Olivia
Online
Hawking
@st3vehawk
Turning crypto chaos into clean workflows | n8n + web3 + AI | Sharing everything in public
^^^^^
Joined June 2026
27 Following
20 Followers
96 Posts
A guy bought boxes of retired Nvidia Tesla cards to build a local AI server for less than one month of subscriptions. The math works. The parts it skips are the ones that bite.
The concept is real. Old datacenter GPUs - the P100, M40, P4 - once cost thousands and now sell used for a fraction. A 24GB Tesla M40 for around $130 is genuinely a lot of VRAM for the money.
But three costs get left out of the "$165 pays for itself in weeks" math.
Power. The M40 draws 250 watts and has no active cooling of its own - that's why it needs the fan shroud. Run it continuously and the electricity plus the heat it dumps into your room is a recurring cost the one-time price hides.
Speed. These cards are from the Maxwell and Pascal era. They lack modern tensor cores and run at a fraction of current inference speeds. A summary that takes seconds on a subscription service can take meaningfully longer locally. You're trading dollars for latency.
Model quality. A local model on an M40 handles summaries, transcription, and overnight batch jobs well. It does not match Claude or GPT on hard reasoning. The post frames it as replacing a $412 stack - but it replaces the boring 70%, not the frontier work you actually pay premium for.
The honest version: this is a fantastic setup for high-volume, low-stakes tasks you run constantly.
Buy the M40. Route the boring work to it. Keep one subscription for the thinking that matters.
The card is cheap. The tradeoff is speed and ceiling, not price.
https://t.co/nOMpBIZbkG
An MIT professor explained the one idea funds pay $500K for in the first hour of freshman calculus. It's been free online for twenty years and almost nobody finishes it.
His point cuts against how everyone was taught. Calculus isn't a hundred formulas to memorize. It's one idea.
Take two points on a curve. Slide them together until the gap vanishes. The slope you're left with is the derivative. Everything else is just consequences.
He's David Jerison, and his 18.01 lectures have taught more people calculus than any room on earth.
Skip to the board where the power rule, the product rule, the chain rule - the fifty things you were forced to memorize - all fall out of that single limit. No formula sheet. No tricks. One definition, reused.
That same slope is velocity. It's marginal cost. It's the gradient that trains a neural network. One operation, hiding everywhere something moves.
The reason it matters: most people learned the fifty formulas and never the idea underneath them. They can compute a derivative and have no sense of what it is. Which means the moment a problem doesn't match a memorized pattern, they're stuck.
Understanding the one limit means you can regenerate all fifty formulas from scratch. Memorizing the fifty means you're helpless the moment one is missing.
The math is free.
What nobody can sell you is the patience to sit with one idea instead of cramming the fifty that fall out of it.
That hour is the only part that ever pays.
An MIT professor explained the one idea funds pay $500K for in the first hour of freshman calculus. It's been free online for twenty years and almost nobody finishes it.
His point cuts against how everyone was taught. Calculus isn't a hundred formulas to memorize. It's one idea.
Take two points on a curve. Slide them together until the gap vanishes. The slope you're left with is the derivative. Everything else is just consequences.
He's David Jerison, and his 18.01 lectures have taught more people calculus than any room on earth.
Skip to the board where the power rule, the product rule, the chain rule - the fifty things you were forced to memorize - all fall out of that single limit. No formula sheet. No tricks. One definition, reused.
That same slope is velocity. It's marginal cost. It's the gradient that trains a neural network. One operation, hiding everywhere something moves.
The reason it matters: most people learned the fifty formulas and never the idea underneath them. They can compute a derivative and have no sense of what it is. Which means the moment a problem doesn't match a memorized pattern, they're stuck.
Understanding the one limit means you can regenerate all fifty formulas from scratch. Memorizing the fifty means you're helpless the moment one is missing.
The math is free.
What nobody can sell you is the patience to sit with one idea instead of cramming the fifty that fall out of it.
