AI, agents and emerging tech sharing what I learn, test and discover member @CELE8EX
Kazakhstan
Joined February 2024
- Tweets1.1K
- Following160
- Followers908
- Likes2.1K
Pinned Tweet
This is f*cking insane
Nvidia CEO Jensen Huang just explained why Wall Street keeps getting Nvidia wrong:
"Sight unseen, who can you place a 50, 100 billion dollar order on? We're literally the only one"
In 4 minutes, Jensen and Brad Gerstner walk through exactly why the consensus was off by 100-200% and the stock still trades like nobody believes it
0:16 The consensus forecast at the start of 2025
0:44 Off by a couple hundred percent in one year
1:13 Revenue forecast doubled, stock only up 20%
1:24 13 to 14 times earnings
1:57 Is the moat increasing or decreasing?
2:08 The $100 billion check thought experiment
3:54 Why Nvidia gives 100% of its business to TSMC
When a stock lags its numbers, most people blame a bubble
Jensen is teaching the trust stack:
proven last generation → confidence sight unseen → $100B purchase orders → a moat that grows with scale
Old way: benchmark the chip before you buy
New way: "you can't even benchmark these things anymore"
Better than any $2,000 finance course I've seen
This is f*cking insane
Nvidia CEO Jensen Huang just revealed two new money machines that have nothing to do with selling chips:
"I just described two new growth vectors independent of the hardware sale of our company"
In 5 minutes, he walks through exactly how Nvidia plans to get paid on every GPU for years after it ships
0:18 - Two new business models
0:27 - Software licensing for the enterprise
1:31 - A couple of thousand dollars per GPU per year
1:37 - 20 million GPUs shipped a year
2:21 - Moving cloud partners to rev share
3:17 - Why Buffett avoids tech companies, and why this is different
4:09 - "Whoever Jensen invests in you got to invest in that company"
4:54 - Why OpenAI and Anthropic won't come around again
When people value Nvidia, most of them only count hardware sales
Jensen is teaching the installed base play:
ship the GPU → license the software → share the revenue → return the cash
Old way: sell the box, wait for the next upgrade cycle
New way: get paid on the box for 4, 5, 10 years
This 5-minute clip beats most $500 MBA lectures
Dmitro Cashback retweeted
This is f*cking insane
Nvidia CEO Jensen Huang just explained why Wall Street keeps getting Nvidia wrong:
"Sight unseen, who can you place a 50, 100 billion dollar order on? We're literally the only one"
In 4 minutes, Jensen and Brad Gerstner walk through exactly why the consensus was off by 100-200% and the stock still trades like nobody believes it
0:16 The consensus forecast at the start of 2025
0:44 Off by a couple hundred percent in one year
1:13 Revenue forecast doubled, stock only up 20%
1:24 13 to 14 times earnings
1:57 Is the moat increasing or decreasing?
2:08 The $100 billion check thought experiment
3:54 Why Nvidia gives 100% of its business to TSMC
When a stock lags its numbers, most people blame a bubble
Jensen is teaching the trust stack:
proven last generation → confidence sight unseen → $100B purchase orders → a moat that grows with scale
Old way: benchmark the chip before you buy
New way: "you can't even benchmark these things anymore"
Better than any $2,000 finance course I've seen
Dmitro Cashback retweeted
This is f*cking insane
Nvidia CEO Jensen Huang just did the AI bubble math live:
"They don't have to be anywhere near a trillion dollars in revenues to be successful"
In 5 minutes, he walks through exactly why the "OpenAI can't pay for it" fear misses most of the build-out
0:00 The bubble question everyone is asking
0:40 Break it down to first principles
1:22 $3-4 trillion a year in infrastructure by the end of the decade
1:59 Moore's law is over, twice the compute now costs twice the money
3:17 $5 billion per gigawatt then, $50 billion now
4:43 Why two labs don't carry the whole trade
When AI stocks drop, most people blame the labs' revenue
The split they land on:
general compute replacement → neoclouds, sovereigns and startups → frontier labs → a few trillion of infrastructure
Old way: one lab misses a number, sell everything
New way: ask which part of the build-out actually depends on it
Better than any $2,000 finance course I've seen
This is f*cking insane
Nvidia CEO Jensen Huang just did the AI bubble math live:
"They don't have to be anywhere near a trillion dollars in revenues to be successful"
