
BITWU
Feed
Feed
GPT-6 Astra 来了!
可能会比 GPT-4 → GPT-5 的升级更具有时代意义。因为 OpenAI 在试图把 AI 从“会回答问题的模型”,推到“能够独立完成现实世界复杂工作的智能体”。
真正走向能干活的 Agent 的重要一步!
我觉得得可以把过去几代 AI 简单理解成:
GPT-3:会写
↓
GPT-4:会理解和推理
↓
GPT-5.x:会深度思考 + 使用工具
↓
GPT-6 Astra:开始真正「干活 + 自主研究」。
所以GPT-6 最大的亮点和里程碑我认为是这几个:
① ARC-AGI-3 7.8% → 99.9%:未知任务的抽象推理能力出现巨大跃迁。
② AI 开始参与产生真实的新数学结果:从“知识消费者”走向“知识生产者”。
③ Computer Use + Agent 能力成熟:从“回答你的问题”进入“替你完成工作”。
Assumption:
Xiao Ming has been investing continuously from January 2026 to the end of July 2026, for 7 months in a row. What is the final winning rate expressed as 10 to the power of a negative number?
Damn it, I can't figure it out!
Who on earth came up with such a question that is so detrimental to unity?

🚨The US debt continues to expand, so what exactly should ordinary investors be afraid of?
High interest rates are not a one-time shock for highly indebted countries; they compound.
Suppose a government owes 100 units of currency, with an average financing cost of 2%, so the annual interest is 2 units. Now, as old debt matures and is refinanced at 4%, the interest becomes 4 units.
What about the extra 2 units?
The government can raise taxes, cut spending, or continue borrowing.
In reality, the usual choice is the third: debt increases → interest increases → fiscal deficit increases → bond issuance increases → the market demands higher term premiums → interest continues to rise.
This process is already visible in the US, and it’s the biggest headache for the Treasury and the Federal Reserve:
The larger the debt, the more policy options seem available, but in fact, the fewer viable paths there are.
So the so-called fiscal policy ultimately results in two outcomes: either US domestic residents bear the debt cost, or global holders of dollars or dollar assets share the burden.
Anyway, expecting the government to voluntarily tighten its belt is probably harder than expecting ordinary people to hold Bitcoin long-term!
Therefore, current investments should be considered based on three principles:
1️⃣ Hold long-term fixed income assets cautiously, favoring short-term over long-term debt. If the risk compensation isn’t high enough, I’m personally not very interested.
2️⃣ Continue allocating to gold, Bitcoin, and high-quality stocks with pricing power, which align with long-term logic.
3️⃣ What I personally consider most important: use low leverage.
In an era of high fiscal leverage, ordinary people are especially tempted to increase leverage too. Many thought this way during Japan’s real estate bubble, before the 2008 US housing crisis, and even more so in 2021 when crypto investors borrowed stablecoins to increase positions.
As the government becomes bolder in leveraging, individuals should become more cautious about leverage.
In short:
Over the next decade, you only need to ensure you hold some gold, bitcoin:native, and high-quality productive assets long-term, while controlling leverage and avoiding excessive exposure to long-term nominal debt claims. I think that’s enough.
The rest, let them keep printing slowly!

Ansem 看好 Robinhood 股价第四季度会大涨!
我个人也很喜欢 HOOD 因为相对而言,他很可能不是大家以为的另一个 Coinbase,而是第一家真正把 Crypto 基础设施吃进传统券商估值体系里的公司。
这件事的想象空间很大,Q4 会不会大涨我不知道,但是美股里加密概念应该重点关注下 $HOOD
能否快速突破 $110–120 是分水岭!
there is always a lag when tradfi companies add new crypto integrations bc the suits dont see the data until late
believe $HOOD will be a great bet here as chart already is coiling off of the bottom & they are adding a significant new revenue line w/ their layer 2 blockchain @RobinhoodCrypto
ATHs +50% in Q4