That hour is the only part that ever pays.
https://t.co/fJgVOFIBhG
Arthur Blank was fired at 35 and opened a store that became a $340,000,000,000 company. it was called The Home Depot.
it was 1978. Blank ran finance at a hardware chain called Handy Dan. a corporate fixer was brought in to clean house - and pushed out both Blank and his boss, Bernie Marcus, on the same afternoon.
a friend told them they'd just been "kicked in the rear end by a golden horseshoe."
they didn't know it yet, but he was right.
over a year of coffee-shop meetings in LA, the two fired executives designed a store built to bury the company that let them go - warehouse-sized, everything under one roof, lowest prices in the market, staff who actually knew what they were selling.
opening day flopped. barely anyone showed up. so they gave their own kids handfuls of cash and sent them inside, just to make the store look busy.
it worked. that store became The Home Depot - now 2,300 locations, 470,000 employees, worth around $340 billion.
Blank walked away in 2001 and bought the Atlanta Falcons for $545 million. the team is worth $6.3 billion today.
the safe job was the ceiling. as long as Blank had a title and a salary at Handy Dan, he'd never have built the thing that replaced it. getting fired didn't end his career - it removed the only thing stopping it.
the worst phone call of his life was the one that made him a billionaire ↓
Richard Feynman picked up one piece of chalk in 1964 and explained the exact problem every AI lab is burning billions on in 2026.
The BBC filmed it. It has been free for 60 years. Almost nobody has watched it.
His point: nature only ever answers in mathematics. Every lab trying to force a language model to "reason" is slamming into the same wall Feynman mapped 62 years ago.
He was 46 here. The Nobel came 11 months later. The reels survived and now sit on YouTube with fewer views than a keyboard unboxing.
Skip to the blackboard in the middle. He takes one of Kepler's laws and rebuilds it from zero, in notation a twelve-year-old can follow.
No slides. No jargon. One piece of chalk.
The reason it matters for AI: Feynman isn't recalling the derivation. He's reconstructing it live from first principles. That's the exact capability separating a model that memorized proofs from one that can actually generate them.
Every benchmark measuring "reasoning" is trying to test for the thing Feynman is doing on that blackboard.
An ML engineer I know paused it four times and made his whole team watch before standup.
62 years old. Still free. Still the clearest picture of the problem nobody has solved.
Richard Feynman picked up one piece of chalk in 1964 and explained the exact problem every AI lab is burning billions on in 2026.
The BBC filmed it. It has been free for 60 years. Almost nobody has watched it.
His point: nature only ever answers in mathematics. Every lab trying to force a language model to "reason" is slamming into the same wall Feynman mapped 62 years ago.
He was 46 here. The Nobel came 11 months later. The reels survived and now sit on YouTube with fewer views than a keyboard unboxing.
Skip to the blackboard in the middle. He takes one of Kepler's laws and rebuilds it from zero, in notation a twelve-year-old can follow.
No slides. No jargon. One piece of chalk.
The reason it matters for AI: Feynman isn't recalling the derivation. He's reconstructing it live from first principles. That's the exact capability separating a model that memorized proofs from one that can actually generate them.
Every benchmark measuring "reasoning" is trying to test for the thing Feynman is doing on that blackboard.
An ML engineer I know paused it four times and made his whole team watch before standup.
62 years old. Still free. Still the clearest picture of the problem nobody has solved.
https://t.co/aqsY18NvVz
Sam Altman says the elites are close to building a genie that grants any wish. And they'll decide what the first wish is.
The framing here deserves scrutiny, because "genie that grants wishes" is doing a lot of quiet work.
A genie implies you state a desire and receive an outcome with no understanding of the mechanism, no control over the tradeoffs, and no ability to check whether the wish was granted correctly. That's not empowerment. That's dependence dressed as magic.
The tell is in who decides the first wish. If the technology genuinely served everyone equally, there'd be no single first wish to allocate. The fact that someone chooses reveals the actual structure - a small group controlling an immensely powerful tool, distributing access on their terms.
