I could not put into words how much today has meant to me. I’ve been following Intel ever since I was a child and it is not an exaggeration to say they are the reason I fell in love with chips and computers in the first place. I wouldn’t have been able to take the path I have chosen in life if it wasn’t for the thousands of brilliant hard working engineers at Intel who laid the foundation for the ground I walk on. When I started following their bold bet on becoming a world class cutting edge foundry about a year ago I was galvanized by the vision that they set out. It wasn’t just important to me that Intel succeeded it was important to me that America succeeded. Silicon Valley was built by them and their cohort and it was critical to me that we not only maintained our standing in the semiconductor industry but for us to grow our industry so 100k other flowers could blossom. There was no way this could be done without an American foundry on the leading edge. I believed in the plan from the start, and as an engineer myself I was absolutely confident that Intel was developing some of the world’s greatest technology and IP from 18A to backside power delivery to the amazing breakthroughs they’ve achieved in advanced packaging which I like to dub as the new Moore’s Law. They are just getting started and there is without a shadow of a doubt more to come. Today I was vindicated for everything I’ve been talking about for over a year now. The semiconductor industry is won by those not who focus on optimizing margins but those who invest in innovation to bring value to the whole world. Lip Bu Tan and the whole Intel team have delivered on innovation… and again this is just the beginning. I will be here waiting, watching, cheering every step of the way. Thank you from the bottom of my heart for the experience, the gifts, and your words of appreciation they mean more to me that you can possibly imagine. With love, Bubble Boi 🫧♥️

Bro starting to look like an Albanian from the Bronx.

The economy is propped up by women, they spend money like nobody else. I've been around billionaires who watch their dollars closer than the average 20 year old in a big city.

Imagine pulling this out after some hot pot and tasting the spirit of a nation

Palantir is fucking worthless.

Software engineers have negated every advancement in transistor density, computer architecture, compilers, and computer networking over the past 30 years.

Hahaha this is soo good hahahahaah. MORE MORE MORE. I WANT MORE YUM YUM 😋

If universities were actually good they wouldn’t have like 3% acceptance rates and 99% graduation rates it would be like 40% acceptance rate and 5% graduation rates like the Navy Seals.

Most insane github I've ever seen in my life lol.

> Be Chinese > Take everybody’s innovations > Improve it 10x > Give it away for free? why do they keep doing this? :

Holy shit Powell really said I don’t really give a fuck about inflation we can’t have 2 million unemployed CS grads.

> checks profile thinking it’s ground breaking technology > it’s tinder for jobs Burn down the Bay Area.

I am an expert on bubbles. So it brings me no joy to say that the AI bubble is popping this time next year. When you promise infinite scaling and don’t produce it the calculus changes. I don’t think it will be bad for most companies but those who built their entire business model around making the best LLMs are unfortunately going to struggle as models become more of a commodity. The end user doesn’t care much if Claude is 5% better than GPT5 they care about costs, speed, and utility especially at the scale things will be going. The obvious play now is shorting Nvidia & dumping your OpenAI options in the secondary market. The new winners will be inference chips with low TCO as everything on the stack turns into commodity network switches and NICs. Other winners will be those who are at the new frontiers of using AI for novel applications from drug development to teaching.

“It’s already difficult to find good engineers” Agreed. I’ve had a LOT of trouble convincing several Berkeley, Caltech, & Stanford PhDs with specialties in photonics, nano fabrication, and metamaterials to turn down tenure track professor job offers to join my company as well. Oh … wait. What do you guys do? An AI app for note taking? And you can’t find people who can write Python?

Disliking McDonald’s is middle class pmc coded. It’s literally amazing food at a reasonable price all over the world.

Probably the most wild paper I’ve ever read… group of researchers proved they can generate electric power from the earth’s own magnetic field. This was thought to be impossible like just a few months ago. The experiment is pretty solid, they created a cylindrical shell of manganese-zinc ferrite and blocked out any RF radiation that could possibly induce a charge. They even show that when the cylinder is orthogonal to the Earth’s magnetic field they get no voltage potential which matches the theory. They only get a very small amount of volts though but like holy shit maybe you can do something with it ?

People who want to get good will get good. These are motivated people, the same people who 40 years ago would go to a library and check out a textbook on a programming language. Now it’s just easier than ever to get good. Meaning if your mediocre your terrible. The new bar is good.

Yeah I guess you got a point. I mean if you’re trying to become a world class mathematician that can win a Field’s Medal you need to start early or else you might not have a chance……. SIKE! Last person to win a field’s medal didn’t even major in math in college he majored in Journalism!!!! He didn’t start doing any real math until his *checks notes* 6th YEAR IN COLLEGE. You can literally GRIND your way to ANYTHING !!!!

