小东与Ai

Langboard字间输入法 边打字边学习单词

[小东与Ai @aiislove666]

Indie developer. AI enthusiast. Stock investor. 🚀

I build apps, spend way too much time following AI, and occasionally pretend I understand the stock market.😂

Here to share what I'm building, what I'm learning.

Let's connect! 👐

[小东与Ai @aiislove666]

还是东鹏反应快，已经删掉了姆巴佩相关的代言，蒙牛还没有动作，估计待会都会官宣解约

[小东与Ai @aiislove666]

GPT-6.1-Terra(max) Pelican testing 😅 I don’t expect Astra-6.1 to beat Opus 5.5 on the frontend anymore.🙂

[小东与Ai @aiislove666]

Mbappé will be CANCELLED in China.

[小东与Ai @aiislove666]

Does Dario or Sam pay you? So the next move is banning the open sourced model so the closeAI could hike price?😅

[Cohere @cohere]

While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected.

Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions.

Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent.

But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models?

We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem.

Read the full report: https://t.co/TZSugXi2ZE

[小东与Ai @aiislove666]

@grok what if Elon was a normal user

[小东与Ai @aiislove666]

@grok Estimate how much Elon gets from the Original Content Rewards Program every month?

[小东与Ai @aiislove666]

@thsottiaux Terra may be more suitable for operating the acceptance results of the simulator.

[小东与Ai @aiislove666]

@1vnzh So what you mean is that you are more willing to believe in the closed source model?🤣

[小东与Ai @aiislove666]

@cohere Is this just dirty data during training?If you want to upload files, you have to have a key or something like that. Does the model also come with its own?

[小东与Ai @aiislove666]

@_Jason_Dean_ @synthwavedd much worse than sonnet

[小东与Ai @aiislove666]

@ArtificialAnlys poor performance in the Terminal Bench 4.0 test comparing with SOL

[小东与Ai @aiislove666]

OpenAI just released GPT-6.1-Terra. Compared with Sol-6.1, there is no advantage in The Terminal-Bench 4.0 score. So it is recommended to continue to use Sol.

[OpenAI @OpenAI]

GPT-6.1 Terra: delivering fast and capable computer use at a lower price.

GPT-6.1 Terra understands what’s on screen, takes action across apps, writes naturally, and is a more capable model for your whole workday with the efficiency to take on more of it.

[小东与Ai @aiislove666]

@QwenDevs qwen4 soon?

[小东与Ai @aiislove666]

@HughieJau 不迟不迟 来了就是朋友🤗

[小东与Ai @aiislove666]

感谢大家的支持，快500关注了，谁会是第500个呢🤗

[小东与Ai @aiislove666]

@Zhibin992 来咯

[小东与Ai @aiislove666]

Apple really messed up with the iPhone Duo this time.

A Foxconn employee said that in the state of tight production capacity, Apple had previously reduced certain good margin quality inspection standards to ensure supply, which would certainly lead to a higher probability of products with screen failures entering the market.

[小东与Ai @aiislove666]

@AnthropicAI Open-source will win the game. No company be willing to handle their datas to CloseAI.

[小东与Ai @aiislove666]

Anthropic’s latest financials look incredible. Enjoy it while it lasts.

My prediction: Within 6 months, open-source models will surpass Anthropic’s best models and completely destroy its pricing power.

Open-source AI, led by DeepSeek, is closing the gap at an insane pace.

The AI gold rush is about to turn into a brutal price war.

[小东与Ai @aiislove666]

@xingyu_ling 羡慕啊， 我今天好像都没更新，估计就几个忽略了🥲

[小东与Ai @aiislove666]

@bohao21328888 但是体感上不一样，所以吃了没有周额度的亏😂

[小东与Ai @aiislove666]

虽然kimi只有月额度，但是你只重置25%只会让人感到小气😅 还不如不重置

[Kimi.ai @Kimi_Moonshot]

We're sorry for the rate limits and unnecessary quota consumption on the K3.1 Pro subscription plan. These were our issues, and we should have fixed them sooner.

Today, every subscriber on every plan gets a one-time usage reset, restoring 25% of their monthly quota (usage below 25% resets to zero), equal to one week of usage. Your usage bar will reflect this adjustment by end of day PST.

We've also adjusted the plan's parameters and are reducing unnecessary subagent calls and output length, so your quota goes further.

Thank you for your patience while we put this right.