"It will benefit humanity" is the phrase every concentration of power has always used. It's not a lie exactly. It's just unfalsifiable, and it conveniently never requires giving up the control.
The genie metaphor also hides the real question. Every classic genie story is a warning about wishes granted literally, consequences unforeseen, the wisher worse off than before. Altman is invoking the image of unlimited power and skipping the part every version of the myth actually teaches.
The people who build the genie decide what it optimizes for.
"More people can wish for anything" sounds like democratization.
But the genie still belongs to whoever built it.
Sam Altman says the elites are close to building a genie that grants any wish. And they'll decide what the first wish is.
The framing here deserves scrutiny, because "genie that grants wishes" is doing a lot of quiet work.
A genie implies you state a desire and receive an outcome with no understanding of the mechanism, no control over the tradeoffs, and no ability to check whether the wish was granted correctly. That's not empowerment. That's dependence dressed as magic.
The tell is in who decides the first wish. If the technology genuinely served everyone equally, there'd be no single first wish to allocate. The fact that someone chooses reveals the actual structure - a small group controlling an immensely powerful tool, distributing access on their terms.
"It will benefit humanity" is the phrase every concentration of power has always used. It's not a lie exactly. It's just unfalsifiable, and it conveniently never requires giving up the control.
The genie metaphor also hides the real question. Every classic genie story is a warning about wishes granted literally, consequences unforeseen, the wisher worse off than before. Altman is invoking the image of unlimited power and skipping the part every version of the myth actually teaches.
The people who build the genie decide what it optimizes for.
"More people can wish for anything" sounds like democratization.
But the genie still belongs to whoever built it.
https://t.co/FVPaJOka6m
Warren Buffett spent 20 years donating $48 billion to the Gates Foundation. this year he gave them nothing.
for two decades it was the biggest philanthropic partnership in history - Buffett funneling Berkshire stock to Bill Gates' foundation, an irrevocable lifetime pledge he made in 2006.
then it stopped. his mid-year donation this year - nearly $6 billion in Berkshire shares - went entirely to foundations run by his own children. the Gates Foundation received none of it.
Buffett called Gates' ties to Jeffrey Epstein "distasteful." the two men, friends for decades and Berkshire board partners for 16 years, reportedly hadn't spoken since the DOJ released the Epstein files in February.
when asked if he'd ever give to the foundation again: "I'll wait and see what unfolds."
$48 billion over twenty years, ended without a public statement - just an omission everyone noticed ↓
Mel Gibson told Joe Rogan he has three friends who all had stage 4 cancer and now have none. Rogan named the drugs. Gibson nodded.
This is worth pausing on, because the format is doing something the claim can't support.
Three anecdotes, no names, no medical records, no timeline, no mention of the chemotherapy, radiation, or surgery those patients almost certainly also received. Just a nod on a podcast that reaches millions.
Ivermectin and fenbendazole are being studied for anti-cancer properties. Some early lab results are genuinely interesting. That's a real research thread worth following.
But "three friends recovered" is not evidence, and the way it's framed here quietly erases everything else those patients did. Stage 4 cancer patients who recover almost never recover from one repurposed drug alone. They recover from aggressive standard treatment, sometimes alongside experimental additions.
The danger isn't discussing these compounds. It's the implication that a dewormer replaced oncology.
Someone with stage 4 cancer watching this clip might delay real treatment based on three stories with no verifiable details attached.
The research deserves proper trials. The patients deserve better than being turned into a nod on a podcast.
Anecdote is where hypotheses start. It's not where medical decisions should end.
If you or someone close is facing this, talk to an oncologist before a clip.
Mel Gibson told Joe Rogan he has three friends who all had stage 4 cancer and now have none. Rogan named the drugs. Gibson nodded.
This is worth pausing on, because the format is doing something the claim can't support.
