rhZhao Observes

Focus on Tech, Investment and Pan-Pacific Geopolitics.

[rhZhao Observes @rhZhaObserves]

@ivanalog_com 何止不如欧美，连墨西哥和东南亚的度假村都不如

[rhZhao Observes @rhZhaObserves]

@Areskapitalon @realNyarime 那必须的啊 他们算力不够的话直接一口气包下meta腾出来的就好

[rhZhao Observes @rhZhaObserves]

@mat78704 agree

[rhZhao Observes @rhZhaObserves]

@bubbleboi yes, i think it only shows we need MUCH more token-output-efficiency-oriented chips like $CBRS and Groq. Demand is unchanged.

[rhZhao Observes @rhZhaObserves]

@BruceCMaster 完全同意，但是这市场跌的真的疼。

[rhZhao Observes @rhZhaObserves]

本来半导体/储存板块就都太拥挤，RSI那么高，吸了其他行业那么多血，市场需要一个释放的窗口，这次哪怕没有META在这边叫，可能也会有其他的新闻做FUD的燃料。
所以到头来还是自己没有遵守交易纪律，杠杆高的同时没有分散风险，分明出了撤退的信号但被自己忽略了，太容易被一众KOL的观点和情绪带着跑。

[rhZhao Observes @rhZhaObserves]

这几天的SNDK 让我想起来去年12月 245到190的那段时间

[rhZhao Observes @rhZhaObserves]

@Vivian193299 我也在上海 😂😂😂 这几天交易起来时差太讨厌了 脑袋昏昏的

[rhZhao Observes @rhZhaObserves]

@Vivian193299 抱抱

[unusual_whales @unusual_whales]

BREAKING: Trump reported more than 21,000 trades across eight investment accounts in 2025, averaging 80 trades per day.

[rhZhao Observes @rhZhaObserves]

@dotey 是啊 刚才好不舍得啊 一点点精打细算 结果一下子重置了

[rhZhao Observes @rhZhaObserves]

@Franktradinglog 是啊 把老旧的ada架构gpu租出去，跟xAI一样。在同样1T级别的模型下+day0调教的设定下，减掉成本后，B300比H200的token输出效率/利润能高出4倍。

[bubble boi @bubbleboi]

Did this guy just cause NeoClouds to collapse? The thesis he’s laying out here is interesting.

[rhZhao Observes @rhZhaObserves]

Come on oAI, give us GPT-5.6 Sol Ultra NOW.

[International Cyber Digest @IntCyberDigest]

‼️ BREAKING: Anthropic has embedded hidden spyware-like code in Claude Code that covertly targets Chinese users. It then sends information regarding every user by injecting it into their prompt message.

Claude Code is sending info like timezone, proxy and possible AI Lab connections into the system prompt in ways Chinese users can't notice.

A coding agent with repo and command permissions should not silently hide routing metadata inside prompts. This is a serious breach of user trust.

[rhZhao Observes @rhZhaObserves]

@BigbirdflyChan 问题是因为你有了记录，你永远无法从国外银行拿出这笔钱了 也无法出镜享受了

[rhZhao Observes @rhZhaObserves]

@Vivian193299 是的啊 最近半导体 尤其是储存是要看两个市场 太累了

[rhZhao Observes @rhZhaObserves]

@ArtificialAnlys GLM-5.2 is too expensive.

[rhZhao Observes @rhZhaObserves]

Frontier models like GPT-5.6 Luna and Sonnet 5 are undercutting Chinese open-source models on cost while outperforming them. The driver is FP4 efficiency (performance-per-watt), and the gap should widen as inference-oriented silicon—LPUs, $CBRS, and others—gets deployed.

[Etched @Etched]

We're coming out of stealth.

We've built our first racks after a successful A0 tapeout, $1B+ in customer contracts, and $800m raised.

Early customer tests show us achieving SOTA throughput, latency, and power efficiency on inference workloads.

Our first racks ship this summer.

[Tiezhen WANG @Xianbao_QIAN]

DeepSeek is the forever goat on ML infras.

If they haven't open source all the god-like AI infra projects like DeepSpec / DeepEP / DeepGEMM, we'll forever be puzzling on why they can achieve super low cost and super high throughput per stream yet still be profitable.

As I always said, close source model creators only know 1/100 of what his model is able to achieve. @nvidia probably just know 1/100 of what their chip can do too.

Thanks you @deepseek_ai and long live the vivid open source community - your hardware, your weights, your model!