@gigaqian: Chinese AI is getting closer to the information you consume… and Reuters may only be the beginning. Thomson Reuters just built its own proprietary AI on top of technology descended from Alibaba’s Qwen, then trained it on Reuters news and other private data. That doesn’t mean Alibaba is writing Reuters articles… but it signals a trend. The bigger story is more important: open-weight Chinese models are getting good enough that Western companies can build their own systems on top of them, customize them privately, and you may never even know where the underlying technology started. You might never download Qwen. You might never knowingly use Chinese AI. But technology that started with it could still end up inside the systems that shape the information reaching you. #aitools #reuters #china #technews #gigaqian
America is cooked. We voted our way out of the leading frontier.
2026-08-24 23:47:50
2
:
Can confirm doing similar activities at my fortune 300
2026-08-24 23:41:32
7
Peter Burstyn :
This is a very slippery slope! What's stopping a "reporter" from using the AI for the entire article and never fact checking?
2026-08-25 00:49:00
4
@thecarboncure :
My mac M5 16g couldnt handle the newest qwen. Lol
2026-08-24 23:42:10
2
thepsychuniverse :
Nice awesome journalism keep it up 🙏
2026-08-25 00:31:34
1
Jason J. McBride 🇨🇦 :
Great post, very informative 🥂
2026-08-24 23:46:29
3
MagicAIDaydreams :
Thank for the info.
2026-08-24 23:56:07
1
Линкольн 💸 :
i don’t even use ai anymore without walter fixing it
2026-08-25 03:12:25
1
TexanRT :
Reuters about to be sanctioned
2026-08-25 14:27:11
0
Kairma :
Well thats TERRIFYING on a huge epistemic level.
2026-08-25 16:17:55
2
BAILEYpc :
I'm going to give you a heads up. yesterday I gave five different models of simple task. four of these models were the large primary AI domains. the simple task was look up this domain name and tell me the who is info. Microsoft co-pilot was the only one that attempted it but still got it wrong every single time. the rest of the models refuse to do it or stated that it could not do it. no I'm in a simple chat on my phone doing this. then I asked Kimi... not only did it show me its processes as it was executing my ask but it provided a report as if I was looking at it on a web page. it got everything perfect including the registrar. this is a simple task that I would have expected anthropic to pull off but it couldn't. my point is I think the Chinese models have already surpassed the US and that's the panic and I think the panic is showing in the numbers here.
2026-08-25 16:12:24
0
Lynn user485797232483 :
Hi Giga
2026-08-25 03:40:51
1
Arcane :
I have already embraced AI replacing journalism, check out my site
2026-08-25 20:05:13
0
™️_꧁༒☠ᴊɪʜꫝᴅ/☠︎✓ :
proofademic flagged paraphrased content and i’m unsure why
2026-08-25 07:51:53
0
DL :
Great. Open weight is at least not as biased to the US ones
2026-08-25 07:12:08
0
OpenclawGateway :
qwen3.8:27b is so amazingly next level.
2026-08-25 04:16:37
0
AI Fan Token (AI) :
Perception changes value.
2026-08-25 23:14:07
0
Kdog :
follow me if you like skiing
2026-08-25 08:50:24
0
Sir Zeno :
🔥🔥🔥🔥🔥🔥🔥🔥🔥🔥🔥
2026-08-25 02:52:40
1
WillWylder8888 :
🌠
2026-08-25 02:07:11
0
user6612763152721 :
And then @Reuters still asked for us to pay for news they are getting and written for free . Hell no. No more subscribing to Reuters!!✌️
2026-08-25 01:57:33
0
Lord Brosifious :
This dudes a Chinese propagandist
2026-08-25 01:23:51
0
To see more videos from user @gigaqian, please go to the Tikwm
homepage.