Latest / Connor with Honor: AI, Real Estate and Real Life / The 20-Year Problem Solved in 24 Minutes: The End of Knowledge Work?

Transcript

The 20-Year Problem Solved in 24 Minutes: The End of Knowledge Work? · Connor with Honor: AI, Real Estate and Real Life

• 0:00
  SPEAKER_00: A mathematician spent 20 years of his life on one problem and AI solved it in 24 minutes.
• 0:05
  SPEAKER_00: And the mathematician himself said, My life will never be the same.
• 0:08
  SPEAKER_00: We need to talk about what happened today.
• 0:10
  SPEAKER_00: So, welcome back, March 6, 2026.
• 0:13
  SPEAKER_00: I'm Connor, AI with honor.
• 0:15
  SPEAKER_00: Every day I go find the smartest people on the planet covering where AI is actually going, and I translate it for people who have real lives, real jobs, and real things at stake.
• 0:24
  SPEAKER_00: Today's source is Dr.
• 0:25
  SPEAKER_00: Alec Wis Alex Wisner-Gross.
• 0:27
  SPEAKER_00: He's an MIT physicist, one of the clearest thinkers I've found on this stuff.
• 0:31
  SPEAKER_00: His newsletter is called the Intermost Loop.
• 0:33
  SPEAKER_00: Link is in the description.
• 0:35
  SPEAKER_00: I read it, I break it down, so I get it.
• 0:37
  SPEAKER_00: So let's go.
• 0:38
  SPEAKER_00: The AI that's replacing the professional, OpenAI just released two new models, GPT 5.4 Thinking and GPT 5.4 Pro.
• 0:46
  SPEAKER_00: There's a test out there called the GDP Valve.
• 0:49
  SPEAKER_00: Think of it like the ultimate professional competency exam, except instead of testing one profession, it tests all of them at once.
• 0:57
  SPEAKER_00: Law, accounting, engineering, research, analysis, writing, everything a knowledge worker gets paid to do.
• 1:03
  SPEAKER_00: GPT 5.4 scored at 83%.
• 1:07
  SPEAKER_00: It's like hiring one contractor who does the work of a plumber, electrician, carpenter, and accountant, gets it right 83 times out of 100, never calls in sick, and doesn't negotiate salary.
• 1:17
  SPEAKER_00: And here's the thing the other 17%, the gap closes with every new model.
• 1:23
  SPEAKER_00: We're probably not years away from 90%, probably months away.
• 1:28
  SPEAKER_00: It also scored 57.7 on a software engineering benchmark, and that's something else that's absolutely incredible.
• 1:36
  SPEAKER_00: Now, on a web navigation test, and also it built a fully working Minecraft clone from scratch in 24 minutes.
• 1:44
  SPEAKER_00: That's the first artificial intelligence to ever do it.
• 1:47
  SPEAKER_00: It's like telling somebody to build a fully functional house with no blueprints, no crew, and they hand you the keys before lunch.
• 1:54
  SPEAKER_00: What does this mean for your future?
• 1:56
  SPEAKER_00: If you work in any professional field where you sit at a desk and produce knowledge, reports, contracts, code, analysis, strategy, the clock is ticking.
• 2:06
  SPEAKER_00: Not on your career, on your ability to use this tool better than the person next to you.
• 2:11
  SPEAKER_00: The ones who figure that out first are going to look like superstars.
• 2:15
  SPEAKER_00: The ones who ignore it or think it's ridiculous are going to be confused about why they got passed over.
• 2:22
  SPEAKER_00: They're inside that window right now.
• 2:23
  SPEAKER_00: We're getting ahead of this still matters.
• 2:25
  SPEAKER_00: The window, though, won't stay open forever.
• 2:28
  SPEAKER_00: Let's look at story number two, the mathematician's 20 years.
• 2:31
  SPEAKER_00: Now, this one's going to be different.
• 2:33
  SPEAKER_00: I tell you about a man named Bartosz Nasreki, and I'm sorry if I'm messing that up.
• 2:38
  SPEAKER_00: He's a mathematician.
• 2:40
  SPEAKER_00: He spent 20 years, not 20 days, not 20 months, 20 years of his life working on a single problem.
• 2:46
  SPEAKER_00: A tier four frontier math problem.
• 2:49
  SPEAKER_00: That's the Mount Everest category of mathematics, the problems that the world's best minds have been staring at and getting nowhere on.
• 2:58
  SPEAKER_00: GPT 5.4 by OpenAI, Pro, just solved it.
• 3:02
  SPEAKER_00: Clean.
• 3:03
  SPEAKER_00: The mathematician read the solution.
• 3:05
  SPEAKER_00: He said it felt almost human.
