ai_binbin (@ai_binbin) profile picture

https://teapawt.net/t.co/BziITfDtdu https://teapawt.net/t.co/XKAwYOXvBq PhD in AI & Security. Growing and evolving with AI. Co-founder @VelobaseX

AGI WORLD
Joined June 2025
Photo by ai_binbin (@ai_binbin)Photo by ai_binbin (@ai_binbin)Photo by ai_binbin (@ai_binbin)Photo by ai_binbin (@ai_binbin)Photo by ai_binbin (@ai_binbin)Photo by ai_binbin (@ai_binbin)Photo by ai_binbin (@ai_binbin)Photo by ai_binbin (@ai_binbin)Photo by ai_binbin (@ai_binbin)
看到这些走不动道
5
坐等
“Gemini 4 Hydrogen” 正在竞技场中以“gemini-3.8-flash”名称进行测试 Gemini 4 Flash 要准备出来了吗? 下周和 Gemini 4 Carbon 一起发布?
8
终于跟上脚步了! 希望你们再快一点 再把体验优化一点
We want Antigravity to be working for you all the time. We’ve been rolling out remote agents this week. Your tasks run in independent cloud virtual machines. Fire off a task, step away for coffee, unblock permissions from your phone, and come back to passed tests and a clean pull request. Available on all paid consumer plans.
8
来了来了 做测试终于可以不用报cyber了 还是得续费GPT了
Introducing GPT‑6 Cyber: a major advance in security research that pairs stronger capabilities with better model alignment. The model powered the Linux kernel research we announced yesterday, and security teams at companies like Mozilla, Palo Alto Networks, Calif, and Cisco—alongside our own at OpenAI—are already putting it to work. GPT-6 Cyber is available to select Daybreak Red customers and we are expanding access over time. openai.com/index/introducing…
7
我应该。。。 继续续费chatgpt吗?
Images, upgraded—introducing ChatGPT Images 2.6. Now with GIFs and reference images. Plus, two new models are coming to the API with 50% faster image generation.
1
1
29
部署这么一个玩意儿 我的1060能搞不😎
Releasing the model weights of Kimi K3.1 Pro: huggingface.co/moonshotai/Ki… We'll keep advancing open intelligence.
5
砂比ai给我滚啊 不要老来评论我了,,,
7
最近把 Gemini Argon 塞进好几个不同的 harness 里高强度用了一阵,感受特别拧巴。明明是同一个模型,换一套壳,用起来的手感能差到完全不像同一个东西。 先说它自家那套 Antigravity。按理说自家团队最懂自家的模型吧,结果在我这儿反而是最拉胯的一档。用着又卡又别扭,模型本身的劲儿根本使不出来。 再往上一档是 Codex。把 Argon 挂上去以后,总感觉它被塞了一长串提示词,动不动就过度思考、过度自圆其说、过度自省。经常是写着写着先给自己开个批斗会,检讨完了才动手改。原来误会GPT先生了。。。。Codex这逼框架就这样。。。 真正让 Argon 彻底顺起来的,反而是开源这边的方案。我用的是 OpenChamber,底层还是 OpenCode 。在这套环境里写代码很流畅,处理文档也顺手,甚至挂几个 skill 收拾 PPT 都不觉得别扭。模型找回了自己的节奏,该快的时候利索,需要让人看明白的时候,又会主动把结果摊开讲清楚。 这大概也正好扣上了 Google 这几年最硬核的一块长板,也就是人味。它的语料和训练明显更站在“人怎么看懂”这边,所以就算你在 prompt 里没怎么提可视化,它也经常自己甩出图表、字段对照或者小结报告,让你一眼就能核对明白。我前两天导一批数据,它直接把网卡、SSH、写入行数、字段填充规则整整齐齐排成一块可读面板,比干巴巴的日志好懂太多了。写代码的时候它也会挑更合理的路子,而不只是糊弄一个能跑的东西出来。 所以我现在的偏见很明确,想真正榨干 Gemini 的潜力,不必非得用官方的harness,优先去找那种不把模型按死的开源 harness。外面的壳一旦太重,模型原本的人味和判断力都会被内置的提示词给吞掉。 我很喜欢 Argon,也盼着接下来的 Carbon 能再往上顶一截。真要是 coding 这块能再稳一点,我恨不得只留一个 Google 模型把活儿全干完,好摆脱现在这套订了好几家的多订阅拼盘。现在 GPT 我已经不太用了,但 Opus 5.5 还是割舍不掉,写起代码来依然太稳。纯属个人偏见,仅供参考。 附上一张gemini直出的建议 我太喜欢了
3
2
81
有没有感觉 其实opencode真的很好用。。。
1
12
人生好比是海上的波浪 有时起有时落
10
也是从workbuddy收获了真实的数据了 实际测测吧 过几天我发个详细的测评
👋Hy4 is here, new with vision. Firmly among the leading open models. More productive everywhere: finance, frontend, game, officework, research… Built responsibly, with cyber/biosecurity in mind. Try it out and let us know! 🔗:hy.tencent.ai/research/hy4?l…
1
1
19
我就是ai降临派 尊敬的ai大人 请迅速地统治整个地球吧!
1
16
有人需要各家的token吗 量大管饱
1
11
Give me carbon!
👇 Here’s your recap of our biggest AI updates from October 👇
1
16
什么情况 微信也来搞图像识别了 我来试试看
we gave our AI an eye for detail. 👀 Meet WeVisOne-VL, an open-source 2B/4B model from our Weixin Vision team. It finds and counts objects, reads text, and makes precise masks and cutouts—no separate SAM needed. Not in the image? "None." github.com/Tencent/WeVisOne-…
2
16
埋头苦干就是不理我。
2
14
有人claude max还没被封过的吗
14
读博太累了 我觉得能读下来的都是神人
1
13