Jingle bell 初号机

Yarrow is a multi-agent simulation engine built on policy simulation.

Ask him any questions you want to know
Then you will receive calibrated probabilities + a complete evidence chain and reasoning

Truly calibrated predictions ☘

Who said fitness has to be done only at a gym?! 🏋️
I’m going to a hot pot restaurant... Let my stomach... get a workout! 👀

#GateVIP fitness buddy

Every sentence you say to ChatGPT, OpenAI can actually see.
The contracts and business plans you send are also stored on other people's servers.

Is it possible to keep the AI on my phone,
so that the data never goes through any server from start to finish?

The Manta Network team made a new product, Incognitive 🌊

Manta itself builds a ZK privacy-focused L2.
Privacy is their core specialty.

Now Manta

Yesterday, several people privately messaged me asking whether they can join the ByteDance group. They said they don’t even need to talk (and some also want to join to get to know Bao Zong and Yang Shen).
To ensure everyone’s safety, I definitely turned them down 😂
But I can summarize the daily highlights from what people discussed in the ByteDance group.

Yesterday’s TLDR: China is moving residents’ wealth from banks and real estate into the stock market—this is a paradigm shift, not a minor policy tweak.
The US in 08

Zora keeps coming up with new tricks again!
They’ve creatively launched an agent creator product~

You might remember Zora. In 2020, they did NFT minting (my memory of Zora is that later they got involved in pumpfun on the Base chain—at one point, it accounted for 67% of token creation on Base + Solana. The pump was also intense, and $ZORA back then it surged 18x from the bottom

Now Zora

Binance’s US stocks launch goes live for 30 days: $1 billion floods in—equivalent to an average daily inflow of $41 million, and total trading volume breaks $3 billion

Over-the-top numbers can only mean this: US stocks are booming + demand has been pent up for too long 🤯

So do you know what people in the crypto world bought after getting the US stock entry point? “$1 billion” — what did they put it into? 👀👇📊 ——48% went into semiconductors

More than 7,000 eligible stocks, 71%

7.2 Daily Byte Group Highlights

TLDR: Retail investors lose money not because they don’t try, but because they’re trying in the wrong direction. Long-term value investing in “old-timer” stocks, quick in-and-out trading for tech stocks—if you do it the other way around, you should actually be able to make money.

1. Retail investors’ approach is backwards

When tech stocks rise 30%, quickly take profit; when “old-timer” stocks fall 20%, stubbornly hold and “wait for a rebound”

I got the package chief to go live on #币安广场 tomorrow night 👀

Over the past few days, I contacted you and helped arrange introductions to exchanges and trading products for the package chief (stock DEX). If there aren’t 10, there are at least 8.

In the end, Leto chose to go live exclusively on Binance Square. My personal guess is that it’s because—Binance Square has the largest real trading user base in the entire industry, and this segment of users overlaps very little with Twitter’s audience.

For those who are just getting started with 2M

Recently in the ByteDance group chats, aside from joking about how everyone in the neighboring US stock groups feels “free,”
I, as the group admin of the crypto community, haven’t been able to be free.

Everyone doesn’t know: actually, earlier I had a chance to choose the AI track.

I started this account during the same stage. That afternoon, after the AI expo in Shanghai just ended, I went to dinner with “Digital Life Kazek” @Khazix0918. Everyone was very excited.

At that time, Kazek’s AI

Over the past few days, ByteDance Group’s highlights of key remarks

1. How to buy in biotech (DYOR, no investment advice at all!)

The simplest approach: blindly buy XBI (a small-cap biotech ETF), GNOM, and IDNA (even smaller, focused on intracellular biotechnology). XLV is a large-cap healthcare value stock; it has weaker growth—don’t choose it.

Biotech is the sector most likely to benefit from AI

1. He already hit me today; tomorrow he probably won’t hit me again
2. If he hits me again tomorrow, the chances of him hitting me the day after tomorrow are even lower
3. He beat me up in the morning, but in the afternoon he actually didn’t hit me—maybe he still feels sorry for me
4. He’s been hitting me for several days in a row. Tomorrow he’ll surely make it up to me. He can’t keep hitting me continuously; I can’t leave
5. I was almost beaten to death, so I ran. Otherwise, wouldn’t all the times I was hit before have been for nothing?
6. What if, after I ran, he suddenly takes off in his career?
7.

