μ μνΈλ§ :Β
"μ λ κ·Έκ²(GPT-5)μ΄ μ΄λ€ μμμμλ μμ²λκ² λμμ§ κ²μ΄λΌκ³ μμνμ§λ§,
λ€λ₯Έ μμμμλ λλκ²λ μ΄μ λͺ¨λΈλ€μ κ²½μ°μ²λΌ κ·Έλ€μ§ λμμ§μ§ μμ κ²μΌλ‘ κΈ°λν©λλ€.
μ΄κ²μ μ°λ¦¬κ° λ€λ₯Έ λͺ¨λ λͺ¨λΈλ€, GPT-3μμ 3.5, 3.5μμ 4λ‘ κ° λλ§λ€ νλΒ λνμ κ°μ΅λλ€."
"I expect that it'll be hugely better in someΒ areas and surprisingly,Β not as much better in others which has been the case with every previous model."
μ μ리 κ±Έμλ§ νλ©΄μ μ£Όλ₯μ리 μ‘΄λ μ³κ°κΈ°λ Έ - dc App
μκ°λ¦¬λ₯Ό μ‘΄λ νΈλλ§
μ΄κ±° μμ μ€μμ. κ·ΈμλΆν° We don't expect that we're near an asymptote, but, you know, this is like a debate in the world, and I think the best thing for us to do is just show, not tell. You know, there's a lot of people making a lot of predictions, and I think what we'll try to do is just do the best research we can, and then figure out how to responsibly release whatever. Whatever we're able to create.
I expect that it'll be hugely better in some areas and surprisingly not as much better in others, which has been the case with every previous model. But this feels like the conversation we've had every other model release. When we were going from 3 to 3.5 and 3.5 to 4, there's a lot of debate about, well, is it really going to be that much better? If so, in what ways?
And the answer is there still seems to be a lot of headroom. μ΄κ±°κ³
λ¨Ό λ»μ΄λλ©΄ μ κ·Όμ μ κ·Όμ νκ³ μμ§ μλ€. (μ 체λκ³ μμ§μλ€) κ± μ°λ¦¬λ κ²°κ³Όλ₯Ό 보μ¬μ£Όλ©΄ λλ€. μ΄λ λΆμΌμμ ν¬κ² μ’μμ§κ³ , μ΄λ€ λΆλΆμ λ κ°μ λ μ μλ€. 3μμ 3.5 κ°λ, 3.5μμ 4κ°λμλ κ·Έλ¬λ€. μ΄λ° λμμ€μ. μ΄κ±Έ μ΄λ κ² μλΌμ μ λ κ² λ²μνλ©΄ μ λ°λμ λ»μ΄ λλκ±°.
μ¬μ ν νκ°μ μμ κ±°κ³ (LLM κ³ μ μ λ¬Έμ μμΌλ‘) νλ₯ ν μΆλ‘ μμ ν¬κ² λ²μ΄λμ§ μμ κ±°λΌλ λ» κ°μ
μ΄λ€ λΆμΌμμ ν¬κ² μ’μμ§κ³ μ΄λ€ λΆλΆμ λΉμ·ν κ±°λ€ <--- μ΄κ² μμ μΈλ° λκ° μ€μμ΄λΌλκ±°?
μ λ¬Έ λ΄λ νλ¦°λ§ μλλ°? λκ° μ€μμ΄λκ±°
μ§μ§ μ΄ν΄λ₯Ό λͺ»νλ 건κ°.. λ°μ μ΄ μ 체λκ³ μμ§ μλ€. μ΄λ€ λΆλΆμ ν¬κ² κ°μ , μ΄λ€ λΆλΆμ μ κ² κ°μ λ κ²μ΄λ€. 3->3.5 λλ, 3.5->4 κ° λλ λ§ λ§μλ€. μ΄μ°λλ λ΅μ λ°μ μ μ¬μ§κ° λ§€μ° λ§λ€λ κ²μ΄λ€. < μ΄ λνμ λ Όμ μ΄ λ°μ μ΄ μ μ κ²μ΄λΌλ κ²μ? μ΄λ€ λΆλΆμ 'μ κ²' κ°μ λ κ²μ κ°μ‘°ν κ²μ΄ μ λ μλ. κ·Έλ° μμν λ Όμ νμ μλ€, κ²°κ³Όλ‘ λ³΄μΈλ€, μ κ·Όμ μ λΏλ κ²μ΄ μλλΌ λ°μ μ¬μ§κ° λ§λ€λ κ²μ κ°μ‘°νλ κ²μ΄μ§. μ λ°λμ λ Όμ‘°μ§ μ΄ κΈμ΄λμ
ν° λ³ν μλ€λκ±°λ€ κ²¨μΈμ΄λ€?
μ λκΈ κ°μ.
GPT-5 νκ°μ λμμ§μ§ μμ κ². λ₯μ μΈλ©΄ λ¬Όμ λ΄λ¦΄ κ².
μμ²λ κ°μ 2027 μΆμ
μ΄μλΌ νμ₯νμ΄ κΈ°μ 1μΌμ°¨