shes not special. its cheaper for these companies to send out products than pay for ads. happening alot with korean skincare, and other asian mass produced brands.

---

mstr will cross 1k before you know it and gay bers will self mass deport

---

TLT down 0.69% today….long term yields going up. I’m curious to see what the treasury market does….if short term rates dump, and inflation expectations go up, that means long term rates go up. If deflationary death spiral then long term rates go down. If we money print our way to victory the feds will force long term yields to go down by mass purchasing LTT and force Americans to pay via inflation; their chosen poison pill

---

Retarded take. There will be mass layoffs, the automation capabilities are immense.

---

Well all coronal mass ejections are usually hot in my research.

---

Frank Herbert: The Arrakeen Calculus: A Due Diligence Triptych on the Guild-Sphere of Google (Alphabet Inc.) FOR THE EYES OF THE C.H.O.A.M. DIRECTORATE ONLY I. Executive Summation: The Spice of the Internet The target entity, known across the galactic market as Alphabet Inc., yet branded to the masses as Google (or The Guild-Sphere in internal parlance), presents itself as a colossal, necessary entanglement. Its domain is the Data-Spice, the essential catalyst of modern economic foresight and communication—a flow both addictive and vital. Our Scrutiny-Mentats report a structure of two primary Houses: Google Services (the Imperium's bulk, rich in advertising revenue and the mass-mind’s subservience) and Google Cloud (the nascent, volatile territory of future-computational infrastructure). A third faction, Other Bets (the Tleilaxu of failed promise and wild-card potential), represents the necessary waste of an empire testing the bounds of the possible.

---

$ALT surging for the 5th day in a row.  Indiscriminately shorted abd they increased the position against Weight loss that retains muscle mass, tolerable, and may be a functional cure for liver disease.  Attacks fat. Huge catalyst upcoming. This baby will fly  22.3$ shorted now 

---

Mass unemployment is bullish. It worked in the Roman times 🤡🤡🤡

---

Imagine trying to pick someone up in a bar and when they ask what you do, you have to say " I bet against America and nothing excites me more than mass layoffs and your parents retirement accounts collapsing"

---

WTF are you talking about? Once you have a model and want to mass-deploy it to production (i.e. inference only), you anyway run it through some kind of a development process with custom-compiled software on the output (serving the customer requests). I'm sure any bigger company can afford to hire devs and have them use whatever non-CUDA framework these TPUs work with. Especially when TPU TCO savings far outweigh the devs' salaries. This is very different from prototyping and training, which you may want to do around CUDA and existing libraries built on top of CUDA (e.g. pytorch based frameworks and libraries), to maximise flexibility.

---