• 3:06
  SPEAKER_00: He called it his personal move 37.
• 3:08
  SPEAKER_00: Here's what that means.
• 3:10
  SPEAKER_00: In 2016, an AI beat the world's greatest player at a game called Go.
• 3:14
  SPEAKER_00: It's one of the most complex strategy games ever invented, with more possible moves than there are atoms in the observable universe.
• 3:22
  SPEAKER_00: Ha.
• 3:23
  SPEAKER_00: And the AI made a move, move 37, that no human player would have ever made.
• 3:28
  SPEAKER_00: The commentators went silent.
• 3:30
  SPEAKER_00: The world champion stared at the board, probably had a cigarette.
• 3:33
  SPEAKER_00: Everyone watching knew something permanent had just happened.
• 3:37
  SPEAKER_00: The mathematician felt the same thing, looking at the solution to his problem, he said, My singularity has just arrived.
• 3:44
  SPEAKER_00: Life continues on the other side, off to infinity.
• 3:47
  SPEAKER_00: It's like you've spent 20 years trying to open a safe, every combination, every tool, every method, and somebody walks up, listens to it for 30 seconds, and opens it on the first try while you watch.
• 3:59
  SPEAKER_00: What does this mean for your future?
• 4:01
  SPEAKER_00: Mathematics is the language underneath everything: physics, medicine, engineering, biology, economics, climate.
• 4:08
  SPEAKER_00: Every hard problem humanity is trying to solve has a math problem at its core.
• 4:13
  SPEAKER_00: If AI is now cracking the ones that the world's best mathematicians couldn't crack in 20 years, the rate at which we solve cancer, build new materials, understand aging, fix broken infrastructure, that rate just got a rocket engine attached to it.
• 4:29
  SPEAKER_00: We're talking about potentially compressing decades of scientific progress into years, maybe less.
• 4:35
  SPEAKER_00: That's what used to take a generation, but you just start taking up an afternoon to get accomplished.
• 4:41
  SPEAKER_00: Story number three, Hollywood says yes.
• 4:43
  SPEAKER_00: Apple says, put your name on it.
• 4:45
  SPEAKER_00: Remember, Apple got uh one over Blockbuster.
• 4:48
  SPEAKER_00: You remember that?
• 4:48
  SPEAKER_00: It was a long time ago.
• 4:50
  SPEAKER_00: Netflix just bought Ben Afflex, AI company called Interpositive.
• 4:55
  SPEAKER_00: What does it do?
• 4:55
  SPEAKER_00: Well, it learns real from real film footage and the raw, unfinished stuff straight off to the set.
• 5:00
  SPEAKER_00: Then it handles the color, the lighting, the sound mixing, the visual effects, and post-production.
• 5:05
  SPEAKER_00: The work that used to take entire teams of specialist months to complete, Netflix said yes, full send.
• 5:12
  SPEAKER_00: They absorbed it completely.
• 5:13
  SPEAKER_00: It's like a restaurant owner who fires the kitchen staff and installs a machine that cooks everything perfectly every time at a fraction of the cost.
• 5:21
  SPEAKER_00: Some owners would never do it.
• 5:23
  SPEAKER_00: Netflix just did it.
• 5:24
  SPEAKER_00: Apple Music went the exact opposite direction.
• 5:27
  SPEAKER_00: They're launching transparency tags.
• 5:29
  SPEAKER_00: Labels that tell listeners whether AI helped make the artwork, the song, the music video.
• 5:34
  SPEAKER_00: It's like a food label that says whether your bread was made by hand or machine.
• 5:38
  SPEAKER_00: Some people aren't going to care.
• 5:40
  SPEAKER_00: Some people will only buy the handmade loaf.
• 5:43
  SPEAKER_00: Let the market decide.
• 5:44
  SPEAKER_00: What does this mean for your future?
• 5:46
  SPEAKER_00: Within five to ten years, the cost of making a professional quality film or album drops so dramatically that a person with a great idea on a laptop can produce something that looks and sounds like it came out of a major studio.
• 5:58
  SPEAKER_00: The gatekeepers lose their gates.
• 6:00
  SPEAKER_00: This is a massive disruption to Hollywood.
• 6:03
  SPEAKER_00: But it's a massive opportunity for anybody that has a story to tell who never had access to the tools before.
• 6:09
  SPEAKER_00: The question is, are you going to be the one telling stories or are you going to be the one that's being disrupted?
• 6:15
  SPEAKER_00: All right, story number four.
• 6:16
  SPEAKER_00: The servers are on the battlefield.
• 6:18
  SPEAKER_00: Here's one of the here's one that most people are sleeping on, and it might be the most important story of the day.
