crowdturtle 🐢 @crowdturtle

Data scientist in big tech. Investor in cash-flowing real estate and special situation public securities. NYC Joined September 2011

Tweets

29K
Followers

5K
Following

2K
Likes

87K

crowdturtle 🐢 @crowdturtle

7 hours ago

@JayShams whatever as long as they are not mugging me / trying to attack me

0 0 0 214 0

View Details

tradfi news @tradfi

8 hours ago

*META CEO MARK ZUCKERBERG SAYS IN INTERNAL TOWN HALL THAT AI AGENT DEVELOPMENT OVER THE LAST FOUR MONTHS HASN'T 'ACCELERATED IN THE WAY WE EXPECTED' $META

10 5 96 30K 8

0 0 2 499 0

View Details

Normally, I wouldn’t pay much attention to statements like this from Sam Altman. But given all the extraordinary developments we’re witnessing right now, I truly believe we’re at the forefront of an unprecedented revolution. Sam Altman: "In another year or two, we expect to have built systems with astonishing power, capable of delivering tremendous value to the world. Artificial intelligence will reshape the material conditions of human life on a scale that no technology has accomplished since the harnessing of electricity, and perhaps beyond even that"

Chris @ChrissGPT

23 hours ago

Sam Altman in the financial times: “In another year or two, we expect to have built systems with astonishing power, capable of delivering tremendous value to the world. Artificial intelligence will reshape the material conditions of human life on a scale that no technology has

70 124 930 226K 275

73 86 1K 127K 377

View Details

crowdturtle 🐢 @crowdturtle

a day ago

oh we have too much memory now?!

Alex Corrino @AlexCorrino

a day ago

Added a "Memory Breakthrough" scenario to the calculator over at Memory Analyst, based on Andrew's prediction. A 60% blanket memory efficiency breakthrough that hits the market in 2027, and a modest 30% resulting increase in token demand (from lower prices) puts the memory

10 8 132 26K 75

0 0 1 618 0

View Details

crowdturtle 🐢 @crowdturtle

a day ago

so $RDDT is flying today. did a licensing deal leak????

2 0 9 4K 0

View Details

Greg @GS_CapSF

2 days ago

I think people have been far too bearish on the improvements we are seeing in cost to serve. Cutting inference costs by half would completely change the economics of a data center. I wrote about algorithmic improvements as one of the key levers to improving data center ROIC back in May. open.substack.com/pub/marginofsa…

Stephanie Palazzolo ✈️ ICML @steph_palazzolo

3 days ago

OpenAI engineers earlier this month developed an optimization that cut inference costs in half for models it was applied to. After the optimization was applied to logged-out ChatGPT traffic, it reduced the number of GPUs needed to power that traffic to a couple hundred.

36 94 1K 479K 348

6 2 29 8K 30

View Details

crowdturtle 🐢 @crowdturtle

2 days ago

there are so many variables

amit @amitisinvesting

2 days ago

$META Two more perspectives here as we digest this news, one bearish and one bullish. Bearish: If META has excess compute that they are willing to sell via a new cloud business…doesn’t that mean we aren’t compute constrained? Isn’t this really bad for neoclouds? Why would META

239 35 927 285K 174

0 0 0 516 1

View Details

crowdturtle 🐢 @crowdturtle

2 days ago

@ACapitalLP trying to build a thesis on what benefits when we go from a million to a trillion AI agents always on...the agentic rails.

0 0 0 289 0

View Details

Fishtown Capital @FishtownCap

3 days ago

@rebound_capital 1, we have no idea what the incremental returns on additional investment will be in 29-30. Fiber buildout looked great until it didn’t, too. 2, very few people will bring workloads back on prem for a multitude of reasons. Most companies can be barely handle desktop support.

2 1 6 458 1

View Details

Brian Armstrong @brian_armstrong

3 days ago

True. A lot of learning to ship is getting over the fear of what others will think. And choosing the right tools + narrowing the scope to the essentials, so you can do it in a reasonable time with high quality.

Guillermo Rauch @rauchg

8 months ago

Learn to ship. Shipping is a skill distinct from coding. Shipping is designing, coding, QAing, story-telling, teaching, marketing, selling, pivoting, iterating… It used to be that coding dominated in importance because of coding ability scarcity. AI will push you to go further.

