Behind the ChatGPT, why is it getting better every day?

(Cause): Have you ever wondered why an AI like ChatGPT knows so much, can answer almost anything!

Many people will probably use ChatGPT or AI help with day-to-day tasks.

But you know that...Behind its "super-sensitive intelligence" is to "train so hard!"

And it's not just one train and it's over...

Building AI like this requires two big processes:

1.Pretraining - Training from huge data (like reading the whole world)

2.Finetuning - adjust the tune exactly to the task to do (like a special class to the point)

Today, Herpes will tell you a simple account of what these two steps are and how we live!

(Formula): What is Pretraining + Finetuning?

✨ 1.Pretraining: Read everything the world has.

* AI will be taught with "enormous" amounts of data such as web, books, code, etc.

* Goal: Let it "understand the language" as broadly as possible.

* School characteristics: No one teaches directly. Learn from Pattern in the data itself (called "unsupervised learning").

Simple Compare: Like a child who reads everything in a world library, and then captures Pattern himself as to what should follow.

✨ 2.Finetuning: Specialized training to excel in the subject.

* After understanding the common language, you have to "tune up" to suit your job, such as being a doctor's assistant, marketing coach, etc.

* Use specialized information + someone gives feedback or teaches directly.

Simple comparison: Like a child who has read every book, but has to come to a special class on "Make content bang in TikKa."

(Effect) So what about us?

AI is smarter every day because he has a system of "learning + tuning" all the time.

But is there anything in our business to learn and develop?

If you still use AI today, it may be time to "Finetune" your own business to win the market!

Who wants to know how AI can be used to make content, sell, or mold brands to be tuned?

Come to talk to Herpes at LINE OA: @ drgim