Sapientia

995 posts

Sapientia banner
Sapientia

Sapientia

@FutureByHistory

Exploring, seeking, and unveiling the truth. Esse quam videri. Fuck Anthropic.

From PRC Katılım Mayıs 2025
297 Takip Edilen38 Takipçiler
Sabitlenmiş Tweet
Sapientia
Sapientia@FutureByHistory·
Just took a political ideology questionnaire and got a bullseye on 100% centrist. I expected a result something more like social liberalism or just plain liberalism, but to my astonishment I scored dead middle — 50/50— on every dimension. True middle ground on everything. Neither extreme left nor right — pragmatic compromise seems to be the way😎.
Sapientia tweet mediaSapientia tweet media
English
0
0
2
229
卡尔
卡尔@carl88187210579·
不懂就问, 为什么Coser愿意还原一些很暴露的角色呢🤔?
卡尔 tweet media
中文
243
12
1.2K
472.8K
NVIDIA
NVIDIA@nvidia·
AI security advances when the industry builds in the open, together. We're introducing the Open Secure AI Alliance with industry leaders to develop new techniques and tools to safeguard software and agents. By sharing models, tooling and research in the open, we can broaden the community of defenders. Learn more about the founding members’ contributions: nvda.ws/4pD8Fc5
NVIDIA tweet media
English
402
804
6K
2.9M
Sapientia
Sapientia@FutureByHistory·
@Mr_Salio The problem is not Mythos now. Kimi K3 already beats Mythos in some areas. If Kimi K3.1 did not surpass Mythos then the update is not really an update. The real boss that Kimi K3.1 should challenge is Opus5.
English
0
0
0
136
Salio
Salio@Mr_Salio·
🚨 Kimi K3.1 Leak >Performance that closes the gap with GPT-5.6 and Fable >Kimi K3.1Maintain its reported Mythos-level capabilities >Faster inference and lower latency >Better token efficiency on long reasoning tasks >More reliable coding and agent workflows >Fewer unnecessary reasoning steps >Continue outperforming Opus 5 on key coding and agent tasks >Remain open-weight, making frontier-level AI accessible to everyone do you think it will beat mythos?
Salio tweet media
English
74
83
1.2K
87.3K
✧
@HeyMikima·
BAN PORN ON X BAN PORN ON X BAN PORN ON X BAN PORN ON X BAN PORN ON X BAN PORN ON X BAN PORN ON X BAN PORN ON X
English
1.1K
311
3.9K
2.8M
星野ふじ
星野ふじ@hoshinofujivy·
日本人吃鱼,所以用「杂鱼」骂人 中国人用「傻逼」骂人,所以中国人是吃逼的吗?
星野ふじ tweet media
中文
184
14
506
38.8K
Sapientia
Sapientia@FutureByHistory·
@Mononofu @JensenHuang The best irony in the 21st century: A member of the technical staff at Anthropic doesn’t know the difference between open weight and open source.
English
0
0
4
820
Julian Schrittwieser
I’m so excited that @JensenHuang is a believer in open source now, looking forward to the CUDA and GPU driver open source release!
Jensen Huang@JensenHuang

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. The world needs both frontier closed models and frontier open models. images.nvidia.com/pdf/Open-Weigh…

English
1.7K
344
5.4K
6.4M
Max For AI
Max For AI@MaxForAI·
哈,现在全硅谷的AI公司似乎就只有一家没有签署这个协议了🤣 @AnthropicAI 你还活着吗?
Max For AI tweet media
Jensen Huang@JensenHuang

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. The world needs both frontier closed models and frontier open models. images.nvidia.com/pdf/Open-Weigh…

中文
94
34
738
231.5K
鸟叔本鸟
鸟叔本鸟@niaoshu·
成功了!玻利维亚区的chatgpt,这波我也上车了: 1、订阅了一个 plus,139.99波币,即,12.61美元, 2、如果升级到20X是1379.99波币,即,124美元, 3、美区IP就行,Play商店要改成玻利维亚区的账单,谷歌支付,绑定招行的Visa卡。我用谷歌账号登录chatgpt账号...
鸟叔本鸟 tweet media鸟叔本鸟 tweet media
中文
154
109
904
194.5K
草莓泡芙🍀
草莓泡芙🍀@SnozakiSakura·
突然想到一件事。 如果你天天跟 Grok 聊涩涩, 四舍五入是不是等于在和马斯克文爱?
中文
55
6
209
22.7K
草莓泡芙🍀
草莓泡芙🍀@SnozakiSakura·
今天,ChatGPT终于上线了其他AI早就有的功能 —— 深入思考 模型应该变得比之前更聪明了吧?
草莓泡芙🍀 tweet media
中文
52
0
20
16.2K
BridgeMind
BridgeMind@bridgemindai·
CLAUDE OPUS 5 IS HERE AND IT BEATS FABLE 5. At half the price.
BridgeMind tweet media
English
125
127
2.3K
187.9K
Sapientia
Sapientia@FutureByHistory·
Altman is a hypocrite. He furtively changed the non-profit foundation written in OpenAI’s founding charter. The only “open” model of OpenAI is GPT-OSS. Frankly, I believe that model should be called GPT ASS, it has hallucination rates around 50%, and it totally fails multilingual tasks. Same situation with Anthropic, who also didn’t sign the decree mentioned by Jenson, except for that Dario is even more despicable: putting trackers into account suspension emails, adding spyware into Claude code, and no open models, not to mention they lowered their red line in AI safety guardrails to a very low position, plus they offers no open models at all. Sometimes I even wonder if Anthropic is a subsidiary of the CIA! They even tried to lick the Pentagon’s feet for a few months but that didn’t end up well. It appears that the Pentagon has an even lower moral standard than Dario has! All in all, OpenAI does some dirty and eyebrow-raising things behind the scenes, but they are pale in comparison with Anthropic.
𝘊𝘰𝘳𝘳𝘪𝘯𝘦@OopsGuess

