Cerberus

I made this for retards. $NVDA

@Pirat_Nation False.

RAM memory companies plummeted and are reporting million-dollar losses after Google's announcement of the creation of an AI algorithm that reduces memory usage by a factor of 6 and improves speed by a factor of 8. RAM prices are projected to go down.

@TrendSpider Algos then a manual fade since humans know it's BS

🚨 BREAKING: $SPY moves +6 points in under a minute as Trump pauses strikes on Iranian power plants for 10 days.

Gadi Hutt leaves Annapurna Labs Amazon burned years and billions trying to make Trainium happen investing in OpenAI, Anthropic, anyone who’d use it Now they’re loading up on Blackwell Turns out you can’t will your way past Jensen ANON $NVDA $AMZN

BREAKING: The S&P 500 closes sharply lower as Iran War pace talks lose momentum, erasing -$1 trillion in market cap today. The S&P 500 as now lost -$3.5 TRILLION since the Iran War began.

@FinanceLancelot Bookmarking for the roast I will do if you later. Don't delete this pls.

Micron Technology $MU is collapsing right on schedule. This will make 2000 look like a cakewalk. The catalyst? Finding out Scam Altman who "secured" 40% of the global ram supply until 2029, is broke.

@jbulltard1 It's obvious that he's in panic mode. He's losing. He's stuck.

This is such a joke truly he’s terrified of every market dip so he changes deadlines after the close twice this week

@jbulltard1 That's awesome. That's the day I can sell my 3rd biggest position without triggering a wash sale 😂

Wickedly bad day

@Mylovanov I think I see a flaw in that plan. Hey guys, let's go hide behind these gigantic vats of oil

A potential scenario: US Marines land on Kharg Island under fire, take cover behind Iran’s oil terminals. This forces Tehran into a brutal choice — destroy its own economy or let the US control 90% of its oil exports, — Financial Times. 1/

@KobeissiLetter Idk how he even comes up with these lies

BREAKING: Bloomberg says tanker tracking data since March 23rd "shows no sign" of the 8 boats full of oil that President Trump said were "gifted" to the US. "If this was some great 'present' you’d think both sides would want it known for their own reasons," Bloomberg says.

$SPX $SPY just 3% away from a correction. -6% in March since the war began on Feb 28th -- flat ytd at that time. Trump was "expecting worse". But he may just get it.

Joshua Meyers from $JPM pulling up some datapoints on $MU / $SNDK today. $GOOGL has been using it from 3.1 at least. It does not change anything for HBM.

@TBU12345678 Maybe they should take some of it and cram it down your throat.

given the FCF generation that $MU / Hynix / Samsung / $SNDK will see next year shouldn’t they invest in the hyperscalers to increase cash gen and keep the party going? look at the $NVDA playbook!

@amitisinvesting Market be like "hold my beer."

well Mr. President, you are now getting a lower stock market and higher oil prices today 😂

@aleabitoreddit Yeah, I was literally thinking... Mr Market: "hold my beer".

Trump in Cabinet Meeting: “I thought Oil Prices would go up more. And I thought the stock market would go down more”. Now that the $SPY index is down -5.25%… Having the President say this earlier today isn’t exactly the best words of encouragement.

@firstadopter The results show why everything is blood red today.

@insane_analyst Is that today's intraday stock price?

@jukan05 But now you can code your own

Fuck, I really needed this, but Sama killed it.

@StockSavvyShay This is what NVIDIA's $2B is for. So they get to build a free fab, with a built in customer. Pretty sweet deal.

$LITE is building a new U.S. laser fab in North Carolina to produce InP-based optical components for AI data centers with $NVDA as a key customer. The project includes hundreds of millions in investment and a production ramp targeted for mid-2028.

Morgan Stanley on "TurboQuant – Implications for Technology" Analyst comments: "This compression algorithm makes AI inference 8x faster while using 6x less memory. It affects only the KV cache during inference and gives much more output per GPU. The read-through is positive for hyperscalers and LLMs given the ROI opportunity. It is a long-term positive for computing and memory. Implications for memory: neutral to positive long term Short-term impact: TurboQuant targets only the Key Value (KV) cache during inference, which is the temporary key/value vector that grows with context length. Model weights, including HBM usage on GPU/TPU, and training workloads are not affected. It allows 4-8x longer context on the same hardware, or much larger batch sizes, without running out of memory. This is not a 6x reduction in memory or total hardware needed, but an efficiency gain that increases throughput per GPU. Long-term impact: The Jevons paradox effect is that efficiency increases total demand. The inference economics are shifting: by shrinking data size and movement, TurboQuant aims to improve throughput per accelerator and lower cost per query. The biggest bottleneck in scaling AI services today is KV cache memory. If models can run with materially lower memory requirements without losing performance, the cost of serving each query drops meaningfully, resulting in more profitable AI deployment. Models that need cloud clusters can fit on local hardware, effectively lowering the barrier to deploying AI at scale. More applications become viable, more models remain active, and utilization of existing infrastructure improves. In that sense, TurboQuant is less about incremental optimization and more about shifting the cost curve of AI deployment. Broader tech implications: another DeepSeek moment Positive for hyperscalers and model platforms: We cite the ROI opportunity from much cheaper per-unit quality in long-context inference and retrieval-heavy applications. Neutral near-term implications for computing and memory: Better compression means lower memory traffic and lower GPU-hours requirement per workload. However, a lower cost per token can also lead to higher product adoption demand, including larger batch sizes and longer context. This may be negative at the margin for the software layer, as compression can be embedded directly into platform infrastructure."