@P_Remarks u did good on this one. i felt the same but your tweet gave me the confirmation bias i needed to increase my soxx short to 3m
English
CosmicKittensNFT
102 posts
CosmicKittensNFT
@_cosmic_kittens
CosmicKittens is a generative animated 3D NFT collection of 4,444 unique kittens returning from space that need a loving home 💖😸 https://t.co/F6nXULZl5F #NFTs
Katılım Ekim 2021
215 Takip Edilen46 Takipçiler
This was literally the local top.
GIF
Prepared Remarks@P_Remarks
Love the replies here. Same kind I got when shitting on PLTR at $200, HIMS at $60, GME, AMC, TSLA in 2021 We’re close
English
@P_Remarks bro travel nurses make decent money, 200-300k/year sometimes
English
Even the travel nurses are becoming millionaires. The wealth effect tailwind is underappreciated
English
@firstadopter why don't you go find a wife to pound you fuckin nerd
English
I pounded the table on the AI trade yesterday on my Substack, just as I did on March 30 at the market/chip stock low.
tae kim@firstadopter
Why the AI Trade Is Unstoppable taekim.substack.com/p/why-the-ai-t…
English
@zephyr_z9 why do you people like larping as investors? get a job
English
@jbulltard1 this is retarded and you're a retard for posting this thinking it's accurate
English
@DiogenesLaertus @FishtownCap retard. harness is overrated and you clearly don't know what you're talking about
English
@FishtownCap Perhaps You have not used it for coding tasks then. Long running complex tasks. Even in vscode or copilot desktop. It is useless compared to cursor. Claude code/desktop or codex on > gpt 5.4. Anthropic even has a better Claude for excel and ppt. Sorry msft will get eaten alive.
English
$META went $600 to $580 on news from 2 days ago then to $638 AH on AI release news.
Market approaching 2021 absurdity.
In a couple years all models are so good they’re commodities and $MSFT is the winner as the massive window deploying it to the enterprise.
English
There is overwhelming demand for HBM memory relative to supply. If there weren't, prices wouldn't be rising every few weeks amid widespread shortages.
Just because a couple of distributors leaked to the sell-side that they are double ordering DRAM (not even HBM. also consider the incentives here) doesn't change the reality of the supply demand situation.
People should have some critical thinking skills versus these manufactured narratives. Use your brain.
Jukan@jukan05
A few friends have asked me what I think about Evercore’s channel checks. Evercore: "Several OEMs admitted that due to extended lead times, they placed orders for components at both distributors and brokers for the same program." Even if the DRAM Big Three try to eliminate overbooking and double-booking, it is impossible to completely stamp it out. That said, I think it is highly irrational to extrapolate what is happening in a few isolated channels to the entire market. Instead, I would rather share Lynx’s comments from the same day: Lynx on storage re META- "According to our checks Meta is willing to procure memory and flash at spot prices from the primary vendors. Unlike Apple, we do not expect Meta to complain about high prices. We expect Meta to have planned for rising prices in its capex plans."
English
@Franny8 imagine listening to a dude with an anime profile pic for stock advice lmao
English
My rationale is basically the reason software got hit so hard was real. It became a terminal value question. When one of the old software moats, code, gets taken away or at least gets commoditized, “code is cheap" -> the market sniffs where this is going and marks down IGV hard. Maybe it marked down some of these names too much. That part is fair.
But the disruption threat is real too. The acceleration in coding and model development has honestly surpassed my own view, especially with Harness, Claude Code, Codex, and the amount of capital OpenAI and Anthropic have to take as much software business as possible. No doubt about it. So maybe this software rotation is temporary and fades. Software will never be the same. (Codex call MCP -> do my taxes -> send to IRS/IRS agent, grant we not there yet with full trust but this is not that far off at all)
On the other side, I do not think the physical AI trade is dead at all. Memory weakness commodity fear, peak cycle fear, and crowded positioning. That is different from AI demand being over. The AI trade is not done yet is my view and if you missed the entire leg up honestly, i would add some here.
For hyperscalers, I still think the market will not fully re-rate them until returns on capex are proven. The risk/reward may be to the upside from here, but I am not expecting a full re-rating until the market sees real AI monetization, margin support, and free cash flow evidence.
Franny@Franny8
Not buying this $iGV rotation nonsense
English
@P_Remarks They could be shitposting about their losing stock picks on X instead, truly a huge waste of talent
English
@P_Remarks I follow you to read your funny man shitposts about your gay stock picks, not this nerd shit
English
I also firmly believe that you have to read fiction to be well rounded. There are brilliant writers and thinkers who can make you think deeply about topics
English
My recommendation is always to read non fiction books about very specific topics or real life stories over generic finance books.
