AI is a floor increase, not a ceiling booster

A reformed learning curve

Before AI, students were confronted with a matching problem: learning resources must be made with a target group in mind. This means as a consumer, learning resources were suboptimal attacks for you:

• You are a newbie $topic_of_interestBut have knowledge in related subject $related_topic. But finding learning resources that teach $topic_of_interest in terms of $related_topic Is difficult.
• To learn effectively $topic_of_interestYou really have to learn for required skills $prereq_skill. But as a beginner you don’t know you really have to learn $prereq_skill Before we learn $topic_of_interest.
• You have basic knowledge of $topic_of_interestbut have plained and have difficulty finding the right resources for $intermediate_sticking_point

Roughly sees the acquisition of control in a skill over time:

What makes learning with AI groundbreaking is that it is possible Meet you at your skill level. Now an AI can answer questions directly at your level of understanding and even work for you. This changes the learning curve:

Mastery: Still difficult!

Experts in a field are usually more skeptical about AI. By Hacker News:

[AI is] superficial. The deeper I go, the less it seems to be useful. This happens quickly for me. God also forbids you to investigate a complex and possibly controversial subject and you want it to find renowned sources or in particular academic.

This is very logical, when considering the data on which AI has been trained. If the training focus of an AI has abundant training data on a subject that says everything more or less the same, it will be good at synthesizing it in output. If the subject is too advanced, there will be much less training data for the model. If the subject is controversial, the training data will contain examples that say opposing things. Mastery therefore remains difficult.

Cheat

The introduction of OpenAI Study Mode Hints on a problem: instead of having an AI learn, you can just ask for the answer. This means that cheaters will offer plates that the AI can offer at every level:

Cheaters will not bloom here in the long term!

The impact of the changed learning curve

Technological change is an ecosystem change: there are winners and losers, unevenly divided. For AI, the impact level is determined by The amount of control required to make an impactful product:

Coding: a blessing for management, less for large code bases

When we try to cod something, engineering managers often come across a problem: they know the principles of good software, they know what bad software looks like, but they don’t know how to use it $framework_foo. This has made it especially difficult for an example of a backend EM to build an iPhone app in their spare time.

With AI they can quickly learn the basics and make simple apps run. They can then use their existing knowledge Refine it in a workable product. AI is the difference between their product existing or not existing!

For developers who work on large, complex code bases, enthusiasm is more muted. AI has no context about the very specific requirements and existing implementations to contend, and is less useful:

Creative works: don’t come to a theater in your area

There is a lot of fear about AI among creatives: shall we all soon read novels generated by AI and watch films generated by AI?

This is unlikely because creative fields are extremely competitiveand beating competition to pay attention novelty. Although AI has made it easier to generate images, audio and text, it has (with with Some exceptions) Not increased production of ears and eyeballs, so the bar to make a competitive product is too high:

Novelty Is a difficult requirement for successful creative work, because people are extremely good at detecting when something they view or read is derived from something they have seen before. This is why, although Avatars in Studio Ghibli style briefly took over the internet, they have not affected the cultural position of Howl’s moving castle.

Things you already do with apps on your phone^{: minimal impact}

An area that has not Given that there is a lot of impact on tasks that already have specialized apps. I will concentrate on two examples with abundant MCP implementations: E -Mail and Order food. AI Doordash agents and AI film producers face the same challenge: the bar for a new product to make an impact is already very high:

E -mail seems like a ripe area for disruption by AI. But modern e -mail -apps already have a wide range of filter and organizational aids that Tech -Savvy users can use to make complex, personalized systems for efficient consumption and organizing their inbox.

Summary Is a core -ai -skill, but it doesn’t help much here:

• Spam is already shaken quietly in the spammap. A summary of the mess is, well, rumbling.
• For important e -mail, not me want to A summary: an AI will probably produce less specifically manufactured information than the sender, and I don’t want to miss any risk of missing important details.

Similar to ordering food: apps such as Doordash have carefully designed interfaces. They find a careful balance between information such as price and ingredients against photos of the food. It is unlikely that AI will produce interfaces that are compiled faster or more.

The future is already there – it is just not very evenly distributed

AI has raised the floor for knowledge work, but that change does not matter to everyone. This goes a long way to explain the very wide range of reactions to AI. For technical managers like me, AI has had a huge impact on my relationship with technology. Others fear and be replaced repeatedly. Still others hear smart people express enthusiasm for AI, struggling to find and think useful I just don’t have to get it.

AI has not replaced how we do everything, but it is a very capable technology. Although it is worth experimenting with, whoever you are, if it doesn’t seem like it’s logical to you, that’s probably not.