Three anecdotes, no names, no medical records, no timeline, no mention of the chemotherapy, radiation, or surgery those patients almost certainly also received. Just a nod on a podcast that reaches millions.
Ivermectin and fenbendazole are being studied for anti-cancer properties. Some early lab results are genuinely interesting. That's a real research thread worth following.
But "three friends recovered" is not evidence, and the way it's framed here quietly erases everything else those patients did. Stage 4 cancer patients who recover almost never recover from one repurposed drug alone. They recover from aggressive standard treatment, sometimes alongside experimental additions.
The danger isn't discussing these compounds. It's the implication that a dewormer replaced oncology.
Someone with stage 4 cancer watching this clip might delay real treatment based on three stories with no verifiable details attached.
The research deserves proper trials. The patients deserve better than being turned into a nod on a podcast.
Anecdote is where hypotheses start. It's not where medical decisions should end.
If you or someone close is facing this, talk to an oncologist before a clip.
https://t.co/t7e1VmXRZH
Terence Tao, the most decorated mathematician alive: "Hedge funds pay you $750K a year for one skill - telling a real pattern from noise. Almost everyone has it flipped."
His lifelong theme is the frame that matters here. Almost nothing is purely one thing. There's flawless structure, like a clock. There's flawless randomness, like a coin toss. Nearly everything real sits between the two, and the entire job is pulling them apart.
The primes are the cleanest test case. They look scattered and lawless. Yet Tao and Ben Green proved they contain evenly spaced runs of any length you ask for. Order was buried inside the apparent chaos the whole time.
That's signal detection with the marketing stripped off. It's also precisely what quant funds pay half a million dollars to do.
The uncomfortable part for anyone chasing that seat: the mathematics is free. Tao delivered this at UCLA and it's stayed public ever since. Every theorem, every proof, available tonight at zero cost.
What no one can sell you is the judgment. Knowing when a pattern is real and when your own eyes invented it is the whole skill - and it takes years of being wrong to build.
The lecture teaches the math. The math was never the bottleneck.
The bottleneck is learning to distrust the pattern you most want to be true.
Terence Tao, the most decorated mathematician alive: "Hedge funds pay you $750K a year for one skill - telling a real pattern from noise. Almost everyone has it flipped."
His lifelong theme is the frame that matters here. Almost nothing is purely one thing. There's flawless structure, like a clock. There's flawless randomness, like a coin toss. Nearly everything real sits between the two, and the entire job is pulling them apart.
The primes are the cleanest test case. They look scattered and lawless. Yet Tao and Ben Green proved they contain evenly spaced runs of any length you ask for. Order was buried inside the apparent chaos the whole time.
That's signal detection with the marketing stripped off. It's also precisely what quant funds pay half a million dollars to do.
The uncomfortable part for anyone chasing that seat: the mathematics is free. Tao delivered this at UCLA and it's stayed public ever since. Every theorem, every proof, available tonight at zero cost.
What no one can sell you is the judgment. Knowing when a pattern is real and when your own eyes invented it is the whole skill - and it takes years of being wrong to build.
The lecture teaches the math. The math was never the bottleneck.
The bottleneck is learning to distrust the pattern you most want to be true.
the man who runs a $1.5 trillion company once had to personally explain why his database kept crashing. he was 19, and the site was called TheFacebook.
in 2004 he ran it off rented machines from his Harvard dorm, describing the whole thing like a hobby.
"when we first launched we were hoping for maybe 400 or 500 people. now we're at 100,000. who knows where we're going next... maybe we can make something cool."
no business plan. no exit strategy. just a coder solving one problem at a time - how to scale to the next university, how to stop the servers from falling over, how to keep people coming back.
that "something cool" now serves over 3 billion people.
but here's what almost nobody in his position does.
he never sold it. Microsoft and AOL reportedly offered him millions for a program he built in high school - he said no. Yahoo offered $1 billion for Facebook when he was 22, with his entire board telling him to take it - he said no. most of his senior team quit within a year.
