STi克莱小汤

If you want to make quick money from investment, the basic outcome is to give away money,
If you want to earn long-term value from investment, it often may bring surprises instead.
If you want others to take over your investment, you will ultimately be swallowed by even darker hands.
If you want to learn and share together through investment, it will be easier to accumulate value and merit.

In short, you must be true to your conscience; only then can some invisible forces be activated in the midst of it all. This is also an interesting deep-level philosophy, the way of balance.

$BTC $ETH $BNB #加密市场观察

#币安神算子 Isn’t this a new event? Where did all these people with so many points come from? Could it be that points earned before #Predict were included? What’s the point of playing then? If you’re running a points event, everyone’s points should be reset to zero so we can start from claiming check-in rewards.

Binance Square is starting to remind me of Bihu back in the day. It’s a real shame what happened to Bihu in its later days—they just failed to crack down on garbage posts and low-effort fluff. That drove away a lot of quality creators. A thoughtful, in-depth article might be thousands of words long, but it couldn’t compete for traffic; split it into 10 shallow posts and it would get more traffic.

Back then, I told people around me that Bihu wasn’t far from shutting down. Later, Bihu tried to save itself and came very close to turning things around, but Gulu didn’t want to keep going.

Square was really good when it first started, but lately it’s been getting stranger and stranger.

Anyway, that’s enough. There’s no point saying more.

$BTC $ETH

My connection with $TSLA goes back to 2017, when I helped manage the charging and power systems for the development of the Model 3. Tesla’s lead R&D engineer struck me as someone who was extremely passionate about his work and highly professional. I personally wasn’t very interested in mechanical R&D, even though I had published three first-author SCI papers and, by the time I earned my master’s degree, met the standard for a PhD graduate.

But I really disliked technical drawings. What caught my interest instead were Tesla’s financial reports and the secondary market. So if we’re talking about crypto, my connection with $BTC really began with Tesla. It was also thanks to that collaboration that, around 2019, after the fires dealt the company a major blow, I wrote to several large institutions saying that I believed Tesla had tenfold upside.

Of course, some institutions took it as a joke, especially some domestic ones, which may have been rather arrogant in their mockery. But one top-tier American investment bank replied very seriously and discussed it with me. At the time, I felt they treated opportunities with great care. You can see the same thing in venture capital: European and American investors seem more willing to get hands-on and try new things, while some Chinese investors prefer to flock in once success is clearly within sight. That’s why I feel some Chinese institutions are pretty average at going from 0 to 1 (with the exception of state-backed institutions), but very good at going from 1 to 2.

But against the broader backdrop, our country’s pace of development has already been a miracle. I believe things will get better and better, and that our openness to innovation and spirit of exploration will keep improving too.

I vaguely remember what a well-known financial adviser said to me in 2018: “Why should I bring on a career-switching newcomer who’s just returned from overseas? That’s too risky. I’d prefer someone with a finance background.” (That company has since gone bust.)

By contrast, the exact words of the European-region managing director at the time—who is now the CEO of a top American bank—were: “You have experience and judgment from mechanical R&D that people who studied finance and investment will never have. As for financial and investment knowledge, if you’re interested, two years of self-study will be enough. Don’t ever give up on pursuing what you want. If you want to go to JP in the UK, you can list me as a reference.”

Those words subtly but truly showed two completely different attitudes toward risk. I can’t say who was right or wrong—after all, there are plenty of charlatans here at home too. But the two people clearly seemed to be on completely different levels.

#股票财报季

This is a very important day. Let's give it our all together ⛽ and break through!

#股票财报季 Stocks sometimes lose a lot of their upside potential because of earnings reports, but having those reports often gives us more to think about when positioning ourselves early.

That’s because the data is delayed. For example, the premium memberships driven by A Record of a Mortal’s Journey to Immortality may not show up until Bilibili’s Q4 earnings report. So Bilibili’s stock tends to rise later, with dips along the way.

This is a crucial opportunity for Bilibili this year—though of course, we don’t know whether BLG will drop the ball.

$BILI.US $NVDAB

There are two kinds of people who build products and invest:

One kind does it purely to make money; the other does it to improve their own efficiency or for fun, sharing it with others along the way. If it attracts a large user base, that's a matter of fate; if it doesn't become wildly popular, having some high-quality users and institutions is still great. It's purely about improving one's own efficiency.

My friends and I are the latter. A product can improve our own work efficiency by around 50%; several products together can boost it severalfold, or help us learn new methods. In this situation, it doesn't really matter how many people use them; it's more about practicing the process. It's not a negative-cash-flow model or one that burns through money recklessly. As for investment views, styles, and approaches, I share my judgments months in advance. Whether people believe them before the trend arrives is a matter of their own judgment. It doesn't affect the short-term line, and therefore doesn't affect the overall trend or the changes in the Dao.

