brandonnguyn
Region: US
Wednesday 17 December 2025 02:54:48 GMT
55132
5632
81
809
Music
Download
Comments
torque :
Ts pmo im not learning mayonnaise for 2% efficiency 💀
2025-12-17 16:00:58
775
sheczaun :
they’re actually most efficient with hebrew
2025-12-19 11:45:19
114
Alex2epic :
How u gonna code in Chinese be
2025-12-17 20:58:11
150
Loki 胡洛奇 ☭ :
Neural networks are fundamentally Chinese
2025-12-17 22:36:32
84
John nuts :
I thought it was polish?
2025-12-17 18:30:55
26
Jonathan :
Ya, but that’s just for the thinking step. The input and output text can be any language.
2025-12-17 04:39:56
269
Purple :
It seems polish might actually be better
2025-12-17 08:45:38
24
. :
A study was just released showing that the accuracy is much lower with symbol based languages
2025-12-28 10:16:32
4
Name :
I'm Hispanic
2026-02-04 20:14:00
1
robofarm00 :
How to learn mandarin
2025-12-17 18:10:10
2
Trent Herring Khaos :
Like pinyin?
2025-12-28 03:53:19
0
Code Snob :
Chinese characters are 4 byte Unicode, so they’re actually not smaller
2025-12-29 16:00:31
3
tensor :
not particularly, and especially not from models which aren't China centric. Some models were evolving to using Chinese for thinking tokens as it's more token efficient, but that doesn't mean that it performs BETTER in Chinese. even if it did, input language and output languages can be different and it doesn't really matter since it's all semantic representations anyways. Transformer models were created actually for language translation, granted these were encoder decoder models instead of decoder only, but mathematically context and an encoder aren't all that different, near identical behavior can be learned
2025-12-17 18:43:41
0
Jason🫗 :
I saw that Twitter post too
2025-12-18 05:00:40
1
Henry :
This is actually false. Because many characters have multiple meaning, it has a harder time with the precise vectorization
2026-01-29 01:15:04
5
kole.labs :
Do I have to first be Chinese to become more Chinese?
2025-12-18 03:18:59
3
dede.talks :
Already started chat
2025-12-18 00:35:07
1
Razza :
It’s Polish not chinese
2025-12-18 23:27:41
0
Pablo :
“Change to mandarin please” ok now what
2026-02-22 07:13:08
0
Jon :
i literally saw a video saying the opposite a few days ago now i'm confused 😭
2026-01-03 06:14:19
0
- :
whats should i major in college
2026-01-11 15:25:44
0
RichieRich53476 :
so this is partially true it is competitionally more effective but llm have a higher chance of losing contacts because of the difficulty of accurately processing Mandarin characters
2025-12-28 22:38:13
0
mik :
我的中文不太好
2025-12-18 00:54:17
0
To see more videos from user @brandonnguyn, please go to the Tikwm
homepage.