Husbant I max out cwedit cah and flee to chyna now we rich

Not to say I saw this coming but the average ML engineer knows nothing about computers I’m not surprised people whose job it is to make hardware perform at its maximum is mogging them. This is the result of everyone hiring people who can solve leetcode hards in Python and don’t know the difference between big endian and little endian. The average software engineer is a glue eater whose job is to glue together libraries and packages not do anything innovative or interesting. We need to change how we train people.

Absolute culling of the herd. The most dangerous thing you can have is idle hands and lots of energy.

There are arbs for those with eyes to see them.

Someone needs to write an anthropological study on the BayArea ABG.

If you are doing low level systems programming or hardware development you need Linux. Every compiler, every tool, every driver, it just works on Linux and there is always a Linux build you can use. Thinkpads are great for that. Good battery life, good I/O, all drivers work. I’m always skeptical when I go into a “hard tech” startup and see 30 MacBook boxes stacked up. You can get 3-4 really good thinkpads for the prices of a new MacBook Pro.

Albanians in UK: Coke dealers Albanians in USA: CTO of OpenAI, Tech Founders lol same people just different incentives

It’s so over for me

Self driving car? > PID loop Rocket that lands itself? > PID loop Hypersonic ICBM? > PID loop Your thermostat? > PID loop

No you didn’t

Every engineer after getting good at C++

To put into perspective how big of a fucking bubble we are in Nvidia is 3.8% of world GDP, and 14.7% of US GDP… 😞

What if we just do this?

There is a random guy named James Groth who the CME lists as a liquidity provider for Treasury Futures Block Trades… The crazy part is it’s just one guy.. not a company. You literally email him directly at his hotmail address to do a 10 million dollar trade. To put into context how wild this is the other liquidity providers are all major banks or billion dollar trading firms lol. All I can find about him is he got fined a few years ago for spoofing the treasury market.

Why am I seeing so many of these posts ? You guys know Stanford is a university and not some VC fund looking to back a SaaS founder.

Behind every electronics company there is some autistic IIT grad or Iranian PhD adding billions of dollars in value.

How is this possible?!

Every AI company founding team: -Twink drop out -Chinese phd -less than 5 years of cumulative experience Every chip company founding team: -most Indian name you can't even start to pronounce -random guy named bob or bill -80 years of cumulative experience

You can’t move Wall Street to Miami because the exchanges are actually in data centers in Secaucus New Jersey lmao.

Hey @sama I am a proud $200 a month payer for ChatGPT Pro and have been trying out your DeepResearch tool extensively. And it’s fucking ass !!! But it’s not your fault. That’s why I’m reaching out… the issue is all these articles it cites are paywalled so DeepSeek is just summarizing the abstracts for me… You can fix this. Fuck these academic publishing companies get subscriptions and train on all those articles bro. Let’s stop gate keeping knowledge and get real AGI !!!

If you are graduating college in 2026 you have three career paths: -Underpaid entry level matcha dispenser to the wealthy white girl SUV caste -YC GPT Wrapper Entrepreneur -Bubble (bad) / Theil (Evil) Fellow -Borderline Schizo Quantitative that can past the 10 round Jane Street Rain Man Interview

I remember my boss at my last job was like: "wow you're 27 you got the big promotion we don't usually promote people that fast, keep up the good work " I was like DAMN you're glazing me CRAZY. I quit the next month.

> can you code ? > no > any academic credentials ? > none > ever shipped a product before ? > nope > lead a team of people? > never > So can you then explain why Garry Tan & Marc Andressen follow you? >They follow me for my unique taste i am like like a sommelier for the TL

The white guy having the lowest depth tree is so funny

All these guys have philosophy degrees.

Yeah CS isn’t a science who cares. Electrical engineering isn’t either. Both of them are a mix of art & alchemy. The best majors are philosophy & math as they actually teach you how to think. The rest are just job programs.

BREAKING 🚨🚨🚨 LEOPOLD ASCHENBRENNER HEDGE FUND SITUATIONAL AWARENESS MAKES OVER $1 BILLION DOLLARS IN A DAY ON NEWS OF NVIDIA INVESTMENT AND COLLABORATION WITH INTEL. SITUATIONAL AWARNESS IS NOW UP OVER 100% YEAR TO DATE.