"I don't really like putting a price-tag on the stuff I do. that's just not the point."
the people who build the biggest things aren't optimizing for the exit. Zuckerberg wasn't trying to get bought. he was trying to build something people couldn't stop using - and refused every offer to hand it to someone else.
every buyout is a bet you're selling too cheap. the price is what someone smarter than you thinks it's worth today. it's the floor, never the ceiling.
the billion he turned down at 22 is now a rounding error on what that "no" was actually worth ↓
the man who runs a $1.5 trillion company once had to personally explain why his database kept crashing. he was 19, and the site was called TheFacebook.
in 2004 he ran it off rented machines from his Harvard dorm, describing the whole thing like a hobby.
"when we first launched we were hoping for maybe 400 or 500 people. now we're at 100,000. who knows where we're going next... maybe we can make something cool."
no business plan. no exit strategy. just a coder solving one problem at a time - how to scale to the next university, how to stop the servers from falling over, how to keep people coming back.
that "something cool" now serves over 3 billion people.
but here's what almost nobody in his position does.
he never sold it. Microsoft and AOL reportedly offered him millions for a program he built in high school - he said no. Yahoo offered $1 billion for Facebook when he was 22, with his entire board telling him to take it - he said no. most of his senior team quit within a year.
"I don't really like putting a price-tag on the stuff I do. that's just not the point."
the people who build the biggest things aren't optimizing for the exit. Zuckerberg wasn't trying to get bought. he was trying to build something people couldn't stop using - and refused every offer to hand it to someone else.
every buyout is a bet you're selling too cheap. the price is what someone smarter than you thinks it's worth today. it's the floor, never the ceiling.
the billion he turned down at 22 is now a rounding error on what that "no" was actually worth ↓
@itsak1to long-term mindset is what really stood out here
Elon Musk on why the US loses AI to China without space-based data centers: "The constraint outside China is electricity. Inside China, it's chips."
The framing is more precise than most of this debate gets.
His argument: Chinese labs are already competitive on relatively little compute. Give them abundant compute and they likely lead. And China already generates more electricity than the US, Europe and India combined.
Export controls limit which chips China can buy. They don't limit which models the rest of the world can use. That distinction matters more than the policy conversation acknowledges.
The lithography point is the real signal. If China solves domestic chip manufacturing at volume, the chip constraint disappears entirely and the electricity advantage becomes decisive.
Space-based data centers solve the American side of that equation - unlimited solar, no grid permitting, no local opposition.
But it moves the bottleneck rather than removing it. Musk says this openly: solve power outside China and the constraint returns to chips. The competition doesn't end. It just relocates.
Every reframe of this race lands on the same conclusion. Whoever removes their binding constraint first sets the pace for the next cycle.
Right now both sides have a different constraint and both are working on it.
https://t.co/dpc9MAVtEZ
MIT put its full multivariable calculus course online in 2007 and charged nobody a dollar for it.
Lecture 34 is the final review. A professor in an orange sweater writes Unit 1 on a chalkboard and works through an entire semester. The camera barely moves. No editing, no music, no visible production budget.
OpenCourseWare launched in 2001 with 50 courses and a plan to publish materials from every course MIT teaches.
Faculty were told upfront: no royalties, no additional pay. Most agreed anyway.
The catalog passed 2,400 courses. Downloads run past 300 million. Millions of visitors a year, many from countries with no equivalent institution anywhere nearby.
Universities charging $60,000 a year watched a peer give away the lectures and kept charging the same amount.
That's the part worth sitting with. If the content were the product, tuition should have collapsed. It didn't move.
What you pay $60,000 for is the credential, the network, the forcing function of deadlines, and the people sitting next to you.
The lecture was always the cheapest input in the entire system.
MIT proved it by giving it away and losing nothing.
MIT put its full multivariable calculus course online in 2007 and charged nobody a dollar for it.
