Focus on 1) Technology and Dividend Blended 2) Drawdown protection. Check highlights and articles. No investment advice. Now building ModelDock
Iceland
Joined October 2011
- Tweets13.5K
- Following774
- Followers17.1K
- Likes17.4K
Pinned Tweet
What each budget can buy?
The Pareto front of local AI right now:
modeldock.net/insights/local…
毕竟一个成年女性，有千丝万缕的社会关系和难以捉摸的背景：
有可能是竞争对手派来的，
有可能是来分家产的，
有可能是老婆闺蜜，
竟然还有可能是个间谍，
万一要是朝鲜间谍，就万劫不复了，
富豪们想了想，禁不住轻抚钱包里爱泼斯坦的遗照。
马达尼当上纽约市长之前的共和党:
"他要搞社会主义，他会让纽约万劫不复。"
马达尼当上纽约几个月后：
1. 万斯出来领功，表示是特朗普政府让纽约变得更好（屁）
2.普通人视角：这个社会主义，搞普惠日托，搞保障房建设，搞市政建设提供工作，控制犯罪，这就是你们共和党一直批评的社会主义？
3.共和党视角：他居然给穷人福利，用白花花的银子让穷人吃饱穿暖，有地方住，银子怎么能用来造保障房而不是数据中心，造孽啊。
@OpenDesignHQ Hey OpenDesign, your Stripe/pay link is off, cannot pay to subscribe.
美国的国民性格里，最特别的一点，就是自大。
明明全国也建不了多少数据中心，但是满脑子想的都是怎么卡中国的脖子。
老登政治家这么想也情有可愿，年轻一辈的AI公司也是这个调门。
吃不下，可是真的会噎死的。
一段往事，美国20年代对苏联禁运，30年代陷入经济危机了把工厂整个往苏联卖。
如果我们客观的去考虑，
OAI如果能保持今天A社2/3的营收就很好了，
那么OAI的算力开支也应该是A社的2/3左右，即4000亿左右。
四千亿也已经很多了，但是还是远远小于OAI这些疯子承诺的1.4万亿。
也就是，现在至少有一万亿的overbuild, 即使不考虑 deepinfer / Strata 带来的生产力暴增。
不知道这是哪个幸运儿负责这一万亿建设。
Big if true..
这位大神宣称可以用更智能的LLM去读取/理解/修改LLM的权重，从而跳过训练，直接让低阶LLM更聪明。
我看了看手边的Opus5.5 和Qwen Flash Next。
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math.
rikiyat.github.io/blog/weight-sp…
github.com/RikiyaT/lm_und…
huggingface.co/Qwen/Qwen3.5-0…
huggingface.co/RikiyaT/Qwen3.…
我认为OAI已经是期货死人了。
巨额commitment + 越来越高的融资价格 + 日渐收窄的发展空间。
OAI：我看中大众AI的长远发展。
GOOG/META/字节：巧了，更巧的是我们还真有用户数据。
Indonesia, where Dax revealed the location of those racks are provoking.
Could it be that in the past year/ months, China CSPs set up shell companies in SEA to run their training/ inference, only to find that they only need a fraction of the setup now as they 1) drop training 2) run DeepSeek/ streamable flash models?
Suddenly there is a -80% reduction in metal demand in SEA?
the reason there's no clarity with gpu markets is because it's completely chaotic right now
lot of effectively fake listings, many layers of brokers etc.
but because it's crazy, most people looking for GPUs aren't very good at it. unless they have a strong network they're not going to find out that 3 racks in indonesia are idle for the next 3 months
it's a good time to be someone who can hustle and find a deal