PID loops can do anything man

Out of all the billionaires in the arena it is @sama I respect (fear) the most. Going from practically a nobody to climbing the hierarchy to head YC and basically making it the brand it is today. Then bowing out to run OpenAI before there was much proof it was going to work out & building without a doubt one of the most valuable companies on earth. If OpenAI was public we would’ve bid it to 3 trillion easy and then lost our shirts when GPT5 came out. He’s been the chief architect of this bubble to the point that some of the phrases and words we use today would literally not be in our vernacular. Scaling Laws, AGI, infinite growth, sentience it’s all him peddling this to us. Working the capital markets like the shrewdest banker on earth, two fundraises at different valuations at the same time… WHO THE FUCK CAN PULL THAT OFF?! To baiting Zuck with FOMO by getting his chief lieutenants to feed poison to his brain. To then rugging him by going to the press to spoof the order book on AI Researcher salaries and at the same time saying everyone who joined his Super Intelligence team was mediocre. The GPT5 failure to the dumping of shares to anesthesiologists in San Francisco… it’s all him. Then finally, he calls it a bubble. He admits he was exuberant. But yet, he can’t be saying OpenAI is a bubble. It’s the other guys who’s models are just slightly worst and don’t have the distribution the infrastructure we do. See it’s not AGI 2027 it’s we own the railroads we own the oil wells. Do you really care about where your token comes from? You just care that it’s cheap and gets you to where you need to be. His next move if I had any guess is vertical integration. The new metric isn’t ARC AGI retards its Profit per Token (PPT). Write it up @SemiAnalysis_ there is a new AI race and it’s not going to be won in a Jupyter Notebook.

They make things

Idk who needs to hear this but stop taking adderall for a company you don’t have any equity in lol.

I have a real job

Proud to announce I am joining @JaneStreetGroup in the Fall as a Hardware Engineering Team Lead in their South Sudan office.

OpenAI is making $4.5 billion in revenue and has a net loss of $13.5 billion a year with OPERATING LOSSES of $7.8 billion. The 10% stake in AMD is a warrant that is gated by gigawatt deployment using their chips. OpenAI gets free shares for buying and building out up to 6 gigawatts of data center capacity using AMD GPUs. That would be worth $34.5 billion dollars in value to them today! Letting them leverage their balance sheet to get loans to build even more data centers ! BUT WHERE ARE THEY GOING TO GET THE MONEY TO BUY AMD CHIPS?!? Nvidia promised them $100 billion with a similar deal based on $10 billion per gigawatt of scale out using their chips. So is the idea build the first 1 gigawatt data center using Nvidia chips and then reinvest that to build an AMD one ??? This is not going to happen. This would mean 17 GIGAWATTS OF AI DATACENTER CAPACITY JUST FOR OPEN AI. This would mean OpenAI would need to build 17 xAI Colossus 2 data centers and would need two sworn enemies to work together. Sama has gone too far. The demand isn’t there yet. Jensen is pulling out of that deal he’s not subsidizing a fucking competitor. Sam is playing both sides and rushing to grab land before others notice with nothing but promises. It’s a house of cards with only one person being able to pull out and make it collapse. Implosion incoming !!!!

This is how I win

I attribute a lot of my success to the Mother Theresa figurine I keep at my desk praying for me as I write Verilog.

Incredibly based and the right way to solve problems not this humanoid trash.

Kinda wild that these guys alone added multiple trillions of dollars to world GDP & increased the standard of living on the planet more than any government or charity.

Michael Burry has called GME, the inflation trade post Covid, 2022 market decline, 2023 bank stocks, rally in Chinese equities in 2024/2025 & the run up in gold. Not saying he will be right this time but to act like he isn’t an actual good fund manager is dumb.

31k commits in 4 months like whaaaat lol

My favorite finding about IQ is that the only thing that maintains a strong correlation with increasing IQ scores is a high correlation with being in financial distress. Not paying bills, going bankrupt or maxing out credit cards, is related to IQ scores not linearly but instead quadratically lmaooo. So if your yoloing your life savings into options every week and blowing a bag at the Rick store your probably a fucking maniacal genius.

It’s low status everywhere except America and even here the tides are turning.

If even a quarter of what Burry is saying is true Oracle, Coreweave, Nebius, Meta, gapping down 30%. I have no reason to not believe it either. Everyone knows the b200’s overheat and at best last 12 months. With water cooling issues as well a failure is a total loss… it’s an insider secret nobody wants to stand up and talk about in fear of retaliation.

