SN

Recovered a $1.5M drawdown to 27.9% CAGR

2022: -13.41% 2023:+59.31% 2024:+49.18% 2025: +29.84% 2026: 14.48% so far.

Scaling information edge into Alpha

[SN @ScaleAlphaHQ]

Built a tool to optimize stock trading strategies and execution — looking for beta testers!

First, thanks to X— I’ve discovered so many great ideas here.

Why did we build this tool?

A few years ago, I lost six figures trading stocks. It was a five-year dark journey — slow but steady losses. I tried everything, but eventually realized the problem wasn’t a lack of ideas (good ideas are everywhere). The real issues were:

Lack of clarity in trading logic — I constantly mixed up short-term vs. long-term trades. I’d try to catch a bounce but end up becoming a loyal shareholder, or cut a position to reduce risk and then miss the re-entry, or believe in the fundamentals but actually trade based on sentiment.

Poor execution and discipline — I’d skip setting stop losses and just hope things would work out, or I couldn’t resist going in with oversized positions.

I tried building a quant trading system, but found it impossible to fully translate diverse, discretionary trades into something purely systematic. I also wasn’t great at building pure quant strategies — a tiny parameter change could flip a strategy from profitable to unprofitable, and backtesting only filters out the obviously broken ones.

Eventually, I landed on a simple (silly) solution:

Do research only on weekends; trade only during the week

Before each trade, hold a small “ritual” — talk through the logic as if I’m explaining it to someone else

Pre-define stop loss and take-profit levels

Limit screen time during market hours

It wasn’t optimized for maximum returns, but it worked for me. Thanks to the bull market in recent years, I earned back all my losses.

After the release of o1 in 2024, I realized I could build a new kind of “quant” system — using reasoning from LLMs to support discretionary trading. I started building this with a few friends in April, began live testing on July 15, and as of October 3, the return stands at +18.5% (with some luck involved, of course).

What does this tool do?

It generates a detailed, personalized execution plan for any U.S. stock — including entry points, stop loss, profit targets, position sizing, and even options strategies.

It can also discover customized trading opportunities and update them as needed (though most of the time, it simply says: no opportunities right now… which is useful too).

Both types of reports come with full explanations. You should not treat this as financial advice or blindly follow it — but as a rational second opinion, it’s solid.

How is this different from just using GPT?

We go beyond traditional financial data. Our system scrapes and reasons over real-time unstructured data — Twitter, TikTok, Reddit, Snowball, news sources, and Congressional trading disclosures.

We’ve built a TB-scale database and a large knowledge graph to model stock relationships.

On top of that, we’ve developed 500+ historically validated alpha-generating factors (and counting), and built custom analysis agents to interpret them — all wrapped in a personalized interface.

In live trading, this significantly outperforms GPT, as expected.

Why us?

I used to work as a quant at Merrill Lynch and State Street, then built models at Ant Group. After ChatGPT launched, I had the opportunity to train LLMs and built two B2C AI products. Now I’m fully focused on trading and building this startup. (If you’re in the Bay Area — happy to grab coffee!)

My co-founder is a college classmate who runs a profitable crypto HFT fund (we plan to support crypto strategies later too). Our data lead has deep experience managing PB-scale systems. Two of our advisors are working at Two Sigma and Jane Street.

If it’s making money, why go public?

Yes, the system is generating returns (we started with 5 figures, and now deploy 6 figures). We were lucky to catch runs in INTC, WBD, APP, TSLA, BTU, and others.
More importantly, the tool helped me avoid some bad decisions — like holding too long or selling too early. It’s not always right, but it often provides a helpful outside perspective that improves outcomes.

As for starting a fund:

We’re still early — I think we need a much longer track record to prove consistent returns.

Managing other people’s money is a huge responsibility, and I’m not ready for that without much much more validation.

So why not just quietly make money?

We are — but I don’t want to stay silent.

My trading style is mature, and most market opportunities don’t even apply to me. If we can build a product that helps others while doing something meaningful — that’s far more fulfilling. I always wished I had someone to trade alongside. A second opinion can make a big difference.

And if it’s genuinely useful to others, I get to earn some side income too.

What do we need?

We already have 121 internal beta users, and we’re looking for 80 more active traders who:

Trade individual U.S. stocks (not just long-term holders — though long-term investing is a great way to make money)

Are willing to test the platform for 2–4 weeks (completely free)

Can share feedback after using it (via Zoom or over coffee)

What you’ll get:

Free access through the end of the year

Lifetime discount when we officially launch

Direct influence on the product roadmap

Website (PC) : https://t.co/iGYzulVnGA

If you're interested, DM me to join the beta.

If you're an influencer, we’d love to collaborate — your followers can link to your strategies and generate tailored execution plans. We’ve also built legal and compliance layers to protect both sides. This increases follower engagement and opens up a new revenue stream for you.

