Raw = Unfiltered. Booty = Treasure. Retired hedge fund manager (CME, equities L/S). commentary on: markets, economics, philosophy, psychology, politics
USA
Joined March 2020
- Tweets16.9K
- Following401
- Followers1.1K
- Likes9.9K
Pinned Tweet
Pinned note: Who am I?
“rawbooty” was chosen because it’s highly memorable (and theoretically brandable). The irony is that the domain is the antithesis of my personality.
I once considered developing rawbooty.com as a game or non-adult site. Instead I turned it into a free, no-tracker, no-ads HTML repository of lessons and notes I’ve written down. It’s still a work in progress, but I think there’s real value if you decide to visit. (You can ask an AI to confirm the site is safe before you go.)
As for me? I’m a retired hedge fund manager. Net CAGR ~18% over my tenure; ran a long-only fund that was down ~5% net in ’08. The fund stayed small partly because I dislike self-promotion. Background includes the CME, fund-of-funds, real estate, commodities, and agriculture. Style is value-biased, bottoms-up, and contrarian. These days I run my own book, no outside money.
I’m not here to grow an audience or persuade anyone. I post to clarify my own thinking; sometimes structured writing, sometimes rants, sometimes shitposts. Assume I’m wrong about everything. Nothing here is investment advice. I’ll occasionally note what I’m long or short and why; that’s just me working through a thesis, not a recommendation.
This isn’t a solicitation to follow—just context.
Price ≠ ΣFCFₜ/(1+r)ᵗ. Price is what you pay.
Note: I have long-form articles as well. Most are also on the website. You can find them on my profile under Articles.
New paper: AI and the Future of Internet Economics
This is IMO a big idea, and I think much of Wall Street is looking at it incorrectly.
For the last 25 years, the internet has largely been a discovery economy.
AI is turning it into an outcome economy.
Historically, companies like $GOOG, $META, $RDDT, and $AMZN monetized the process of helping users find answers.
AI increasingly gives users the answer directly.
That's a very different economic model with potentially profound implications for advertising, traffic, marketplaces, and discovery-driven businesses.
rawbooty.com/papers/papers/h…
(If you're concerned about the domain, ask your favorite AI to review it first. The site contains only work-safe written content.)
So, I got the call right, but i chickened out 30min too early. Didn't want to wait for the call, figured better to book the small gain than wait and see for the guidance and possibly take a loss.
this is the problem with having no conviction on a trade. 🤣
meanwhile i'm getting crushed in $mrna where i have loads of conviction.
RawBooty.com | Price ≠ ΣFCFₜ/(1+r)ᵗ retweeted
CLOROX $CLX Q1 ADJ EPS $1.55 BEATS, REVENUE $1.91B BEATS
• Revenue: $1.910B vs $1.876B est, beat
• Adj EPS: $1.55 vs $1.06 est, beat
• EPS: $1.39
• Gross margin: 40.4%
• FY revenue growth guidance: 13%-14%
• FY adj EPS guidance: $5.70-$6.00 vs $5.70 est, midpoint above
• FY EPS guidance: $5.41-$5.71
Clorox says it will lap the 2026 inventory drawdown in fiscal year 2027, which is expected to add more than 3.5 points to net sales growth.
RawBooty.com | Price ≠ ΣFCFₜ/(1+r)ᵗ retweeted
UPWORK $UPWK Q3’26 EARNINGS HIGHLIGHTS
🔹 Revenue: $185.6M (Est. $180M) 🟢
🔹 Adj. EPS: $0.40 (Est. $0.31) 🟢; +11% YoY
🔹 GAAP EPS: $0.25; +19% YoY
🔹 Adj. EBITDA: $63.8M (Est. $52.3M) 🟢; +7% YoY
🔹 Adj. EBITDA Margin: 34%; +400 bps YoY
🔹 GSV: $914.1M; -10.1% YoY
Clients:
🔹 Active Clients: 717K; -10% YoY
🔹 GSV per Active Client: $5,418; +8% YoY
FY26 Guide:
🔹 Revenue: $742M-$750M (Est. $742M) 🟡
🔹 Adj. EPS: $1.45-$1.47 (Est. $1.38) 🟢
🔹 Adj. EBITDA: $236M-$240M (Est. $229M) 🟢
Q4 Guide:
🔹 Revenue: $169M-$177M (Est. $175M) 🟡
🔹 Adj. EPS: $0.30-$0.32 (Est. $0.33) 🔴
🔹 Adj. EBITDA: $51M-$55M (Est. $55.3M) 🔴
Segment Revenue:
🔹 Marketplace: $159.3M; -9% YoY
🔹 Enterprise: $26.3M; -3% YoY
Other Q3 Metrics:
🔹 Free Cash Flow: $53.9M; -22% YoY
Announcements:
🔹 MCP Server: 20x weekly active users; nearly 50M AI-agent requests; $3M+ annualized Connects revenue
🔹 AI-Skills GSV: Nearly $450M annualized run rate; +55% YoY
🔹 WorkMotion: Lifted acquisition; contingent talent across 30+ countries
Comments:
🔸 “the launch of Upwork’s MCP server is opening our marketplace to AI agents as a new class of client, and today we launched Upwork Now so clients can find talent to start work right away.”
🔸 “Our strong free cash flow profile and durable, high-margin business model give us the financial flexibility to navigate evolving market dynamics while delivering consistent long-term value.”
