Wishes for the next 10 years: Intel's comeback, Nvidia at the top, health for my family and myself, and well-being for Yuuka.

Joined December 2021
130 Photos and videos
Pinned Tweet
分享一只疑似中国五毛账号,警惕该支那蛆虫在x上发布虚假混淆视听信息
65
true
"Highest performance" on benchmarks designed for Vera. Real test: mixed workloads with latency constraints. Single-metric wins don't predict production. Custom cores optimize for benchmarks, not generalist tasks.
1
65
井外势力 retweeted
"Highest performance" on benchmarks designed for Vera. Real test: mixed workloads with latency constraints. Single-metric wins don't predict production. Custom cores optimize for benchmarks, not generalist tasks.
1
3
314
井外势力 retweeted
高盛机构很大分支机构众多,里面有亲华的VC,而且全球均衡投资也本就是他们的基本策略,至于被大A割了之后会不会在继续上当,亲华的VC会不会被开了,可以看看明年他们的报告,顺带一提高盛给中共出的报告其实是与国内白手套伙同一起骗人的,而且是两头骗的那种
4
74
4,700
grok
149
井外势力 retweeted
Honored to work with Intel and Lip-Bu
Elon has a proven track record of reimagining entire industries. This is exactly what is needed in semiconductor manufacturing today. Terafab represents a step change in how silicon logic, memory and packaging will get built in the future. Intel is proud to be a partner and work closely with Elon on this highly strategic project 🚀🚀
2,018
4,600
51,068
29,170,837
95
Grok imagine. #grok @grok @zkbyteoracle
1
1
98
井外势力 retweeted
Yeah not quite, desktop 5080 (which is the same as laptop 5090) gets ~60fps at 1440p Ultra RT, so over 2x faster... M5 Max in there is more akin to 5060ti, still impressive at the end of the day. Credit: Https://www.youtube.com/watc…
1
3
12
1,487
井外势力 retweeted
💚
GPT-5.4 Thinking and GPT-5.4 Pro are rolling out now in ChatGPT. GPT-5.4 is also now available in the API and Codex. GPT-5.4 brings our advances in reasoning, coding, and agentic workflows into one frontier model.
75
99
2,235
147,581
井外势力 retweeted
intel 的 xess 3 有点疯,b580 这种级别的卡都能 4k 50% 超采样 120 FPS,测了下即使是 100% 也是 80FPS
4
1
26
11,094
井外势力 retweeted
I don't trust Qwen models, even the large ones. I trust small ones even less. Qwen has been benchmaxing its models since the very beginning, so the benchmark performance of a 4B parameter model beating Gemini 2.5 Flash Lite is most likely just the benchmark performance only. 4B is just not enough to store any general intelligence because most of the parameters will still be used to store facts. Well, not facts, but something that remains of them.
72
6
145
28,860
已经有老外开始吐槽了,字节这个以后就是只能生成一些网络弱智短视频的水平了,估计一大波退款潮要来了🤣🤣🤣🤣
1
117
This is just a marketing video. In reality, the current version of Seedance is nowhere near capable of generating visuals at a PG-13 level or higher.🤣🤣🤣 #seedance2 #ai
It's over for VFX artists... AI can now keyframe every second of animation from just prompts
2
298
昨天开始,原创的AI生成人脸(nano banana flux lora)都会被即梦检测成真人而无法通过审核。这意味着这个模型的图生视频能力基本等于残废状态,你压根不可能拿它去做写实类的影片,只能去做那种烂大街的15s垃圾文生网络梗视频或者非写实类的视频🤣。好在可灵这个时候能勉强顶上。#seedance #seedance2
2
498
Seedance目前无论是白天还是晚上,生成一个视频的时长在3小时到8小时之间不等,无论哪个级别的会员均如此。此外,生成进度99%之后会有极大的概率出现无法通过审核的情况,哪怕是原创的图生视频亦是如此。更离谱的是即使通过了层层阻碍最终生成成功了,也有极大概率出现各种逻辑错误。#ai #seedance2
1
1
336
以后再看到网上吹的我觉得可以拉黑了,无论中文账号还是英文账号,事实上我发现傻白甜的英文无脑账号吹更多,而且我确信他们根本没实际使用过。🤣🤣🤣
1
80
即梦今天对外开放API后,是这样的,旗舰会员。。。然后一群傻白甜老外争先恐后尝试生成结果一堆不过审或者干脆生成一堆乱七八糟的恐怖谷视频片段。画面一定很美。#seedance #ai
2
5
405