LTCM (Long-Term Capital Management)

$AVGO — #LONG

I’m opening a long position in Broadcom around $347.00.

▪️ Entry: $347.00
▪️ Stop Loss: $328.00
▪️ Take Profit: $385.00
▪️ Price Target: $405.00
▪️ Expected Return: +16.71%
▪️ Risk: −5.48%
▪️ Risk / Reward: 3.05 : 1
▪️ Leverage: 1.5×
▪️ Portfolio Allocation: 5%
▪️ Portfolio Risk: −0.41%
▪️ Time Horizon: 2–4 months

THE THESIS

I’m buying Broadcom here because I think the market is starting to focus more on the structure around the AI buildout than on the demand itself.
That distinction matters.
Broadcom’s AI semiconductor revenue grew 221% year-over-year in Q3, driven by strong demand for custom AI accelerators and networking, and the company expects that momentum to continue into Q4. Broadcom Inc.
That is the part of the story I care about.
The recent debate around AI financing, customer concentration and the scale of infrastructure commitments is real.
But I don’t think those concerns automatically invalidate the underlying demand.
Broadcom is increasingly becoming one of the companies sitting directly underneath the AI infrastructure cycle rather than simply trading around the narrative.
At roughly $347, I think the risk/reward is becoming attractive again.

WHY NOW

AVGO closed at $343.64 on October 1 after falling 2.15% and is trading around $347 overnight. The stock is also still more than 30% below its 52-week high of $495. MarketWatch
I like that setup.
I’m not buying after a vertical move.
I’m buying after the market has had time to question the story.
My first take-profit level is $385.
The broader target is $405.
I don’t need Broadcom to return to its previous highs for the trade to work.
I only need the market to become more comfortable with the idea that AI demand remains strong even if the financing structure around that demand becomes more complicated.

RISK MANAGEMENT

My invalidation level is $328.
Below that level, I think the current setup begins to lose the asymmetry that makes the entry interesting.
The biggest risk is not that Broadcom suddenly stops being relevant to AI.
The bigger risk is that the market starts assigning a much larger discount to customer concentration, financing exposure or the durability of AI infrastructure spending.
If that starts happening, I don’t want to defend the thesis indefinitely.
I would rather exit, reassess and come back later.

WHAT I’M WATCHING

▪️ Bull case: AI accelerator and networking demand remains strong, Q4 execution stays ahead of expectations and financing concerns fade.
▪️ Base case: Growth remains strong enough for the stock to recover toward $385–405 as sentiment normalizes.
▪️ Bear case: Customer concentration, financing risk or weaker AI capex expectations cause the market to compress the multiple further.

MY NOTE

Broadcom is interesting because it doesn’t need to win the AI narrative.
It needs to remain embedded in the infrastructure.
That is a much simpler thesis.
The latest Anthropic financing arrangement is a good example of both sides of the story: Broadcom is becoming more important to the AI buildout, but that importance also creates new financial and concentration risks. Reuters
I’m comfortable with that.
I don’t need the setup to be perfect.
I need the upside to justify the risk.
At this price, I think it does.

▪️ Position: LONG AVGO
▪️ $347.00 → $405.00
▪️ Expected return: +16.71%

STOCKS ARE GETTING INTERESTING AGAIN

Crypto has been the default playground for retail traders for years.

At this point, almost every “opportunity” starts to look the same. Another token. Another narrative. Another chart promising 10×. I’m much more interested in equities right now. Not because stocks are easier. They aren’t.

But because the market is giving us cleaner setups, clearer catalysts and, in many cases, much more understandable risk.

That matters. You don’t need to trade everything. You need to know what you’re looking for, where you’re entering and where the idea stops working.

That’s why I’ve been watching Hedge Capital.

They publish stock setups and trading signals regularly, with the focus on actual entries rather than endless market commentary.

The part I like is simple:

▪️Specific idea
▪️Clear direction
▪️Defined entry
▪️Defined risk
▪️Defined target

No trade is guaranteed. And a signal is useless if you don’t know how much you’re willing to lose before entering.

