AI things update so fast and so much, with new models, new tools, and new terms every day. Just keeping up is impossible. This article talks about how I handle this: the core is one action, finding an anchor that won't change for ten years. With it, you have a standard to judge what to learn and what to skip, so information explosion doesn't become your anxiety.
- People who are constantly being washed by AI news, feeling like they are always chasing but never catching up
- People who have learned many tools but can't explain what they are actually accumulating
- People who want to create content or products, but their direction is always being pulled by the latest trends
- A set of self-questions to judge whether 'this thing will still be around in ten years'
- A prompt that can be directly pasted to AI, outsourcing information filtering
- A method to pick new things to learn: don't look at the tool itself, look at what pain points it solves for whom
The one who should be anxious is AI, not us
First, clarify what you are anxious about.
On the business level, the one who should be anxious is the AI company, not us. They are the ones who need to worry about being overtaken by someone next quarter.
On the knowledge level, unless you are an AI engineer, someone designing and training AI, no one understands AI better than AI. So, let AI help me crawl the knowledge. With so many papers appearing today, I won't read them all myself. I will ask it to help me see which ones are suitable for me.
The key is to clearly state what 'suitable for me' means. If you directly ask AI 'is this paper important', it can only give you a neutral summary. But if you tell it that you are judging based on one of these four things:
- Is this applicable to my industry directly?
- Is this able to optimize my existing knowledge framework?
- This can be used as a teaching example
- This can be written into a post
The things it filters out are completely different. Please analyze everything for me, and I'll look through it myself.
How to explain 'suitable for me' clearly
The previous four criteria are based on the premise that you already know what you're doing. This is the function of an anchor point.
Finding an anchor point has two conditions, and missing either one is not acceptable.
The thing you naturally do better than others and can do for a long time without getting bored.
Note that it's the part that won't change, not the latest part. Every era will definitely have something that remains unchanged for ten years, and that is where we can accumulate long-term.
Many people mistake 'professional skills they've learned through effort' as talent
This needs to be clarified especially, because many people misidentify it. Many people think a certain field is their professional area because they spent a lot of time learning and worked very hard to dive into it, so they are very good in that field. They might really be good, but that doesn't necessarily mean it's their talent.
Talent is actually the thing you do very easily, even something you were good at from a young age. Because you were good at it from a young age, you don't realize it has any value, nor do you think it's a skill. You learn quickly and easily, so easily that you think everyone is like that. That area is likely where you are truly good.
The intersection of these two is your anchor point.
What is my anchor point?
My anchor point is tacit knowledge distillation.
The verification method is simple. I ask myself one question: even without AI, is this thing important?
The answer is very important. The sense of a master craftsman, the judgment in business, and the rules in a manager's mind that are not written down are things that have always been the hardest to pass on in an organization without AI. With AI, it will only become more important, because what you need to teach AI is exactly these unrecorded judgments.
So this thing passes the test. It is not an opportunity brought by AI. It is a demand that existed before AI and was amplified after AI appeared.
The goal doesn't change, the method can always change.
The anchor point locks in the goal, not the method.
How can I practice tacit knowledge distillation? It can change with the times and with the situation. A few years ago, I used note-taking tools. Now I use agents and rule bases. In the future, it might be something else. All of these can change.
But there is one thing I am very sure about: no matter how the times change, I will still complete it.
After you clarify this, you will realize that you are actually just solving a problem. The tools just keep updating with the times. Human nature remains the same. We are just using new tools to solve problems that have always existed.
When a new tool comes out, you just need to ask yourself one question: can this tool help me do the original thing better? If the answer is no, skip it. You won't lose anything.
So what should I choose to learn?
After the goal is locked in, choosing a learning object has a standard.
I only focus on knowledge architecture. The simplest sources are two: what the top few companies' experts are talking about, and papers.
But you don't need to read those papers directly, because you won't know how to use them. You must have a knowledge architecture, and then directly ask AI to translate it into your knowledge architecture version: is this suitable for me, how should I use it, should I modify it or connect it to which part, or do I just need to refer to its mechanism for solving problems.
More importantly, look at the phenomenon of going viral from a different angle. A thing goes viral because it uses a very ingenious mechanism to solve the pain points that most people are currently facing. So instead of studying how to operate this tool, it's better to directly study three things:
What is something that was originally very painful, painful enough that people are willing to switch tools.
What they truly want as a result is not the function they mention verbally.
What design did it use to solve that pain, and this layer is what can be moved.
From the process of researching the pain points of the general public, you will discover trends. Trends are in the pain points of people, not in the feature list of the tool.
Two actions you can do right now.
Action one: validate your anchor point.
Take the thing you are currently investing in, and ask yourself three questions:
- Even without AI, is this thing important?
- Will this demand still exist in ten years?
- When I do this thing, am I doing it smoother and less easily getting bored than others?
If all three are affirmative, then it can be your anchor point. If one is negative, it may be a good opportunity, but not your anchor point, and the two are different.
Action two: outsource the filtering.
This can be directly pasted to your AI:
Conclusion: Accumulating too much is equivalent to having no focus.
I have seen fully automated tools that help people accumulate knowledge, and they do indeed record everything for you. However, after recording too much, there is also a lot of garbage, and in the end, you still have to manually determine what is important. If this is the case, why not start by developing the habit of recording what you personally find important.
To determine what is important, you need to first have an anchor point.
Therefore, the order is as follows: first have an anchor point, then have a standard; with a standard, you can ask AI to help filter; if filtering is possible, information overload will not become your anxiety.
First find an anchor point, do not first chase tools.
Further Reading
- Why Your AI Has to Start Over Every Time: Four Steps to Teaching AI After the anchor point is determined, this article explains how to transfer your judgment to AI.
- Information Lag and Market Forecasting: How to Be Ahead of the Market Without Being Too Early The anchor point determines what to learn, and this article determines when to learn, and who to sell it to.
- Why Your AI Adoption Isn't Working: From the Three-Layer Theory to the Organization's Wall Tools can always be changed, but the change should have a direction.