239 964 8K 1.0M 3K

96 115 1K 193K 536

View Details

David Pan @davep

4 days ago

Everyone is writing about agent loops right now. Including us at Cursor, because they're so powerful. But here's a prediction: a year from now, nobody will be talking about them. Not because they weren't useful. Because they'll work right out of the box. Batteries included. No instructions necessary. Feels a lot like prompt engineering two years ago. It was incredibly important. People wrote courses on it. Now you just talk to your agent like a normal person. That's the strange thing about AI right now. You learn something critical, get huge gains, and before long it's the new normal and something else is the bottleneck. So the alpha isn't what you know. It's how fast you learn it, and how easily you can let it go.

251 304 4K 262K 1K

View Details

crowdturtle 🐢 @crowdturtle

3 days ago

someone please convince me to buy more $NRP. i feel like it's bleak

1 0 5 892 2

View Details

Ihtesham Ali @ihtesham2005

3 days ago

Jeremy Grantham has called every major bubble of the last 50 years and went on CNBC to say AI investors will lose their shirts before the technology ever pays off. The man with the track record is not betting on the believers. Everyone arguing about AI right now is arguing about whether it is real. Grantham thinks that is the wrong fight. He agrees it is real. That is the problem. He put AI next to the two biggest inventions of the modern era. 1. Railroads. 2. The internet. Both changed the world. Both made everyone certain. Both pulled in more money than anyone could absorb. His words: every great new invention arrives with the promise that it will rewrite the rules. Then the same thing happens every time. The certainty pulls in too much capital. The capital builds too much too fast. The technology survives. The investors do not. Railroads rewired the world and ruined the people who funded them. The internet did the same. Fortunes were lost laying fiber that nobody would use for a decade. The Amazon that came out of 2000 first fell 92 percent. The people who believed earliest took the deepest losses on the way to being right. Grantham is not telling you AI will fail. He is telling you it will work, and that the working is exactly what bankrupts the people who got there first.

9 8 37 7K 17

View Details

crowdturtle 🐢 @crowdturtle

3 days ago

the biggest welfare program in the history of mankind is $SNAP's stock comp to its execs

1 0 6 729 0

View Details

crowdturtle 🐢 @crowdturtle

3 days ago

$EVC unbelievable

Lee Roach @leevalueroach

3 days ago

$EVC hit nice $13 handle. Need this at $69 handle or my wife my leave me.

14 1 53 16K 4

1 0 6 1K 3

View Details

Dissecting the Markets @dissectmarkets

3 days ago

I love this toll road analogy with the hyperscalers and token costs 🙌 This analogy is why I was excited for $MSFT when they recently allowed users to use DeepSeek on Copilot Or $AMZN or $GOOGL when they’re hosting lower cost AI models onto their servers

Rihard Jarc @RihardJarc

4 days ago

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…

86 262 2K 1.5M 4K

2 2 8 2K 1

View Details

Chubby♨️ @kimmonismus

4 days ago

Meta is now facing the exact problem every AI company will soon face. It wants to replace expensive external coding tools like Claude Code and Codex with its own internal system, MetaCode. But to build a better coding model, Meta has to make sure it is not accidentally training or evaluating on outputs from rival models. That is the distillation trap: The more companies rely on frontier models to build internal AI infrastructure, the harder it becomes to prove where the intelligence actually came from.

105 76 1K 150K 307

View Details

Barchart @Barchart

4 days ago

Everyone cites EPS momentum as a reason to remain bullish, but we've been here before 🤯 👀

29 55 360 55K 72

View Details

crowdturtle 🐢 @crowdturtle

4 days ago

i'm starting to get interested in the mag-7. which means they $META $AMZN etc prob sell off through EOY as i am 6-8 months early usually.

Mike Zaccardi, CFA, CMT 🍖 @MikeZaccardi

5 days ago

Mag 7 P/E premium at the lowest level in more than a decade... Tortsten at Apollo And my full @StockCharts $MAGS writeup: articles.stockcharts.com/article/tech-f…

1 11 48 16K 19

0 0 4 768 2

View Details

Neil Sethi NEW account, please refollow @neilsethinew

5 days ago

Nomura's McElligott continues to warn of the negative supply from hyperscalers (at this point all GOOGL with no new buybacks authorized): For nearly two decades, “corporate buyback flows paired with de minimis issuance acted both as a latent bid under the market and a general equities-wide vol suppressor, as buybacks tend to get most active into drawdowns,” McElligott said, describing the “generational regime shift.” “In this evolved ‘AI Trade 2.0,’ that virtuous tailwind risk[s] becoming a vicious headwind on the ‘supply > demand’ inflection shock for US equities,” he went on, reminding investors that in the past, Mag7 buybacks were “20-30% of the total.”