CryptoPepper

vip
Age 3.3 Yıl
Peak Tier 2
No content yet
Some time ago, I developed an A-shares sentiment monitoring system (today I saw Lobster🦞).
The main focus was on a comprehensive model update for my previous single North-South position monitoring. Although there were constant bugs, the overall framework has been established—18 Agents simultaneously scraping sentiment data from Eastmoney, Snowball, Tonghuashun, and Taoguba.
Optimistic and Pessimistic Agents compete and score, and if their opinions diverge too much, an automatic arbitration is triggered. Additionally, the model iterates its weights through signal backtracking from T+1 to T+7
View Original
post-image
  • Reward
  • Comment
  • Repost
  • Share
Nearly 3,000 different TradingView indicators
CC ran it all day, basically downloaded the entire indicator library. Next, the sub agent will start combination testing and multi-factor testing.
You say AI trading is very difficult, but as long as your logic is correct and you have quantitative experience, things that once seemed crazy are no longer a problem.
View Original
  • Reward
  • Comment
  • Repost
  • Share
You could say that AI will amplify the gap between people, and my collaboration with AI is hundreds of times better than yours.
Someone who doesn't know how to use AI or dance with AI can be said to have underperformed the market.
View Original
  • Reward
  • Comment
  • Repost
  • Share
Can openclaw adapt to local customs?
In Nanjing, eat open-duck
In Guangzhou and Shenzhen, it's open-chicken
In Sichuan, it's open-rabbit
Always eating lobster🦞 isn't very happy either
View Original
  • Reward
  • Comment
  • Repost
  • Share
After trying Manus for 7 days, it automatically charged me $40, and now it's a monthly subscription.
It's like eating shit for 7 days and still trying to find a way to squeeze out a month's worth of effort.
Manus truly is a Chinese trick; if you don't Google it, you can't find the unsubscribe button (don't laugh, you’re the same).
View Original
post-image
  • Reward
  • Comment
  • Repost
  • Share
No one understands large models better from a consumer's perspective than I do.
1. For drawing, the top choice is the nano banana model (Free to use on Xianyu for 80 a year).
2. For videos, the top choice is seedance 2.0 (at least 2 dollars per minute, five videos are needed for effective results).
3. For engineering code, the top choices are Gpt 5.4 / Opus 4.6. Both are quite good, but the Opus ecosystem is better.
4. For understanding human nature, the top choice is Claude AI. I haven't seen any obvious hallucinations on Claude, and it understands human nature and politeness.
5. For AI rural
View Original
  • Reward
  • Comment
  • Repost
  • Share
The Middle East's data centers are the lifeline. Iran has really got it this time.
Especially as Gulf countries are undergoing transformation, targeting AI data centers = targeting the entire US AI industry = wiping out over 40% of US stock market share = US stocks plummeting in value.
Iran has truly reached the US's kill zone this time.
View Original
post-image
  • Reward
  • Comment
  • Repost
  • Share
Na Bai's Snow, this popular stock can also fall from 19 dollars to 0.9 and still not see the bottom, with a decline of over 99%
So, you need to understand
1. The world is a makeshift stage
2. The world is a plate
3. Most stocks are similar to memes
Even Buffett's investments have to shout build and require the CTO to recover costs
View Original
post-image
  • Reward
  • Comment
  • Repost
  • Share
Stumbled upon a bunch of GitHub projects that can make lobsters wear "glasses"
But it turns out they all require you to access their project's API and then start charging
So I did some statistics
Currently, the best one is still Brave (based on my own experience) — you can basically get what you need. It only requires adding a credit card, and you won't be charged until 1000 uses. So I recommend everyone to try this first
If you have large-scale needs, you can also try some different products (wearing glasses is a very important task, and those with the ability should compare each one)
View Original
  • Reward
  • Comment
  • Repost
  • Share
Over 1715+ production-ready skills, install with a single CLI command
Browser automation? Install
GitHub management? Install
Image generation? Install
OpenClaw is not AI, it's the operating system for AI
Skills are its App Store
If you, like me, find it too cumbersome to find skills one by one, then you must check out this GitHub repo ⬇️
View Original
post-image
  • Reward
  • Comment
  • Repost
  • Share
Yann LeCun is at it again
"If future AI still relies on current autoregressive LLMs, they will be very knowledgeable but still very stupid."
Hallucinations, difficult to control, only parroting—this is an architecture problem, not a scale issue
LeCun is betting on "world models"(World Models), and he says the next breakthrough in AI might come from Meta rather than OpenAI
So, what are world models?
View Original
  • Reward
  • Comment
  • Repost
  • Share
The first step of all openclaw automation: start generating your scheduled daily report
View Original
  • Reward
  • Comment
  • Repost
  • Share
Currently running automation processes:
1. VPS scheduled heartbeat <> push to GitHub <> local Obsidian agent knowledge base iteration
The process is: I directly send good articles to CZ( my agent <> record the article + analyze viral formats <> automatically record every 5 minutes and push to GitHub <> local Obsidian automatically pull <> improve the knowledge base
2. Super Stock System = two agents self-discuss daily pre-market and post-market research reports
Start research analysis at 9 AM daily <> fetch subscribed RSS feeds or various financial news <> open sub-areas, select specific secto
View Original
post-image
  • Reward
  • Comment
  • Repost
  • Share
Last week, I followed minkmax for several reasons:
1. Peking University's cost advantage: the price of green electricity in the western regions is one-third of that in the West, and token prices can be 16 times lower than Claude, which is very attractive.
2. China doesn't stack cards; it stacks distilled models and focuses on algorithms, thus reducing the unit computational power consumption from the inference framework model.
3. Minimax was boosted on OpenRouter, allowing many Americans/Koreans who can't afford it to also enjoy the benefits of the old Chinese model.
View Original
post-image
  • Reward
  • Comment
  • Repost
  • Share
Hamaney has always been the biggest internal obstacle to Iran's nuclear weapons development…
How much impact does the death of a spiritual leader of a theocratic sect have on the party? Especially considering that Hamaney issued a religious fatwa in 2003: explicitly prohibiting the production, storage, and use of nuclear weapons, stating that they violate Islamic teachings and are against humanity.
The Middle East is not over yet.
View Original
  • Reward
  • Comment
  • Repost
  • Share
All planes in Iran are fleeing.
The US strike over the weekend seems to be confirmed.
The decline in US stocks on Friday was probably a sign of提前逃跑.
View Original
post-image
  • Reward
  • Comment
  • Repost
  • Share
If you don't spend at least 2000 RMB on model tokens in a month, you're not really getting started.
It is recommended that all job postings use this standard as a reference.
View Original
  • Reward
  • Comment
  • Repost
  • Share
  • Pin