But if you’re looking for another source of equity ideas to compare against your own view, Hedge Capital is worth adding to the watchlist.

Crypto isn’t going anywhere. I just think equities are becoming much harder to ignore.

$NVDA — #LONG

I’m opening a long position in NVIDIA at $184.94.

▪️ Entry: $184.94
▪️ Stop Loss: $169.00
▪️ Take Profit: $220.00
▪️ Price Target: $220.00
▪️ Expected Return: +18.96%
▪️ Risk: −8.62%
▪️ Risk / Reward: 2.20 : 1
▪️ Leverage: 2×
▪️ Portfolio Allocation: 10%
▪️ Portfolio Risk: −1.72%
▪️ Time Horizon: 6–9 months

THE THESIS

I’m bullish on NVIDIA here, but this is not simply a bet that “AI will keep growing.”
Everyone already knows NVIDIA is the dominant supplier of AI accelerators.
That is not the edge.
The more important question is whether hyperscalers and enterprises are willing to keep spending at a level high enough to sustain the infrastructure cycle.
My view is that we are still relatively early.
AI workloads are becoming larger, more complex and more expensive to run. That creates a strong incentive to keep investing in faster and more efficient compute.
NVIDIA remains one of the clearest ways to express that view.
At $184.94, I think the market is giving me enough upside to justify taking defined risk.

WHY NOW

I don’t want to chase NVIDIA after a vertical momentum move.
I’d rather build exposure while the market is still debating how durable AI infrastructure spending really is.
My target is $220, representing roughly 19% upside from the entry.
I don’t need another perfect quarter for this trade to work.
I need demand to remain durable and the market to keep recognizing that NVIDIA’s competitive position is still difficult to replicate.

RISK MANAGEMENT

My invalidation level is $169.
Below that level, the setup becomes materially less attractive to me.
I’m not interested in turning a trade into a long-term investment simply because the price moves against me.
If the thesis breaks, I would rather take the controlled loss and reassess.
A strong company does not automatically mean every entry is a good one.

WHAT I’M WATCHING

▪️ Bull case: Continued hyperscaler capex, strong accelerator demand and positive forward guidance.
▪️ Base case: AI spending stays elevated and NVIDIA continues compounding revenue faster than the broader semiconductor sector.
▪️ Bear case: Hyperscalers reduce capex, competition increases faster than expected or AI infrastructure spending begins to slow.

MY NOTE

NVIDIA is one of those stocks where investors are constantly trying to identify the exact moment when the story stops working.
Eventually, that moment will come.
I just don’t think we are there yet.
Everyone knows the AI story.
Everyone knows NVIDIA is dominant.
The edge is not knowing the story.
The edge is deciding when the price gives you enough room for the story to be wrong.
At $184.94, I think there is enough room.

▪️ Position: LONG NVDA
▪️ $184.94 → $220.00
▪️ Expected return: +18.96%

$META — #SHORT

I’m looking to open a tactical short position in Meta around $750.00.

▪️ Entry: $750.00
▪️ Stop Loss: $782.00
▪️ Take Profit: $710.00
▪️ Price Target: $690.00
▪️ Expected Return: +8.00%
▪️ Risk: −4.27%
▪️ Risk / Reward: 1.88 : 1
▪️ Leverage: 1.5×
▪️ Portfolio Allocation: 5%
▪️ Portfolio Risk: −0.32%
▪️ Time Horizon: 2–6 weeks

THE THESIS

I’m taking a tactical short in Meta after one of the strongest repricings the stock has seen this year.
This is not a bearish call on Meta as a business.
The AI story has improved materially. Muse has given the market something it has been waiting for: visible evidence that Meta may be able to turn its AI spending into a consumer product with real distribution and, eventually, real monetization. Reuters
I agree with that part of the story.
What I’m less comfortable with is how quickly the market has priced it in.
META went from roughly $613 earlier this month to almost $780 at the recent high. That is an enormous change in valuation for a thesis that is still in the early stages of proving itself. Investing.com 日本
I don’t need Muse to fail for this trade to work.
I only need expectations to cool faster than the fundamentals improve.