In 5 minutes, he walks through exactly why the "OpenAI can't pay for it" fear misses most of the build-out
0:00 The bubble question everyone is asking
0:40 Break it down to first principles
1:22 $3-4 trillion a year in infrastructure by the end of the decade
1:59 Moore's law is over, twice the compute now costs twice the money
3:17 $5 billion per gigawatt then, $50 billion now
4:43 Why two labs don't carry the whole trade
When AI stocks drop, most people blame the labs' revenue
The split they land on:
general compute replacement → neoclouds, sovereigns and startups → frontier labs → a few trillion of infrastructure
Old way: one lab misses a number, sell everything
New way: ask which part of the build-out actually depends on it
Better than any $2,000 finance course I've seen
Dmitro Cashback retweeted
This is f*cking insane. Ray Dalio on the US debt crisis:
"Is a debt crisis imminent in the US?" "I think within the next three years"
In 5 minutes, the founder of Bridgewater explains exactly how countries go broke:
0:24 - Why it's a mechanical process
0:34 - Countries work like individuals, "except countries can print money"
1:28 - Spending $7, taking in $5
2:49 - Why the market goes down
3:38 - His timeline
4:15 - The boom that's now "heavily debt financed"
"Think about this like a business or an individual. They spend $7. They take in $5, spending 40% more than they're taking in"
Debt → Debt service → Less spending → Crisis
Most people watch the interest rate
Dalio watches how much of the cash flow goes to paying the debt
Better than any $1,000 finance course I've seen
This is f*cking insane. Ray Dalio on the US debt crisis:
"Is a debt crisis imminent in the US?" "I think within the next three years"
In 5 minutes, the founder of Bridgewater explains exactly how countries go broke:
0:24 - Why it's a mechanical process
0:34 - Countries work like individuals, "except countries can print money"
1:28 - Spending $7, taking in $5
2:49 - Why the market goes down
3:38 - His timeline
4:15 - The boom that's now "heavily debt financed"
"Think about this like a business or an individual. They spend $7. They take in $5, spending 40% more than they're taking in"
Debt → Debt service → Less spending → Crisis
Most people watch the interest rate
Dalio watches how much of the cash flow goes to paying the debt
Better than any $1,000 finance course I've seen
Dmitro Cashback retweeted
CINDERELLA HAS A NEW REASON TO LEAVE AT MIDNIGHT
A palace ballroom, chandeliers, a gown made of pure glitter
→ the clock hits midnight
→ she runs out under the full moon
→ past the crystal carriage
→ and leaves one glass slipper on the street
The creator's caption: "when it's midnight and you still have skincare to do"
633,259 likes, 12M views
This is AI and it's INSANE. The sequins, the moonlight, the slipper on wet stone, all generated
A full Disney-level scene from one creator. No set, no costume team, no budget
Seedance 2.5 on Picsart puts this kind of shot on your phone
Would you have left the ball for skincare?
Dmitro Cashback retweeted
This is f*cking insane. Citadel CEO Ken Griffin on the AI bubble:
"There's almost certainly bubbles within this moment in time. And any radical transformation of our economy is generally intertwined with a bubble"
In 4 minutes, one of the best traders alive explains how to think about it:
0:27 - Why every revolution comes with a bubble
0:53 - "I just wish I'd kept my Amazon stock"
1:36 - The railroads lesson
1:51 - Who wins after the bubble
2:33 - "I've seen the entire industry deleverage to zero at least 10,000 times"
3:25 - The rule that keeps him learning
"The companies that are on the other side of that bubble are the market leaders for decades to come"
"The minute I start to use an excuse, I actually stop learning"
Railroads → Internet → AI → New market leaders
When markets drop, most people blame hedge funds deleveraging
Griffin puts that at option 10 on the list, not option 1
Better than any $1,000 finance course I've seen
Dmitro Cashback retweeted
This is f*cking insane. Nvidia CEO Jensen Huang on AI and jobs:
"The narrative about jobs is almost precisely backwards. It's precisely wrong"
In 90 seconds, he flips the biggest fear about AI:
0:19 - Why the jobs narrative scares people
0:25 - "AI has created more software jobs"
0:44 - Why energy is booming
1:22 - "It's precisely wrong"
"The facts are that AI has created more software jobs. The facts are that AI has created more healthcare jobs. The facts are even AI is creating legal jobs"