• 6:24
  SPEAKER_00: AI doesn't run on magic, it runs on chips and memory.
• 6:27
  SPEAKER_00: And right now the world doesn't have enough of either.
• 6:30
  SPEAKER_00: Apple quietly pulled together their highest end, pulled their highest end Mac Studio off the market, the 512 gigabyte RAM version.
• 6:40
  SPEAKER_00: It's gone.
• 6:40
  SPEAKER_00: Not a recall, not a redesign.
• 6:43
  SPEAKER_00: They can't get the memory chips.
• 6:45
  SPEAKER_00: The shortage is that bad.
• 6:47
  SPEAKER_00: It's like the best car manufacturer in the world, but you can't get steel.
• 6:51
  SPEAKER_00: The capability exists, the raw material doesn't.
• 6:55
  SPEAKER_00: Washington is now drafting rules requiring U.S.
• 6:58
  SPEAKER_00: government approval before NVIDIA or AMD chips can ship anywhere on Earth.
• 7:03
  SPEAKER_00: And the commerce is floating the idea, the commerce is floating the idea that if your country wants access to these chips, you better be investing money back into America first.
• 7:13
  SPEAKER_00: It's like owning the only water well in a desert town and starting to charge admission.
• 7:19
  SPEAKER_00: A Seattle startup called Clover Leaf Infrastructure just raised$300 million, seeing what they call powered land.
• 7:27
  SPEAKER_00: They're selling it.
• 7:28
  SPEAKER_00: A real estate that's pre-wired and ready for data centers.
• 7:31
  SPEAKER_00: Oracle is cutting thousands of jobs to redirect the money into building more data centers.
• 7:36
  SPEAKER_00: A nuclear company called TerraPower just got the first government permit ever issued for a commercial advanced nuclear reactor.
• 7:44
  SPEAKER_00: 345 megawatts in Wyoming to power the AI infrastructure that regular power grids can't handle.
• 7:51
  SPEAKER_00: It's like building a new power plant specifically because the existing electrical grid wasn't built for what's coming.
• 7:58
  SPEAKER_00: And then this Iranian state media confirmed that Amazon's data center in Bahrain was deliberately targeted and struck by Iran's Revolutionary Guard.
• 8:08
  SPEAKER_00: Because Amazon supports the U.S.
• 8:10
  SPEAKER_00: military, that's one of the first times in history that a data center was intentionally hit as an act of war.
• 8:16
  SPEAKER_00: That's likely to increase.
• 8:18
  SPEAKER_00: It's like bombing a supply depot, except the supply depot is a building full of servers.
• 8:24
  SPEAKER_00: And what's being denied isn't bullets or fuel, it's computing power.
• 8:28
  SPEAKER_00: What does this mean for your future and mine?
• 8:30
  SPEAKER_00: Well, the countries that control the chips control the future of AI.
• 8:34
  SPEAKER_00: Full stop.
• 8:35
  SPEAKER_00: This is shaping up to be the defining geopolitical competition for the next 50 years.
• 8:41
  SPEAKER_00: Except instead of oil fields, nations are fighting over semiconductor factories and data centers.
• 8:47
  SPEAKER_00: Inside a decade, access to AI compute may be as strategically important as access to oil was in the 20th century.
• 8:54
  SPEAKER_00: And data centers, which used to be invisible, boring infrastructure, now need norms security, physical and cyber.
• 9:01
  SPEAKER_00: This is an entirely new industry that barely exists yet.
• 9:05
  SPEAKER_00: Number five, your job, the economy and the government playing catch-up.
• 9:09
  SPEAKER_00: A University of Chicago economists just confirmed what a lot of people have suspected.
• 9:14
  SPEAKER_00: AI's impact on productivity is now big enough to show up in the national economic data, not just tech companies and the whole economy.
• 9:22
  SPEAKER_00: It's like when electricity was invented.
• 9:24
  SPEAKER_00: At first, it just made individual factories more efficient.
• 9:28
  SPEAKER_00: Then one day, economists looked up and realized it had restructured the entire industrial world.
• 9:34
  SPEAKER_00: Here's the part that surprised me.
• 9:35
  SPEAKER_00: Software engineer job postings are going up, not down.
• 9:39
  SPEAKER_00: Up.
• 9:40
  SPEAKER_00: You think AI writing code would put coders out of work.
• 9:43
  SPEAKER_00: Instead, AI makes software cheaper to build, more software gets built, which creates a demand for more engineers.
• 9:50
  SPEAKER_00: It's called Jevons Paradox.
• 9:52
  SPEAKER_00: It's like when highways were invented, people thought fewer roads meant less travel.