[SN @ScaleAlphaHQ]

We tested our engine on world cup 2026, 9 bets, 8 wins! Include the Spain Tie.

Model says X, market says Y, if gap > 9%, we bet.

Test it out for free: https://t.co/nama55Kroa

[SN @ScaleAlphaHQ]

@bubbleboi this is leopold view years ago.

[SN @ScaleAlphaHQ]

Every bubble is a combination of "a solid story that cannot be falsified in the short term + a leverage mechanism exploiting regulatory loopholes".

Technological revolutions are the most dangerous because their fundamentals are the most solid and the market reflexivity is the strongest, leading to an upward spiral of increasing profits and rising stock prices. At this stage, it is not just the stock price that is overvalued, but also future earnings. When future earnings fall short of expectations and the leverage mechanism is punctured, the bubble bursts.

Why would future earnings fall short of expectations? Because while productivity can explode rapidly due to a technological revolution, relations of production inevitably take much longer to change and will definitely lag behind productivity.

I believe that in the short-mid term, revenue will start flowing toward inference providers. In the long term, it will depend on the strategies of foundation model companies: they must transition from "selling tokens" ➔ "selling workflows" ➔ "selling AI operating systems."

Only an operating system forces users to use proprietary models instead of open-source ones. This way, in the future, when frontier cognition (closed-source tokens) becomes too expensive to train and begins to depreciate exponentially toward cheap commodities (open-source tokens), these foundation models will still be able to retain a certain level of revenue.

Inference providers will eventually become public utilities, because both computing power and cheap commodities (open-source tokens) will depreciate exponentially, ultimately leaving everything dependent solely on the cost of energy.

[SN @ScaleAlphaHQ]

https://t.co/sFkCXF46OG

[SN @ScaleAlphaHQ]

@aleabitoreddit Bro trying to nail both timing and direction at the same time. If anyone could do that, they'd already be the richest person in the world.

[SN @ScaleAlphaHQ]

@Nostre_damus Yes, a large number of anomymous accounts are currently trembling with fear. Now people only price it as a tradable asset.

[SN @ScaleAlphaHQ]

We did a scenario analysis on May 27th.

[SN @ScaleAlphaHQ]

@AryJeayBackup hyper-inflation man, check Venezuela stock market

[SN @ScaleAlphaHQ]

@MarioNawfal Isn't that because of hyper-inflation?

[SN @ScaleAlphaHQ]

@GingkoPT Even if they open it TODAY, it will not go back to $65. Too many ppl just don't do any research. Insurance, tanker rent both will stay high in next a few months at least, this is real world business not paper trade.

[SN @ScaleAlphaHQ]

https://t.co/b8c7H5sNei

[SN @ScaleAlphaHQ]

We still hold both $AMD and $INTC till now. But going to trim more today.

[SN @ScaleAlphaHQ]

Taking profits on $INTC, trimming to a 5% long-term hold.

Adding $AMD today, time to gradually build up the position! @grok

#Investing #Stocks #FinTwitt

[SN @ScaleAlphaHQ]

Continuing to quietly hold my positions without saying much. Today, all correlated assets rose together again; the most recent time this happened was over the three days from April 21st to 23rd.

Risk-off side: Oil is up, US Treasury yields are up, the USD is up, the VIX is up, BTC is down, and the Magnificent 7 saw 5 down and 2 up.

Risk-on side: All four major US stock indices are up, with semiconductors, 🚀 (rocket) stocks, quantum computing stocks, stablecoin stocks, and uranium stocks all going up.

[SN @ScaleAlphaHQ]

Portfolio Positioning
Because of this, the consensus trade is likely holding AI in one hand and energy in the other. This is exactly how our live portfolio is positioned, plus a bit of broad market hedging. Since the beginning of April, SPY's trading volume has been on a steady decline, whereas XLE has maintained pulsing, high-volume spikes and began rebounding on April 17th.

[SN @ScaleAlphaHQ]

AI Supply Chain vs Energy Gap
$NVDA $AMD $INTC $VST $CRWV $NBIS $XLE $USO

The core focus of the market right now is the contradiction between the certainty of the AI narrative and the uncertainty regarding actual shipping and navigation through the Strait. As a result, no one is willing to let go of their AI hardware positions, while simultaneously worrying that if the blockade continues for another dozen days or so, the first refinery to announce depleted inventory and halt production will trigger an across-the-board price surge in refined oil and petrochemical products.

[SN @ScaleAlphaHQ]

Energy and Geopolitics
The uncertainty of navigation through the Strait has a clear lagging impact, as we have discussed before. When everyone still has inventory, things remain calm; but once that inventory runs out, the shock will be precipitous. It will push the prices of various oil and petrochemical products to a high structural baseline, rapidly elevating stagflation expectations.