Lecture 34 is the final review. A professor in an orange sweater writes Unit 1 on a chalkboard and works through an entire semester. The camera barely moves. No editing, no music, no visible production budget.
OpenCourseWare launched in 2001 with 50 courses and a plan to publish materials from every course MIT teaches.
Faculty were told upfront: no royalties, no additional pay. Most agreed anyway.
The catalog passed 2,400 courses. Downloads run past 300 million. Millions of visitors a year, many from countries with no equivalent institution anywhere nearby.
Universities charging $60,000 a year watched a peer give away the lectures and kept charging the same amount.
That's the part worth sitting with. If the content were the product, tuition should have collapsed. It didn't move.
What you pay $60,000 for is the credential, the network, the forcing function of deadlines, and the people sitting next to you.
The lecture was always the cheapest input in the entire system.
MIT proved it by giving it away and losing nothing.
https://t.co/aqsY18NvVz
Jay Leno says he made $30,000,000 a year hosting The Tonight Show and never spent a single dollar of it.
The structure behind that is the interesting part.
He lived entirely on his stand-up income. The NBC money went untouched. Not invested cleverly, not managed by a wealth advisor - just never spent.
Which is what made the next decision possible.
When NBC said they'd have to cut half the crew, Leno took a 50% pay cut instead. His reasoning: "I already have $15 million. I'm doing pretty good."
That's only a reasonable option if you've spent decades not needing the money you were making.
The line underneath it all: "I never thought TV was a job that would last your whole life."
He treated the biggest paycheck of his career as temporary from day one. Most people do the opposite - they get the big contract and immediately build a life that requires it to continue.
Leno kept his lifestyle pinned to the income he trusted, and let the fragile income accumulate untouched.
The financial discipline wasn't the achievement. The achievement was refusing to believe the good years were permanent while he was living in them.
That's what bought him the freedom to protect his crew instead of his salary.
Jay Leno says he made $30,000,000 a year hosting The Tonight Show and never spent a single dollar of it.
The structure behind that is the interesting part.
He lived entirely on his stand-up income. The NBC money went untouched. Not invested cleverly, not managed by a wealth advisor - just never spent.
Which is what made the next decision possible.
When NBC said they'd have to cut half the crew, Leno took a 50% pay cut instead. His reasoning: "I already have $15 million. I'm doing pretty good."
That's only a reasonable option if you've spent decades not needing the money you were making.
The line underneath it all: "I never thought TV was a job that would last your whole life."
He treated the biggest paycheck of his career as temporary from day one. Most people do the opposite - they get the big contract and immediately build a life that requires it to continue.
Leno kept his lifestyle pinned to the income he trusted, and let the fragile income accumulate untouched.
The financial discipline wasn't the achievement. The achievement was refusing to believe the good years were permanent while he was living in them.
That's what bought him the freedom to protect his crew instead of his salary.
https://t.co/czKVk7dRVw
Last Seen Users on Sotwe
Trends for you
Most Popular Users
Elon Musk
@elonmusk
241.1M followers
Barack Obama
@barackobama
119.1M followers
Cristiano Ronaldo
@cristiano
112.3M followers
Donald J. Trump
@realdonaldtrump
111.8M followers
Narendra Modi
@narendramodi
107.1M followers
Rihanna
@rihanna
98.1M followers
NASA
@nasa
92.2M followers
Justin Bieber
@justinbieber
91.3M followers
KATY PERRY
@katyperry
88.6M followers
Taylor Swift
@taylorswift13
82.5M followers
Lady Gaga
@ladygaga
74M followers
Virat Kohli
@imvkohli
71.4M followers
Kim Kardashian
@kimkardashian
70.2M followers
YouTube
@youtube
68.7M followers
Bill Gates
@billgates
64.4M followers
Neymar Jr
@neymarjr
64.3M followers
The Ellen Show
@theellenshow
62.4M followers
CNN
@cnn
61.8M followers
Selena Gomez
@selenagomez
61.7M followers
X
@x
60.8M followers