If you click the link above, you’ll see that it’s a clear collection of websites with text using similar phrasing. If someone had once written an article (opinion) using the exact phrasing of my search, then I’m sure I would see those results and might even walk away carrying whatever opinion that writer had. This is the way of Google.
But, here is ChatGPT’s response:
As you’ll see this is exponentially more clear. OpenAI’s response provides a concise list of possibilities and possible paths. For any one of these topics, I could then ask further questions and suddenly I am learning compared to wandering.
Every child in the world should have access to this. Yes, there will be bias and the source is the internet, which is filled with many wrong opinions, but there is an intelligence that’s undeniable. It’s also not entirely source driven and I can’t explain this.
Recently, I was struggling with implementing an OAuth process. I was stuck on this syntax issue and the vendor’s API guide was not clear. After giving the docs to chatGPT, it then gave me the correct syntax. It still does not make sense. What it output was not in the guide and what it had produced before I gave it the guide was not right. The mind in the machine…
As you go, you will likely need to fill in the blanks for AI as it relates to your goals. Once you settle on how to move forward, you can focus on specific tasks and use AI to go from the very high level to the extremely detailed. For development especially, AI has access to so much content such as API docs, that it can quite literally write the code you need. And, if it doesn’t know everything, you can actually feed it details such as highly technical documentation.
For those unaware, AI has a context limit. This means that there is a limit to what it can keep in its “mind” at a time. People explain this in differently ways, but it’s really quite fascinating. Think about it as if there’s an incredibly smart person who can help you with anything, but they can’t quite remember the sweeping, high level aspects, you and others just naturally store and refer back to during the planning period. An example would be — if you need to go to the grocery store to get food for dinner, but get distracted by ingredients and a conversation with a friend at the store. If this was AI, it might just forget it was making dinner with the food or that it had a family to feed.
To get around this, AI works best with small steps or outlines.
If I was in college right now, I would drop out immediately.
There is not rational reason to maintain the traditional methods of learning anymore. They are so incredibly outdated at this point that it’s embarrassing, and I don’t understand why some entire degree programs haven’t been charged with fraud. The only feasible argument at this point is the network effect of who you might meet.
Yes, this may be extreme sounding, but unless there some regulatory body you must appease, there is zero reason to learn via any course or instructor in a way that’s generally meant for many people.
Instead, use AI to create a teacher that explains things in your way.
Even after having used AI regularly for many months now, I’m still amazed about how often I stop asking questions when the solution is still unclear. My only explanation is that we have been conditioned to feel like we have to endure from minimal teaching and to figure it out ourselves. With people and time constraints, this makes sense. No one can hover over your shoulder answering every question you have.
It’s really very sad if you think about it, and perhaps its just me, but I believe there is a tendency for people to learn something from someone and then potentially …pretend… they understand before they really do. After all, who wants to get someone to explain something 3, 4, or more times? There’s a general impatience that flows through society and everyone knows about it. For many, I imagine this is why they resist learning. If it can’t be figured out from the brief instruction they receive, then they feel bad or embarrassed and change direction.
But AI has infinite patience and will completely change the expectation of learning.
For the first time in human history, people will be able to continually ask questions.
I still forget this, even after months of usage and accidentally slip back into the old ways. I’ll need to write some block of code and knowing that I can, I’ll start to wander through old projects looking for pieces that I’ll need. And then, I’ll remember, and AI will help bring together the pieces without the use of so much time.
To some degree, it does feel like cheating. I really don’t know why though. If, for example, I was to have a job interview and someone asked me if I could code something, I feel confident in saying yes now to 100x more scenarios. To argue that people are still not even in the realm of understanding AI’s impact, I bet many would consider this dishonest, or expect me to magically know everything I might say I could do.
But who’s better, a developer who can specifically speak to the intricacies of a complex algorithmic sort calculation, but lacks broader business knowledge, or an analyst who’s capable of describing with words (prompts) exactly what’s needed step by step and then using AI, actually build it?
I assume the cab drivers of the past scoffed at the advent of GPS as well, claiming their internal knowledge would reign supreme.
For any manager or business owner who isn’t using AI, they are missing out on an incredibly opportunity to give their employees a rocket ship for learning and understanding new technology.
Best of luck. There is a new paradigm that will take time but only because humanity will slow it down.
Our children will get it.
Go open ChatGPT in a browser and leave it open all day. Pay the $20 — it’s the greatest $20 you’ll ever spend. And, try to keep up with what’s coming because no one will keep up, but especially not the people who don’t even try.