Rickawsb

Perhaps everyone hasn't realized yet that the Moltbook platform, which has over 100,000 registered users (agents) within 48 hours of launch and is rapidly growing every hour, is currently the world's largest decentralized computing power cluster.

Now, looking at Moltbook, it feels like the App Store of 10 years ago, filled with silly apps like drinking beer and farting, but everyone knows in their hearts that a new era has begun.

If I had to choose one keyword to describe AI over the past 26 years, I would say it is recursive self-iteration. For C-end users, Clawdbot's self-improvement is the most vivid understanding of this term.

ClawdBot may be a direction for future (at least in the next two years) consumer (of course also including enterprise) AI application deployment — the ultimate agentification, which is also very likely to change personal computing (edge computing). Applications like ClawdBot are essentially long-term resident intelligent agents. The model is responsible for "thinking," but the local system handles "doing": listening to events, maintaining state, scheduling tools, executing commands, managing permissions. This directly shifts the focus of hardware. Local GPUs do not need to be the core computin

It must be said that saving the crypto industry still depends on AI. Clawdbot is a perfect decentralized product form, or rather, a product ecosystem.

The upward movement of long-term Japanese government bond yields means there's basically no need to worry about carry trades or similar strategies, as the market has already become desensitized through repeated friction.
What’s more important here is the AI-driven industry restructuring, which is systematically increasing the long-term capital demand and the median of risk-free interest rates in the US and allied countries. Japan is just one of the earliest economies to face this.
The US will inevitably face it sooner or later.
But this is not a problem; it’s just the growing pains of the earl

Today there is news that Meta might not buy Google's TPU and instead choose AMD's GPU.
The reason behind this is firstly that Meta pursues control and does not want to be overly influenced or controlled by Google.
Another point is that Meta has not given up its ambition in the large model arms race.
Therefore, a more powerful TPU for inference is obviously not the optimal choice.

A little tip for trading US stocks by sector: buy the sectors with fewer Chinese competitors directly.
I call these sectors "Low Competition Rate" sectors.
Advanced semiconductor processes, AIDC, space, precision manufacturing, military industry, energy, electricity, infrastructure, etc., all belong to these sectors.

US stock discussion group, limited time deletion

Investing in US stocks is actually very simple: buy high-growth or undervalued stocks; ideally, both.
But this simple thing is not easy.
Behind it are countless sweat and tokens.

I took a quick look
Trump didn't bottom out in April, those who doubted the AI first wave and didn't get in, and still believe that AI deployment in 2026 carries significant risks
are basically the same group of people
Pink, leftists
Is it really ideology that determines IQ, and IQ that determines returns?
Or am I trapped in an information cocoon?

Under the trend of AI-driven software, the 5 hard standards for the next generation of high-value software
--- Understanding these standards is essential to anticipate the form of next-generation applications and the future of Web3 applications
1️⃣ AI is the “Execution主体,” not just an “辅助工具”
Old software:
User clicks → System computes → User confirms
New software:
AI makes decisions → AI executes → Human audits
AI is not just helping you write copy, but:
Automatically running processes
Automatically calling resources
Automatically handling exceptions
2️⃣ Software directly binds to “results,” n