EPISODE · Feb 3, 2025 · 7 MIN
EP17｜AI界有版權的谷阿莫：Deepseek？
from 神蓉世 · host 陸蓉
生成式建模有兩步驟：訓練(training)、推理(Inference) 訓練成本遠大於推理成本。 Deepseek的爭議來自於用「蒸餾」手段沿用OpenAI訓練結果後， 在推理領域有創新，但推理仍是建立在OpenAI的訓練。 ❌降低98%成本 Deepseek直接沿用OpenAI的成果，這全部都能算是Deepseek省下的成本嗎？ ❌繞開輝達護城河 在推理階段，Deepseek離確繞開輝達，但深層算力仍是建立於輝達GPU (OpenAI 數據） ❓沒有電影，谷阿莫還能解說電影嗎？ 沒有OpenAI，天才少女還是天才嗎？ FB：www.facebook.com/rong.rx.fb IG：www.instagram.com/rong.rx -- Hosting provided by SoundOn
Embed this episode
Ready to play
EP17｜AI界有版權的谷阿莫：Deepseek？
0:00
7:54
1×
No transcript for this episode yet
Similar Episodes
No similar episodes found.
Similar Podcasts
No similar podcasts found.
Frequently Asked Questions
How long is this episode of 神蓉世?
This episode is 7 minutes long.
When was this 神蓉世 episode published?
This episode was published on February 3, 2025.
Can I download this 神蓉世 episode?
Yes. Use the download control on the episode player to save the publisher-provided media file.
URL copied to clipboard!