Many VCs suffer serious losses because they always like to look at decks full of hype, indulging the dopamine rush of their own fantasies, while being more wary of decks for products that have already been completed. If a VC keeps losing money, perhaps it's time to reflect on its decision-making process.

$BNB $NVDAB $BTC #Aİ

Now, AI Agents can automatically use #UniPublisher desktop software as an AI entry tool to create and distribute creators' posts and articles across multiple channels as needed. UniPub itself can already become a super-skill for agents. Cross-pollinating high-quality creators across each other's platforms. #Agents #AI #Web3 skills

AI Agents can automatically use #UniPublisher desktop software as an AI entry tool to create and publish creators' posts and articles across multiple channels, attracting high-quality creators to each other's platforms. UniPub itself can already become a super-skill for agents.

To my surprise, over 1,000 users have already downloaded and used it. Recently, I’ve had quite a few products in development and optimization, and several apps are under review. I haven’t had any time at all for marketing or product optimization, and the algorithm model is still being optimized. I’m still pleasantly surprised by these download numbers. $MSFTB $NVDAB $GOOGL

Hahaha, this image is so relatable. A lot of FDEs really will be like that in the future. #AI #PROMT But people who are good at prompting can actually be really good at Skills too. The products they come up with can be pretty good as well.

Binance Square isn’t going to use Twitter’s algorithm as-is, is it? 😅 That dumb algorithm is something else.

$BNB $BTC

Currently, #AI Web3 all-in-one creation tool UniPublisher supports 17 channels, including @币安广场 Binance Square and Telegram. I think the newly added feature for syncing both short posts and long-form articles to Telegram is really convenient.

You can sync messages to any number of groups you manage at the same time. UniPub has also activated its #Telegram long-form article system. This has always been an awkward feature on TG: without a UI, publishing is difficult. Now, people who enjoy running TG channels can greatly improve their efficiency. We’re also expanding support for Twitter based on user feedback, from 1 to 5 accounts.

$BTC $ETH $BNB

This day is incredible $BTC

Tokenization bundles lots of small assets. If it takes off, it’ll have a very familiar feel. $SOL

In learning and self-discipline, compare yourself with those above you; in joy and happiness, compare yourself with those below you. This, too, is a way of balancing yin and yang ☯️.

An open-source project for organizing Skills, Skilldocker makes it easy for AI users to organize, develop, create, and diagnose their own AI Skills. It also adds a learning section for beginners, making it easy for newcomers to quickly understand everything. With AI skins popping up everywhere, the texture of this WIN98 skin feels so satisfying haha. Let’s play around with the open-source project and make something fun 💡 I feel like we’re about to connect some things.

An open-source project for organizing AI Skills, making it easy for AI users to organize, develop, create, and diagnose their own skills. It also adds a beginner-friendly tutorial section so newcomers can quickly understand how everything works. With flashy #AI themes everywhere these days, the texture of this Win98 retro skin feels so satisfying haha.

A few trending topics from the past couple of days:

1. Abusive language in conversations with Claude is now considered a policy violation. Anthropic updated its usage policy on Wednesday, stating that starting November 12, sustained and unnecessary insults or abuse directed at its models will be prohibited. Normal venting of negative emotions, dark creative writing, and red-team testing are unaffected. If a violation occurs, Claude will end the conversation immediately.

2. OpenAI published 722 mathematical paper drafts, then retracted three of them the following day. The papers claimed progress on 372 unsolved problems and were all produced by an internal OpenAI model that has not yet been publicly released. The core conclusions of 162 papers passed formal verification in Lean. The retractions were due to notation errors.

3. Eleven companies pledged $2.4 billion worth of computing resources to the U.S. Department of Energy (DOE). Nvidia is contributing $1 billion, AMD $500 million, and OpenAI $200 million. All contributions will be in the form of computing credits rather than cash, supporting the DOE's Genesis initiative.

4. Nous Research raised $90 million at a $1.5 billion valuation. The round was led by Robot Ventures, with participation from Nvidia and Samsung. The company said its open-source agent Hermes has been cloned 24 million times, accounting for about 2.5% of global AI token usage.

5. Google unveiled a new embedding model that brings video, audio, images, code, and text into a unified space. Called EmbeddingGemma 2, the model has 740 million parameters (740M) and can run locally on phones, using only about 567 MB of memory on the Pixel 11 Pro. It is open source under the Apache 2.0 license, and its weights have been uploaded to Hugging Face.