This is such a funny tweet that blew up because it’s literally just cope. Like oh wow you didn’t learn analysis at 13 years old? Well better late than never retard. The mindset is so cringe. I’m a GRIND MAXXER, I’m spiritually Chinese. Anything I want to learn, I can learn. Ask anybody who knows me they are like bubble how the fuck did you learn this? I just get the textbook sit down and read it. You can learn ANYTHING. Strength, will power, and never giving up. This is the difference between mediocre and greatness.

Stanford & Berkeley are like the only universities left in the world that can design fast chips. There is also like this one professor at Yale who can too but it’s just him he’s the whole ECE department. It’s a real problem in my eyes because we need to train the next generation of computer engineers to make a ton of chips. And outside of Broadcom, Intel, & Marvell very few are capable of the physical design skills necessary to make fast chips. In the next twenty years it’s possible that physical design will be so automated but still mediocre that we are just stuck with mid basic ass chips. It keeps me up at night I’ll tell you.

Citadel wrote a really good paper where they reverse engineered GPU architecture and did benchmarks. This was back in 2018. Of course much has changed architecturally now and the networking of the GPUs matters a lot as models get bigger. Point being that yeah this probably was actually a side project. This is the result of ML researchers constantly abstracting away the hardware. Citadel and DeepSeek’s business relies on them understanding it at a deep level. Stop coping and realize hardware is the frontier not some model architecture that does not matter.

Some points to consider: -These machines run 24/7 & can’t ……ever fail, down time is millions of dollars an hour. Believe it or not ASML needs to pay for that down time to the fab. -The forces on all these parts are EXTREMELY HIGH. An F-22 raptor pulls max 9 Gs & after that needs extensive maintenance. Just the stage on these machines are pulling over 10 Gs & while that thing is accelerating that fast it has to be perfectly aligned with the wafer below. This isn’t even considering the laser… -Also at least for EUV this is all done in a vacuum as well where you have tons of physics problems to take into consideration. You need more voltage, buffering on any moving part or else it will accelerate indefinitely … it’s literally like engineering for a space station. -Another thing many people don’t consider is throughput is a big big deal to a fab. It’s not just about the cost of the machine but also cost per wafer. The DUV machines do 300 wafers per hour. For a 300 mm wafer that’s 1,052 chips or 315,600 chips an hour lol.

Guys I think some of these H1Bs are pretty good.

Every good FPGA engineer I try to hire: -Sorry I’m in rehab for next few months let’s talk next summer when I’m clean -I’m not really looking to work anymore, have been turning down jobs left and right and living off my Intel profits.

Your unemployed friend on a Tuesday:

Shitting on someone’s business ideas lol

Me on my way to train a linear regression model in 2033

If your circle isn’t talking about: EUV Lithography, Advanced Modular Nuclear Reactors, Photonic Biosensors, or Energy Based Models …. You need new friends

Nvidia will deny this lol. Of the top ten models 2-3 are not trained on Nvidia GPUs… for inference it’s probably worst. 🫧

OpenAI doesn’t have $100 billion dollars and that money is gated by Nvidia $10 billion at a time based on gigawatt buildout. It’s on OpenAI to deliver it. They don’t have the money for the AMD deal, it’s an option for 10% of the company they don’t own it yet. Sam has crossed the one person you don’t cross. He only exists because Jensen says so. It’s going to get ugly here.

Not to be a hater but idk how you can get anything done in this kind of environment. Maybe these people are really good at it though.

You guys are midwits. Case statements compile into jump tables while if statements are comparisons and rely on the branch predictor. This guy is pulling a pro gamer move.

Does anyone ever get tempted to just blow up their entire life from time to time or that only a me thing?

Lmaooo this is hilarious

It is wild to me how little Deep Learning researchers know about basic statistical theory. Everyone acts like all to all attention is a free lunch while basic stats has shown many better ways to capture long range dependencies instead of comparing every token to each other. Ironically, I don’t think transformers success is as much to do with attention as it is to do with discretization (tokenization) which HMM used to great effect and makes modeling long range dependencies much more tractable. Dense attention is compute hungry ESPECIALLY under autoregression. Just using basic intuition a model that first zooms out, then drills down mirroring natural dependency hierarchies feels mooore right... Maybe it’s time we actually think first before wasting a quadrillion flops.