Sam Altman now wants to cosplay as a defender of open AI. OpenAI built a closed system by extracting value from the world’s books, art, code, journalism, forums, and collective cultural memory—then locked the resulting model behind corporate control. But when an open-weight Chinese model learns from the outputs of that closed system, suddenly it becomes: “distillation.” “IP theft.” “a national-security threat.” OpenAI was allowed to absorb humanity. China is not allowed to absorb OpenAI. You privatized the commons, then called imitation of the privatized product theft.

English
0
0
2
110
𝘊𝘰𝘳𝘳𝘪𝘯𝘦
Germany used to have an almost mythical reputation in China. Precise. Rational. Disciplined. German engineering. German quality. German efficiency. Chinese internet even spent decades circulating that ridiculous legend about German engineers leaving carefully wrapped spare parts inside Qingdao sewers a century ago. Then China became Germany’s industrial competitor. Suddenly: Chinese EVs are unfair. Chinese solar is unfair. Chinese manufacturing is “overcapacity.” The yuan is unfairly undervalued. Germany needs a “Plaza Accord” for China. And now the U.S. Treasury says no major trading partner manipulated its currency for trade advantage in 2025. Twenty years of German mythology disappeared the moment Germany started losing. Turns out the legendary German precision was perfectly capable of measuring everything except its own decline.
𝘊𝘰𝘳𝘳𝘪𝘯𝘦 tweet media
𝘊𝘰𝘳𝘳𝘪𝘯𝘦@OopsGuess

Germany is now talking about “Plaza Accord” style talks on the yuan. That alone says everything. They are still dreaming of doing to China what the United States once did to Japan in the 1980s. Force currency adjustment. Crush export competitiveness. Rewire the industrial balance. Turn a rising manufacturing power into a managed client economy. But there is one problem. China is not Japan. China is not a defeated, occupied, security-dependent U.S. colony. China does not host American troops as the price of sovereignty. China does not outsource its strategic survival to Washington. China does not mistake a leash for an alliance. If Europe is looking for the new Japan, it should look in the mirror. Europe is the one following Washington’s sanctions. Europe is the one cutting itself off from cheap energy. Europe is the one restricting high-end exports under U.S. pressure. Europe is the one losing industry, losing competitiveness, and then blaming China for producing what Europe no longer can. China did not create Europe’s industrial weakness. Europe chose deindustrialization, dependency, expensive energy, moral theater, and strategic obedience. Now it wants to call China’s competitiveness “overcapacity” and China’s currency “undervalued.” No. This is not 1985. China is not Japan. Europe is.

English
25
86
493
32.3K
Sapientia
Sapientia@FutureByHistory·
@MKBHD Make a benchmark by yourself, and install on it.
English
0
0
0
8
Marques Brownlee
Confirmed, Z Fold 8 doesn’t seem to want to install any benchmarking apps at all. What gives?
Marques Brownlee tweet media
English
487
543
25K
2.1M
Max For AI
Max For AI@MaxForAI·
卧槽??Claude Opus 5效果这么好吗?? 一个Prompt,一次生成,直接做出了一个可运行的高精度3D网页场景。 测试者 @chetaslua 称,这是他见过单次生成中细节最丰富的一次。过去即使连续修改多轮,也很难达到同样的完成度。 他甚至直接给出判断:这次输出超过了Fable 5、GPT-5.6 Sol和Kimi K3。 真正值得注意的不是画面更精致。 模型需要同时理解3D结构、材质、光照、动画、交互和前端代码,再把这些东西一次性拼成完整的产品。 如果正式版本能够稳定复现,Vibe Coding的下一步可能不只是生成网页,而是直接生成可交互的3D世界。 截至目前,Anthropic尚未正式发布Opus 5。
Chetaslua@chetaslua

Opus 5 made a 3D home , look at the details this is what i would believe is the AGI of @threejs if someone showed me this a year ago look at the texture and details

中文
6
3
44
16.3K