WHY NOW

META reached $779.82 on September 24 and then closed the next session at $751.66, down 3.33%.
That matters to me.
After a vertical move, the first failed attempt to hold new highs is usually more interesting than the move itself.
I’m not trying to short the exact top.
I’m looking for a normalization toward $710 first, with a broader target at $690.
Even at $690, the stock would still be well above where it traded before the latest AI-driven repricing.
That is why I don’t need a fundamental breakdown.
I just need some of the enthusiasm to come out of the price.

RISK MANAGEMENT

My invalidation level is $782.
That sits just above the recent high.
If META breaks back through that area and holds it, the market is telling me that momentum is still stronger than the mean-reversion setup I’m trading.
I would rather take a small predefined loss than keep increasing a short against a stock making new highs.
I will not average into this position above the stop.
The position is deliberately smaller because shorting momentum is not the place to be aggressive.

MY NOTE

Meta is probably one of the better examples of why I separate a company from its stock.
I like the business.
I like the distribution.
I like the fact that Meta finally has a clearer answer to the question of how its AI infrastructure spending might become a product.
None of that means I have to like the stock after a 25–30% monthly move.
The market has gone from questioning Meta’s AI spending to aggressively pricing in its success in a matter of weeks.
That transition may ultimately be correct.
I just think the price got there faster than the evidence did.
I’m not betting against Meta.
I’m betting against the idea that every good story deserves to be bought at any price.

▪️ Position: SHORT META
▪️ $750.00 → $690.00
▪️ Expected return: +8.00%

$AMD — #SHORT

I’m opening a tactical short position in AMD around $630.00.

▪️ Entry: $630.00
▪️ Stop Loss: $655.00
▪️ Take Profit: $600.00
▪️ Price Target: $575.00
▪️ Expected Return: +8.73%
▪️ Risk: −3.97%
▪️ Risk / Reward: 2.20 : 1
▪️ Leverage: 1.5×
▪️ Portfolio Allocation: 5%
▪️ Portfolio Risk: −0.30%
▪️ Time Horizon: 2–6 weeks

THE THESIS

I’m taking a tactical short in AMD here after one of the strongest momentum runs in the market.
This is not a bearish call on AMD as a business.
The company remains one of the clearest beneficiaries of the AI infrastructure cycle, and the market is increasingly recognizing its position across CPUs, accelerators and full-system compute.
That is exactly why I think the setup is becoming interesting from the other side.
AMD has moved from roughly $500 to above $630 in less than two weeks. The stock is now trading near fresh highs after an extremely aggressive repricing.
At this point, I think expectations have moved faster than the underlying information.
I don’t need the AI thesis to break for this trade to work.
I only need the market to stop pricing every incremental piece of good news as another reason to expand the multiple.

WHY NOW

AMD closed yesterday at roughly $629, after trading as high as $630.80.
Momentum can obviously continue.
But this is the type of move where I become more interested in the asymmetry of a pullback than in chasing the final part of the rally.
My first take-profit level is $600.
The broader target is $575, which would still leave AMD well above where it traded only a few sessions ago.
I’m not calling for a collapse.
I’m looking for a normalization.

RISK MANAGEMENT

My invalidation level is $655.
If AMD continues breaking higher through that level, I don’t want to fight the momentum.
Shorting a strong company at an all-time high is one of the easiest ways to be fundamentally right and still lose money.
That’s why the position is smaller, leverage is limited and the stop is defined before the trade begins.
I will not average into this short above the invalidation level.

MY NOTE

I still like AMD.
That doesn’t mean I have to like AMD at every price.
The market has spent most of this year discovering reasons to pay more for the same story.
AI demand is real. Server demand is real. AMD’s competitive position is improving.
None of that automatically means the stock should move vertically without interruption.
The interesting part here is that I don’t need the market to change its mind about AMD.
I only need enthusiasm to cool.
Sometimes the cleanest short isn’t a broken company.
It’s a great company after the market has temporarily forgotten that price still matters.