I don’t think people understand how big of a deal this is and what the ripple effects are. If OpenAI starts now pivoting from shipping big new models to focusing on infra and serving which they will have to do to be profitable btw this means 1) Less resources dedicated to big training runs & the death of the $100 million dollar AI researcher who only knows PyTorch. 2) More resources going to optimizing inference and that’s a huge sea change for hardware and the rise of the gigachad line rate byte parsers. This is bad for Nvidia, very bad. They have a monopoly on training but you can do inference on anything even a plain old CPU and if we are optimizing costs this means people will be more than willing to struggle and work with any inference stack because any energy savings per token is billions in profit at this scale. Nvidia networking with BlueField and Infiniband is not the moat it looks like we can modify Ethernet to do much of the same thing and even better cause at the end of the day PHY is just a PHY and Fiber is just Fiber… There is a huge opportunity here now and I’m not going to lie I saw this months ago and that’s why I named my company “Bubble Computing” @bubble_comp. Me & my team knew that the writing was on the wall the second Sam Altman started looking for HFT engineers of all places .. hmm.. wonder what AI inference and HFT have in common… The bubble has popped, it’s picks and shovels all the way down & that’s what we are going to do. RIP AI Research Bubble 2023-2025

I am really against gambling generally.. but Kalshi & PolyMarket are really the worst of the worst. The model where you have to travel to Vegas to do your debauchery was ideal. You would go there lose a few grand maybe they comp your room, your dinner, get some free drinks and then leave never to talk about it again. There was this unwritten understanding that yes the odds are not on your side but you’re paying for an “experience.” But with these apps you just literally sit there staring at your phone all day losing money with none of the glamour or ritual. It’s really disgusting stuff & just a sign of the trend of companies that exist solely to rape the middle class.

Who's going to tell him that most of this wasn't algorithmic.

Economy Under Obama: “Yeah let me go to Vassar and get an English degree and put myself in 150k of debt there are tons of jobs out there for me” Economy Under Trump: “If you get a CS degree from CMU you will end up behind the counter at Chipotle fighting with a gym bro for an extra scoop of the honey chicken”

So many issues people have with their careers is that they want the lifestyle and income of someone in their 40s when they are 25 lol.

Is this Pittsburgh?

Jane Street traders be like “I’m stealing billions of dollars from India to donate mosquito nets to South Sudan to decrease the overall suffering in the world”

The best programmers I ever meet have one thing in common. Has nothing to do with what school they went to or when they started coding… They all had such a strong autistic interest to build something they didn’t give a fuck what it took. They had a religious level of motivation to make it happen. Terry Davis is the arch type example. But I literally know a guy who wanted to make a mod to a video game he played so he can add teleportation portals and ended up learning all the physics of a game engine and C++ on his own…. He had no exposure to programming or math beyond high school level before this.

Girls always come into your life right before your about to solve the Riemann Hypothesis

This guy has aura but it ain’t the good kind

LMFAOOOOOOOOOOO

Update I replaced (X/Y) with X*(1/Y) and now it passes timing check. I learned this trick from some girl from MIT at my last job lol. Everyone is blown away with my skills. There is no reason why it’s inherently faster just synthesis tools optimize it better.

Dudes write one CUDA kernel and start talking like this

Honestly we should just give her another chance. If shipping a demo that doesn’t really work was illegal half of the founders in SF would be in prison right now.

Instead of non competes Jane Street decides to make its own programming language so far from anything mainstream that you can't get a job anywhere else.

PR firms man… despite what you think every fortune500 company C suite exec has them.

It’s not impossible. You just don’t want to compute the derivative so you don’t. Just like how it’s impossible for me to memorize the Krebs Cycle. If I wanted to do it I could. I just don’t want to. People on this app think because they can do undergraduate math that they are like the Oracle of Delphi or some shit. Like you would probably barely pass a biochemistry class. Stop falling for this psyop. You can do whatever your heart desires.

This is why I own Intel stock.

I am SHORT: -Nvidia -Coreweave -Palantir I am LONG: -Intel

Imagine being an unemployed CS grad reading this. The mindless normie rush to CS has probably been the biggest rugpull in history.

All this for Amazon?

The trick to succeeding at any company is just to act like you own the company. You wouldn’t say “oh well that’s not my job” if it affected your income that year but many people are so wagie brained they only care about their own paycheck. When you realize it is your job to make sure everyone is doing what needs to be done to hit your target it actually becomes easier to be a superstar. If someone is struggling it’s your job to help them, if you don’t like that then you better learn how to motivate them or train them to get good quickly so you can force multiply your efforts. You need to take responsibility for everything just like an owners does and thinking your job is to only do your job is the reason you are probably not moving up. Again no one will say this explicitly to you but it’s kind of fucking obvious. Good engineers aren’t just good engineers cause they do their work fast and log off lol. You need to make everyone better and if your not doing that your not going anywhere on your life and career.

Had to be done sorry…