His five layer cake: energy, chips, infrastructure, construction, manufacturing, all hiring
Energy → Chips → Factories → New jobs
Most people think AI is coming for their job
Jensen is pointing at every layer that's hiring because of it
Better than any $1,000 AI course I've seen
This is f*cking insane. Citadel CEO Ken Griffin on the AI bubble:
"There's almost certainly bubbles within this moment in time. And any radical transformation of our economy is generally intertwined with a bubble"
In 4 minutes, one of the best traders alive explains how to think about it:
0:27 - Why every revolution comes with a bubble
0:53 - "I just wish I'd kept my Amazon stock"
1:36 - The railroads lesson
1:51 - Who wins after the bubble
2:33 - "I've seen the entire industry deleverage to zero at least 10,000 times"
3:25 - The rule that keeps him learning
"The companies that are on the other side of that bubble are the market leaders for decades to come"
"The minute I start to use an excuse, I actually stop learning"
Railroads → Internet → AI → New market leaders
When markets drop, most people blame hedge funds deleveraging
Griffin puts that at option 10 on the list, not option 1
Better than any $1,000 finance course I've seen
Dmitro Cashback retweeted
This is f*cking insane. Anthropic CEO Dario Amodei:
"I looked at it and I said, my God, you just scale this thing up another factor of a million and it'll be better than humans at everything"
In 3 minutes, he explains the moment he saw it coming:
0:18 - Why GPT-2 was the first real test of scaling laws
1:29 - "I personally was super freaked out by it"
2:12 - Only 10% of people got it, "often the smartest ones"
2:46 - Why the skeptics keep saying the same thing
3:09 - The 6 month / 12 month rule
"This is the defining economic and scientific moment of the modern age. We just don't know it yet"
"Six months from now, the things that seemed hard, they're gonna be easy. 12 months from now, the things that seemed impossible, the models are gonna be doing them regularly"
Scale → Generality → Hard → Easy → Routine
When AI looks dumb at one task, most people dismiss it
Dario is teaching you to watch the curve, not the task
Better than any $1,000 AI course I've seen
Dmitro Cashback retweeted
This is f*cking insane. Elon Musk on merging Tesla and SpaceX:
"This is definitely the kind of question where I feel like I'm tap-dancing through a minefield"
In 3 minutes, he explains exactly where the biggest merger talk on Wall Street stands:
0:38 - Why he has to be careful
0:49 - Both boards formed special committees
1:18 - The "odd kabuki dance" he can't be part of
1:38 - Why calling it a certainty is "overstating it"
1:57 - The 50% vote that makes it hard
2:28 - Why he thinks it's a good idea
"Do I think it would be a good idea? Yes"
But Tesla needs "a 50% vote of all shares, not just those who vote"
His case: superintelligence software, chip design, Terafab, robotics, mass to orbit, Starlink connectivity
Rockets → Cars → Chips → One company
When a merger headline drops, most people buy the hype
Musk is telling you the real hurdle is the shareholder vote
Better than any $1,000 finance course I've seen
This is f*cking insane. Elon Musk on merging Tesla and SpaceX:
"This is definitely the kind of question where I feel like I'm tap-dancing through a minefield"
In 3 minutes, he explains exactly where the biggest merger talk on Wall Street stands:
0:38 - Why he has to be careful
0:49 - Both boards formed special committees
1:18 - The "odd kabuki dance" he can't be part of
1:38 - Why calling it a certainty is "overstating it"
1:57 - The 50% vote that makes it hard
2:28 - Why he thinks it's a good idea
"Do I think it would be a good idea? Yes"
But Tesla needs "a 50% vote of all shares, not just those who vote"
His case: superintelligence software, chip design, Terafab, robotics, mass to orbit, Starlink connectivity
Rockets → Cars → Chips → One company
When a merger headline drops, most people buy the hype
Musk is telling you the real hurdle is the shareholder vote
Better than any $1,000 finance course I've seen
Dmitro Cashback retweeted
This is f*cking insane. Gavin Baker on the Magnificent 7:
"The aggregate market cap, which today is probably 10 to 12 trillion, will for sure go to 50 to 100 trillion
But one or two of those companies will go extinct"
In 5 minutes, he explains exactly why:
0:00 - Why Mag 7 goes to $50-100T
0:25 - Why one or two of them die