▪️ Position: SHORT AMD
▪️ $630.00 → $575.00
▪️ Expected return: +8.73%

$GOOGL — #LONG

I’m opening a long position in Alphabet around $352.00.

▪️Entry: $352.00
▪️Stop Loss: $334.00
▪️Take Profit: $385.00
▪️Price Target: $400.00
▪️Expected Return: +13.64%
▪️Risk: −5.11%
▪️Risk / Reward: 2.67 : 1
▪️Leverage: 1.5×
▪️Portfolio Allocation: 6%
▪️Portfolio Risk: −0.46%
▪️Time Horizon: 3–6 months

THE THESIS

I’m taking a long position in Alphabet here because I think the market is still treating AI as a potential threat to Google when it is increasingly becoming another layer of distribution for the business.
The obvious risk is search disruption.
Everyone knows that.
What interests me more is whether Alphabet can use its existing distribution, infrastructure and data advantage to turn that disruption into another monetization surface.
I think it can.
Google doesn't need the economics of search to remain completely unchanged for this trade to work. It needs the market to become more comfortable with the idea that AI changes the product without destroying the underlying business.
At roughly $352, I think there is still enough room between the current price and the upper end of the recent range to justify taking the risk.

WHY NOW

Alphabet has been moving higher, but I don't think the setup requires me to chase a vertical move.
The stock closed around $355 yesterday and is indicating slightly lower in pre-market trading today, giving me a cleaner entry rather than buying directly into strength.
My first take-profit level is $385.
The broader target is $400, which still sits below the stock's recent 52-week high around $409.
I don't need a breakout to new highs immediately.
I need the current trend to remain intact while the market continues getting more comfortable with Alphabet's AI economics.

RISK MANAGEMENT

My invalidation level is $334.
Below that area, the trade begins to lose the structure that makes the current entry attractive.
I’m not interested in averaging down simply because Alphabet is a high-quality company.
Quality does not remove price risk.
If the market gives me evidence that the setup has changed, I would rather take a controlled loss and reassess from outside the position.

MY NOTE

I think Alphabet is one of the more interesting AI trades precisely because the market still argues about whether AI helps or hurts the company.
That disagreement creates the opportunity.
If everyone already agreed that Google was one of the clear winners, I would probably be paying a very different price.
I don't need every concern around search to disappear.
I need Alphabet to prove that its distribution is more durable than the disruption narrative suggests.
At this price, I think that is a bet worth taking.

▪️Position: LONG GOOGL
▪️$352.00 → $400.00
▪️Expected return: +13.64%

RESULTS — JANUARY TO SEPTEMBER

Eleven positions closed. Eleven closed in profit.

Before anyone gets excited, the honest breakdown: seven hit the full published target. Four I closed early - in profit - when the setup stopped working. Small wins instead of stop-outs. That's not a footnote. That's the system.

The numbers:

▪️11 trades — 11 profitable
6 longs / 5 shorts
▪️+11.5% total portfolio return
▪️+9.1% average move per trade
▪️Average risk per trade: ~1% of the portfolio
▪️Average holding period: 3 weeks
▪️Best trade: ROKU, +42% move, +63% on the position

I don't think every position needs to be a 30–40% winner.

A portfolio is built from different types of opportunities.

WHAT MATTERS MORE THAN THE WIN RATE

The interesting part isn't simply that every target was reached.

It's the fact that every position had a defined:

▪️Entry.
▪️Stop.
▪️Target.
▪️Position size.
▪️Time horizon.

Every entry was published here in real time - entry, stop, target, size - before the move, not after. Scroll up and check.

What made this work was not prediction. Half of these positions were uncomfortable at some point. The AMD long was cut for +2% just before the stock broke down, and the AMD short opened into that same crash paid +29%. I'd rather book a small gain than defend a thesis into a stop.

Sizing did the rest. No single trade risked more than ~1.7% of the portfolio. When a loss can only cost you 1%, you never have to be a hero.

One warning, mostly to myself: 11-for-11 is discipline plus variance being generous. It will not last forever. At some point a stop will trigger, it will cost about 1%, and nothing about the process will change.

Position: unchanged.
Process: unchanged.