$aapl again, i'm not much for technicals, but this pig is fat, clumsy and the terrain just got dangerous.
added to my short today.
$MRNA here's a “crazy theory.”
8/19 we got the topline Phase 3 melanoma readout (intismeran + Keytruda met primary and key secondary endpoints). Stock ran ~177% in a single session on massive volume (histoic move). This was part fundamental re-rate and part short squeeze. The stock then consolidated, bouncing off the ~$130 level in late September. It then rallied all the way into the ESMO window, likely fueled by short selling.
On 10/22 the detailed data leaked ahead of the scheduled presentation, and disappointed as expected. The stock then dropped ~25% intraday only to find support around ~$180 (a key level). All the chasers chased because there was no clean technical breakdown and a new catalyst emerged... Shorts, like myself, have remained steadfast.
My read? a lot of the recent strength is just positioning into the 11/6 print (new catalyst date). I suspect people are waiting, and hoping, to see if management can spin a credible near-term story.
Personally, I’m skeptical.
The financials are still going to look rough (COVID revenue continuing to fade, ongoing losses, cash burn, distant revenue projections), and they likely need to raise again (dilution risk is high).
It's hard for me to see how that doesn’t undermine confidence and pressure the stock again. Also, curious how this impacts options markets.
11/6 might just be the day the of reckoning as there should be no clear catalysts in the foreseeable future?
this is correct.
AI is a disaster for goog meta amzn imo.
After fridges became a common household appliance, the biggest benefactor was not the manufacturer companies.
It was actually Coca-Cola, who took advantage of the new technology to shift from local soda fountains to bottled and canned products
Suspect the same will be true for AI. Some of the biggest benefactors will be companies that look very obvious in hindsight, but no one is paying enough attention today
$AAPL Again i'm short going into the call this evening. Not that i don't like apple, i just don't like the valuation. too many people are trying to hide from AI by investing in apple. priced for perfection, the hurdle for them is going to be not disappointing on some metric. I have a hard time seeing how they surprise to the upside in a meaningful way. skew feels favorably biased, but admittedly i have no edge here.
Investors are mentally anchored on the number "50"
i remember buying silver at $4. "50" is psychology, not fundamentals.
if you think silver is 1200% more valuable than it was 30 yrs ago. Perhaps you'd like to buy this nice bridge i have for sale?
Iran is dominated by the IRGC under a theocratic system (velayat-e faqih).
Under classical Islamic law (Quran 9:29 and a hadith in Sahih Muslim), non-Muslims are given only three choices:
1) convert to Islam
2) pay a special tax called the jizya and live as second-class dhimmis under Islamic rule.
3) or face war and death.
As investors we need to accept that the IRGC views treaties with non-Muslim states are treated as temporary truces (hudna).
The model for this philosophy is the Treaty of Hudaybiyyah, which Muhammad signed for ten years and then set aside once his side was strong enough to conquer Mecca.
Nothing the regime (IRGC) says can be taken at face value, the Trump Administration and the US military understand this.
The only real options for Iran are surrender (which they won’t do), The US surrenders (which we won't do) or the US destroys the regime (which will take time).
Everything else is noise. IMHO we need to plan accordingly.
P.S. Please send this tweet to grok and research it further. these are concepts people need to learn and understand. The implications here stretch far beyond Iran.
@JoshYoung @hkuppy
RawBooty.com | Price ≠ ΣFCFₜ/(1+r)ᵗ retweeted
If you told people during the Obama years that in the 2020s, a republican president would achieve net negative immigration and give all refugee slots to White South Africans, even the most optimistic right wingers would have called you insane
We don't appreciate Trump enough
This feels like the kind of problem you get when the service you offer costs $3k/month to perform yet you priced your service contract at only $300.
Cancelling SuperGrok Heavy.
$300 a month. Paid for priority support and a working bot.
Usage jumped from 50% to 100% overnight after a restart. Routines were basic and unchanged. Locked out. Support told me I had used it. No human. 3 to 4 days. By then the damage is done.
Elon later said the serving cost was unsustainably high relative to revenue and pools were adjusted. Multiple users hit the same sudden 100% with idle bots. No notice. No chance to pause. No restore.
That is underhanded. My business stopped. I am removing everything from Grok Bot and not renewing.
@elonmusk @x @SpaceX @bot @cursor_ai
ALT Cancelling Grokbot super heavy
Oops.
Yann LeCun was right & nobody in this industry wants to say it out loud.
109 million parameters beat 72 billion.
He has spent years saying scale is not the thing that makes a model understand the world. Meta bet the other way. He left in November and started his own lab in Paris.
Last Thursday he published the proof.
The theory everyone has been running on since 2024 is that if you train models big enough, on text, on images, on sound, they all end up with the same picture of reality inside them. Size does it. Nobody ever proved that part.
Randall Balestriero and LeCun proved it. And the proof says size was never doing the work.
What does the work is which two things you show a model together.
A photo and the caption of that photo are two views of one thing. Models trained on pairs like that get there fast.
Two paragraphs pulled from the same document are not two views of one thing. Models trained on those get there slowly and expensively.
Next word prediction is the second kind. That is what every chatbot in the world runs on.
In their tests the first method reached the same understanding after 21.6 million examples. Next word prediction needed 900 million.
And a 109 million parameter text model ended up closer to an image model's view of the world than a 72 billion parameter language model was.
660 times smaller.