2:06 - The only 3 model companies he thinks survive
2:47 - Where there is no bubble
3:56 - Why SaaS is a bet against ASI
They used to compete in different events
"They're all competing in the same event and it's called AI. You either win or die"
His survivors: Google, Meta and xAI, because they have unique data
"At the infrastructure layer, there is no bubble"
Apps → Models → Data → Chips
Per-seat software is the old trade
Infrastructure is the safe one
Better than any $1,000 investing course I've seen
This is f*cking insane. Gavin Baker on the Magnificent 7:
"The aggregate market cap, which today is probably 10 to 12 trillion, will for sure go to 50 to 100 trillion
But one or two of those companies will go extinct"
In 5 minutes, he explains exactly why:
0:00 - Why Mag 7 goes to $50-100T
0:25 - Why one or two of them die
2:06 - The only 3 model companies he thinks survive
2:47 - Where there is no bubble
3:56 - Why SaaS is a bet against ASI
They used to compete in different events
"They're all competing in the same event and it's called AI. You either win or die"
His survivors: Google, Meta and xAI, because they have unique data
"At the infrastructure layer, there is no bubble"
Apps → Models → Data → Chips
Per-seat software is the old trade
Infrastructure is the safe one
Better than any $1,000 investing course I've seen
Dmitro Cashback retweeted
This is f*cking insane. Yann LeCun:
"There is a mistake that people in AI have kept making for the last 70 years"
If a machine can do a task humans see as intellectually complex, "automatically this machine is smarter than humans"
"And we know this is completely false"
In under 2 minutes, he explains why:
→ computers already beat us at equations, chess and Go
→ "this does not mean that we are anywhere close" to human or even animal intelligence
→ it's "much simpler" for LLMs to handle language, code and math
→ "The real world is much more complicated. It's messy, it's not entirely predictable"
When AI wins a benchmark, most people call it smarter than us
LeCun is pointing at what's still missing:
Symbols → Language → Code → The real world
LLMs mastered the easy part
The physical world is the next one
Better than any $1,000 AI course I've seen
This is f*cking insane. Yann LeCun:
"There is a mistake that people in AI have kept making for the last 70 years"
If a machine can do a task humans see as intellectually complex, "automatically this machine is smarter than humans"
"And we know this is completely false"
In under 2 minutes, he explains why:
→ computers already beat us at equations, chess and Go
→ "this does not mean that we are anywhere close" to human or even animal intelligence
→ it's "much simpler" for LLMs to handle language, code and math
→ "The real world is much more complicated. It's messy, it's not entirely predictable"
When AI wins a benchmark, most people call it smarter than us
LeCun is pointing at what's still missing:
Symbols → Language → Code → The real world
LLMs mastered the easy part
The physical world is the next one
Better than any $1,000 AI course I've seen
Dmitro Cashback retweeted
This is f*cking insane. Elon on Grok 5:
"It will be the largest model to the best of my knowledge
So this is a six trillion parameter model, whereas Grok 3 and 4 are based on a three trillion parameter model"
In 1 minute, he explains exactly what changes with Grok 5:
→ 6T parameters, double Grok 3 and 4
→ "much higher intelligence density per gigabyte"
→ "the quality of the data that we're training on with Grok 5 is much higher"
→ "inherently multimodal, so it's text, pictures, video, audio"
→ "much better at tool use and in fact, creating tools"
When people compare AI models, most of them look at size
Elon is pointing at a different metric:
Parameters → Data quality → Intelligence per gigabyte → Tools that build tools
Bigger models are the old race
Smarter per gigabyte is the new one
Better than any AI course I've seen this year
Dmitro Cashback retweeted
OpenAI co-founder Andrej Karpathy wrote the loop and just released a prompt built on it. Now Opus 5.5 finishes jobs it used to quit halfway.
This is f*cking unreal...
Any task runs as a graph: one check per node, a baseline before any change, and only the wins survive.
Every dead end gets written down, so tonight's run starts where last night's gave up.
$2,000 agent courses teach less than this one prompt.
Send it to Opus 5.5 and thank me later.
The full prompt and the full article are both below ↓