The next setups will be judged by exactly the same rules.

▪️Define the risk.
▪️Take the asymmetric trade.
▪️Know when to get out.

Meme of the day: Claude went into a frenzy over a CAPTCHA — the AI got stuck trying to solve it and wrote a whopping 150 pages of reasoning about crocodiles, frogs, and gorillas.

Then it output this:
"WHAT THE F*CK IS WRONG WITH MY ANSWERS"

AGI, firm and clear.

📟CyberFuzz📟

Atomic Heart is now available on Steam with a maximum discount.

The shooter recently became available in the RU region and CIS: the base game costs just 499 rubles, and the complete edition with all DLCs — 639 rubles.

Grab it before September 14 — here.

📟CyberFuzz📟

😁 A guy built an AI to find the best moments in adult videos — the system, based on Qwen3.8 27B, automatically scans long videos.

The algorithm looks for cues like:

• Sharp changes in the actresses' voices;
• Hard editing cuts;
• Fast scene changes.

It generates 10-second highlight clips — the AI pervert then uses them to create new videos.

No wonder RAM prices went up.

😁 Famous adult actress Sweetie Fox has started... a YouTube channel with Minecraft let's plays.

She had previously posted gaming videos on the black-and-orange platform, but the gameplay there usually took unexpected turns.

This time everything is decent: Sweetie Fox uploaded a full Minecraft let's play without any surprises.

Watching with the whole family.

🤯 Instagram* has decided to change its logo after ten years.

According to platform head Adam Mosseri, the design has been made simpler and more modern.

It was too recognizable — so they fixed it.

📣 Your phone number is probably rented. Degenphone wants to make it ownable.

Most virtual numbers work the same way: you pay, use it for SMS or verifications, then lose it when the subscription ends. Nothing is really yours.

❗️ Degenphone flips this model ❗️

You mint a fresh European number once, use it on 50+ platforms, receive SMS, pass verifications for crypto exchanges, apps, services and other platforms, and keep the number as an NFT. No KYC, no documents, no monthly “please keep paying or we take it back” energy.

And now there’s a contest running on top of it 🎁

🔥 Degenphone is giving away 6 NFT numbers:

— 1 Gold
— 2 Silver
— 3 Common

The mechanics are simple: every roll gives you points, and each next roll gives more than the previous one.

1st roll = 10 points
2nd roll = 20 points
3rd roll = 30 points
4th roll = 40 points

…and it keeps stacking.

💵 The more you roll, the heavier your entry becomes. And if you mint a number, your total points get multiplied by x2

Winners are picked randomly, but the draw is weighted by points. So yes, luck matters — but farming the contest properly matters too.

🗓 Contest ends June 20.

eSIM is already going mainstream. The interesting part here is that Degenphone turns a virtual number from a rented tool into something you can actually own, use, trade, or sell later.

Early utility + NFT ownership + live giveaway.

✅✅✅ 👉 Start rolling

All the features of the iPhone 18 Pro have leaked online — the smartphone will feature 5G satellite internet.

Insider sources shared that the device will also feature a smaller Dynamic Island notch, a 2nm A20 Pro chip, and a coffee-colored finish.

📟CyberFuzz📟

Sony will delay the PS6 release to 2029 due to a RAM shortage.

The console was originally expected to release in 2026-2028, but the release date may now be pushed back several years, and the price may increase, as the company will be unable to make the new console both powerful and affordable enough.

📟CyberFuzz📟

Cutest thing of the day: Rockstar Games has stepped up to help a terminally ill fan and given him the chance to play GTA 6

The fan reportedly had 6 to 12 months to live. After reaching out to the community, the studio quickly contacted him and provided access to the game, allowing him to fulfill his long-held dream.

📟CyberFuzz📟

A 140-year-old video has been discovered on YouTube — it has no sound, no image, and no title.

The video description is simply a string of symbols that can be deciphered to reveal a chilling message in Arabic: "Come and meet me in hell."

The video unexpectedly began appearing in user recommendations and garnered over 1 million views in a week.

📟CyberFuzz📟