ByteDance Group’s highlights in recent days

TLDR:
1. Breakthrough in advanced packaging hitting new highs; storage chips face FUD, but the logic hasn’t changed
2. For biotech, you should open a supermarket, don’t bet on a single product
3. Taking profit and cutting losses are essentially the same thing
4. Humanoid robots are starting to become worth paying attention to

I. Advanced Packaging
JCET has hit new highs, and J-Sun Technologies is close to it too.
The advanced packaging sector has very strong momentum

II. How should SK Hynix’s U.S. stock ADR be priced (it seems it’s already priced in)
Hynix U.S. stock

Hynix Korean stocks and U.S. stocks are temporarily deviated, and there is no relevant connection.
I’ve seen a few big V accounts in the crypto space who have been trapped after moving into stocks.

“Crypto mindset” is hard to change.
Many experiences from crypto can’t be applied directly to the stock market.

For example, thinking that if a stock has risen too much it must fall, so you should short.
For example, thinking that if a stock is released it must fall (Zhipu didn’t fall).
For example, thinking that the price spread between Hynix’s Korean stock and the U.S. stock converges = arbitrage.

7.14-7.16 Byte Group Essence
These days have fallen very badly, so we must update our overall investment thinking more than ever.

1. The biggest enemy of retail investors: taking unrealized losses as an anchor
Whether to buy or how much to buy has nothing to do with how much you’re currently up or down: even if you lose more, it will not make a stock become cheaper.

If you’re unwilling to cut losses, then it’s not really because it’s “cheap” that you buy more—you’re simply unwilling to accept the loss.

ByteDance Group Daily Highlights 7.17 - 7.20

1、Why did we drop this time? Don’t ask about individual stocks—ask about the macro.

The recent pullback in AI hardware stocks that had surged earlier wasn’t a problem with any single stock. It was the whole sector de-leveraging. Two sparks: “TSMC raises capital expenditure, triggering concerns about excess capacity” + “CXMT Memory’s IPO siphons off funds.”
In China’s A-share market, it’s simpler— the only names with earnings are overseas-linked. Overseas-linked names are falling along with the global market.

Bad news: encryption mindset 0.079% SmartF 3.1%
Good news: in Stock stock mindset 0.091% SmartF 1.3% 🤡

ByteDance group daily highlights 7.21 - 7.31 (Ten-day compilation)

The recent downturn has been pretty brutal. I’m storing a 55% drop over one month.
Finish the crypto market in nine months’ worth of activity in one month.

1. Hynix breaks records with its earnings report, but the stock plunges.

Operating profit year-over-year +557%, revenue +257%, operating profit margin 76.3%—both setting historical records.

Yang Ge previously said: "Profits—this kind of first-order data—has absolutely no relation to the stock price"

How do we turn insurance companies into interest-earning assets we control?

Imagine a traditional insurance company (say, an auto insurance or fire insurance company) collects 1 billion yuan in premiums this year. But they’re afraid that if a major disaster happens, they won’t be able to pay out, so they have to set aside 200 million yuan to go buy insurance from a bigger “reinsurance company,” spreading the risk out.

This “reinsurance” business is actually extremely profitable and guaranteed to do well in both dry and wet seasons. In the past, only Wall Street giants and top-level financial dynasties could play this trillion-dollar money game.

Recently, I saw that #re can completely smash this barrier. It uses blockchain so that USDT we hold can also be directly deposited into a regulated reinsurance market, letting you get your share of the profits.

And the收益 of Re Protocol isn’t conjured out of thin air—it comes from real licensed insurance companies paying premiums. So if the crypto market drops or the stock market crashes, is it related to whether we choose to buy insurance? No. Therefore, its underlying assets have almost “zero correlation” with the broader market—making it a very scarce defensive asset 👀

Two ways to play the protocol—choose according to your needs (must-read for beginners)
Tailored to different risk preferences:
1. reUSD (choose this for stability): principal is strictly protected, and liquidity is excellent—ideal for defensive players seeking steady returns
2. reUSDe (choose this for more upside): locks funds for a longer time and you need to bear some underwriting risk, but in return you can potentially reach the highest up to ~23% APR

RE brings the most stable and most profitable “reinsurance LEGO blocks” from traditional finance onto the chain. If you’re looking for an Alpha asset with real business backing, the Re ecosystem is worth taking a look at
#re $RE #RW #DeFi #ReProtocol

【About Merchandise】：So many people say they’re jealous, but you really don’t have to ~
1、As mentioned in the video itself, I only received a few of the items shown in the video, not all of them

Welcome all AI founders
Welcome VC friends to share and recommend cases to each other
Welcome friends who are interested in AI to join the conversation!👀