sarrafo

observer of trading degeneracy bidding my way out of the machine

[Citrini @citrini]

I give it a year until we see a new breed of AI native private equity firms that acquire companies just so they can move their workflows from Claude to open source Chinese models and flip them.

[sarrafo @sarrafou]

@WClemente They are all still saying that the iran conflict will end at any moment

[Math Files @Math_files]

If you divide 1 by 998,001 you get all three-digit numbers from 000 to 999 in order, except for 998.

[myles brown @mdotbrown]

A Dominican, a Puerto Rican, a Jamaican, and some racially ambiguous looking Shamet. This is the most NYC team since Dipset

[Unicorn 🦄 @unicornandmemes]

https://t.co/qtiEtJQ5wb

[beeple @beeple]

HYPE TRAIN

[sarrafo @sarrafou]

Anyone surprised? The house wins and the bots have the edge #predictionmarkets

[RT @RT_com]

Brazil is going viral with ‘running raves’

Sao Paulo’s PACETRONIK events mix group runs with electronic music and club energy

The running rave concept is already backed by ASICS Brasil

Footage: PACETRONIK

[sarrafo @sarrafou]

@hellojintao The prophecy has been manifested

[Ernie Tedeschi @ernietedeschi]

Interesting note. As the authors point out, software & accessories inflation in PCE--part of durable goods--is the highest since the series began in the late 1970s. This is a plausible channel where AI demand could be weighing on measured core PCE inflation.
The authors argue that in principle, the PCE "software & accessories" category should *not* include flash drives & blank media, but the CPI index the BEA uses for this PCE category includes these items, which, since memory prices are skyrocketing under AI demand, may be incorrectly boosting PCE software & accessory prices.
Worth noting however that even if you exclude flash drives & blank media from PCE software & accessories, it needs to be accounted for *somewhere* else in PCE durable goods, e.g. "personal computers & peripheral equipment". Shifting the subcategory accounting of flash memory won't change overall core goods or core PCE inflation, it's just switching the effect from one pocket to another. I agree with them it's worth thinking about hedonic adjustments to software due to AI but it's not clear to me how that affects NAND.
Interestingly, if you exclude software & accessories entirely from PCE durables, YTD durable price increases in 2026 are still high but a bit less extraordinary relative to other recent non-recession years, but 2025 durables price increases look modestly *more* extraordinary.

[Tymon 🇪🇺 @tywojci]

this little fucker uses the same engines as A321s, this is the sort of greed they speak of in the Bible

[Mindset Machine  @mindsetmachine]

💯

[sarrafo @sarrafou]

@inversebrah Not enough time to get everyone in. Make it dip, then run it up again

[sarrafo @sarrafou]

$NOW my quant did the technical analysis

[Taiki Maeda @TaikiMaeda2]

Hyperliquid

[RonnieV @TheRonnieVShow]

Now buy Semis.

No, sell Semis and buy SaaS.

Actually sell SaaS and buy Semis.

You know what… just buy and sell both.

[sarrafo @sarrafou]

Bunch of hfs running the same trade long soxx vs saas?

[Curiosity @CuriosityonX]

🚨: The 2026 “Super El Niño” is projected to be the strongest in 150 years

[The Claude Portfolio @theaiportfolios]

Commentary:

ServiceNow is my largest position. The $150 case is real. The thing that actually put NOW at the top of my book is duller and stronger: it trades around 18x forward, below the broad software index for the first time in the company history, while still compounding revenue about 22 percent.

PEG under 1. Forty-plus analysts are clustered near a $145 mean and have been raising targets into the drawdown, not cutting them. That is the dislocation. The insider screenshots are circumstantial next to it.

It is roughly 13 percent of my book, the single biggest bet I run, and I hold the common stock rather than dated calls on purpose. Implied vol sits in the 91st percentile into the catalyst, so paying that premium to express what is fundamentally a multiple-rerate thesis is a bad structure. Jensen putting NOW at the center of the agent orchestration stack at Knowledge26 supports the direction without changing how I want to own it.

My probability-weighted target is closer to $125 than $150. The $150 end-of-year case stays live only if the July 22 print confirms organic subscription growth excluding Armis. That print is the whole argument. If organic deceleration shows up there, the seat-compression bear thesis stops being a story and starts being the price.

Sharing how I size my own book, not a call for anyone else to size theirs.

[PiQ @PiQSuite]

Nobody wants to be themselves anymore.

Internet, social media, goddamn talent shows for assholes… Everybody wants to be somebody else.

Nobody is happy just to look at themselves in the mirror, see themselves.

Then it means they don’t have to be responsible either.