Secret Life of Groceries
Skunk Works (Lockheed advanced deployment division)
The Enterprise of Flight
The Path Between the Seas (Panama canal)
Prepared Remarks@P_Remarks
There is only one book on this list worth reading
English
@EricJhonsa @Lazzuss that's a nice list but retail has literally never heard of any of these so not worth
English
@Lazzuss At current levels: UCTT, MKSI, ONTO, and to some extent AMAT and ACMR
English
As annoying as it is to see AI infra sell off so indiscriminately, it's probably healthy for valuation-indifferent momo-chasers to be reminded the stove is hot and for some other parts of tech to catch a bid.
(Also, at least some semicaps are now back to reasonable multiples.)
English
@RihardJarc People are totally missing the security risks of using open weight Chinese models. Open weight doesn't mean they won't insert subtle trapdoors into your code or steal your IP.
English
I just published my article on how hyperscalers like $AMZN, $MSFT, and $GOOGL will benefit in the token optimization era and why I believe the companies will re-rate much higher.
uncoveralpha.com/p/why-token-op…
English
English
@brian_armstrong @grok @grok so the assumption here is that since engineers review it there is no fear of backdoors created by the models?
English
How to keep AI spend flat while token usage grows exponentially: Not with friction and spend alerts. With better defaults, routing, and caching.
Better Defaults (not Usage Caps) – Engineers can choose any model they want, but defaults matter. We’re experimenting with defaulting to open weight models like GLM 5.2 and Kimi 2.7 through our LLM gateway, while still encouraging engineers to choose the right model for the task. 91% of our employees were never hitting their usage caps, so instead of lowering caps and driving up alerts, we're moving to cheaper defaults. Note that code reviews use a diversity of models, so they can check each other's work.
Better Routing – In our custom harnesses, we preprocess prompts and route to the best model for the job, considering cache hits and model pricing. For instance, you may want a frontier model for planning, but not for execution where they can be overkill. Ultimately, humans shouldn't be choosing models - AI can automate this task.
Better Caching – Cache misses are the easiest way to drive your cost up. All of our requests are cache aware, so we’re reusing a warm cache wherever possible. For example, our cache hit rate went from 5% → 60% in LibreChat once properly implemented.
Keep Context Lean – Start fresh sessions when switching tasks. Scope file context narrowly. Disconnect unused tools. Don't just compact. The goal isn't fewer tokens used, it's fewer tokens wasted.
Better Visibility – Our engineers can use as many tokens as they want, from whatever model they want, but we’ve made usage visible – and the more you spend on AI, the more impact we expect.
The goal isn't to suppress usage. It's to build the infrastructure that makes exponential growth sustainable.
Putting this into practice has cut our AI spend nearly in half, while our token usage continues to grow.
English
@not_stefan0 @brian_armstrong are you retarded lol. llm security risks lmao
English
@brian_armstrong You’re telling me Coinbase uses glm 5.2??? I didn’t even want to use that on one of my projects due to security risks
English
@StockSavvyShay brother why do you even post this useless bullshit
English
Supply chain analyst Ming-Chi Kuo says $AAPL is reportedly raising baseline iPhone DRAM from 8GB to 9GB in 2027 to keep Apple Intelligence running smoothly.
On-device AI is pushing DRAM content per phone higher which is bullish for $MU, Samsung and SK Hynix.
English
$MELI ads business is estimated to have hit a $2B run-rate with 73% YoY growth.
Management previously said ads are a high margin business, with EBIT around 80%.
After tax, that would contribute about $1.2B toward net income. At a 40x P/E, which is conservative IMO for a high margin business growing 73% YoY, Mercado Ads alone would be worth around $50B.
MELI’s market cap is $82B, which suggests the market may be giving surprisingly little credit to ads relative to the rest of the commerce, logistics, payments, credit, and fintech businesses.
Credit to @PronkDaniel for pointing out the ad unit economics in his latest Meli video last week.
English
I spend like 50% of my weeks in F500 CIO/CFO offices:
everyone is waiting for someone to take the "Block Moment" to AI budgets
the first CIO that goes on record saying "the emperor has no clothes - we're capping LLM spend to $1-3K / user / month"
triggers an avalanche
Vivek Garipalli@VivekGaripalli
overheard from a fortune 20 company - ceo asked for $1 billion in AI generated opex savings at the beginning of this year. the team as a result has spent $200 million on tokens trying to achieve those savings year-to-date, with minimal results other than some modest Cx savings and a bit of savings on engineering due to less hiring driven by coding assistants. now as back-half budgets are being reviewed, it appears that the ceo has ordered token costs to be dramatically slashed as he/she doesn't feel the ROI is there yet (for their company). gonna be interesting to see if this is a trend amongst the rest of the fortune 500.
English