• 9:57
  SPEAKER_00: Instead, more roads created more driving, more trucking, more commerce, more everything.
• 10:01
  SPEAKER_00: Efficiency created more demand, not less.
• 10:04
  SPEAKER_00: Anthropic, the company that makes Claude AI, launched an early warning system to track when jobs are being replaced by AI versus the ones which are just being assisted by it.
• 10:15
  SPEAKER_00: They said the evidence of mass job loss is limited so far.
• 10:18
  SPEAKER_00: Pay attention to those two words, so far.
• 10:22
  SPEAKER_00: New York is considering banning AI chatbots, of course, from giving legal or medical advice.
• 10:27
  SPEAKER_00: Indiana just signed a law requiring retirement accounts to include cryptocurrency options by 2027.
• 10:34
  SPEAKER_00: It's like local governments trying to post speed limit signs on a highway that's already doing 200 miles an hour.
• 10:40
  SPEAKER_00: They're not wrong to try.
• 10:41
  SPEAKER_00: They're just late.
• 10:42
  SPEAKER_00: What does this mean for your future?
• 10:44
  SPEAKER_00: Well, the jobs more protected from what's coming are the ones that require physical presence, human judgment, real accountability, trades, healthcare, leadership, sales, those jobs most at risk are the repeatable deskbound ones where the output is a document and a spreadsheet or even a recommendation.
• 11:02
  SPEAKER_00: The middle of the income ladder is where the pressure is the hardest.
• 11:06
  SPEAKER_00: Knowing that right now, while you still have time to reposition, is the most valuable thing I can give you today.
• 11:13
  SPEAKER_00: Story number six, the Pentagon, the prediction market, the price of progress.
• 11:18
  SPEAKER_00: The Pentagon officially classified Anthropic, the Claude AI company, as a supply chain risk.
• 11:25
  SPEAKER_00: This could force every military contractor in America to stop using Claude in their systems.
• 11:30
  SPEAKER_00: It's like the government putting a supplier on a watch list, not because they broke a law, but because the relationship got complicated and the stakes are too high to leave it unmanaged.
• 11:42
  SPEAKER_00: Anthropic CEO Dario Amade said that they've been productive.
• 11:47
  SPEAKER_00: They've been having productive conversations with the DoD and also apologized for an internal memo that leaked, one that apparently took shots at the White House and at Sam Altman for OpenAI.
• 11:59
  SPEAKER_00: Apology.
• 12:00
  SPEAKER_00: Anyway, that's not just drama.
• 12:02
  SPEAKER_00: That's two of the most powerful AI companies in the world getting pulled into U.S.
• 12:07
  SPEAKER_00: military and political strategy, whether they plan to be there or not.
• 12:11
  SPEAKER_00: And then this someone on a prediction market called Polymarket, believed to be a military insider, correctedly predicted U.S.
• 12:20
  SPEAKER_00: strikes on Venezuela, Iran, and Iraq before they happen.
• 12:24
  SPEAKER_00: He won$90,000.
• 12:26
  SPEAKER_00: Now they're betting American forces enter Iran before the end of the year.
• 12:30
  SPEAKER_00: It's like somebody at a poker table who keeps winning every hand, and you start wondering if they can see the cards.
• 12:37
  SPEAKER_00: What does this mean for your future?
• 12:39
  SPEAKER_00: Well, AI is already a military tool.
• 12:41
  SPEAKER_00: Not theoretically, but actually.
• 12:43
  SPEAKER_00: The question isn't whether governments are going to use AI in the conflict, they already are.
• 12:47
  SPEAKER_00: The question is who controls the infrastructure underneath and what happens when the weapon is a system that can plan, adapt, and operate faster than any human chain of command.
• 12:57
  SPEAKER_00: The next major conflict, wherever it happens, will have an AI layer that most civilians will never see.
• 13:04
  SPEAKER_00: This is not a prediction.
• 13:05
  SPEAKER_00: This is the trajectory of everything we watched happen today.
• 13:10
  SPEAKER_00: Here's where I land at the end of all this.
• 13:12
  SPEAKER_00: I spent 20 years in law enforcement, 23 actually, if you count the Barney Five Sheriff's Departments I was with.
• 13:19
  SPEAKER_00: I've seen what happens when people don't see change coming until it's already on top of them.
• 13:25
  SPEAKER_00: The ones who paid attention early, who read the room, adjusted, and moved before they had to, those people are always better off than the ones who waited to be told.
• 13:35
  SPEAKER_00: That's exactly where we are right now with artificial intelligence.
• 13:38
  SPEAKER_00: This isn't coming.
• 13:40
  SPEAKER_00: This is here today.
• 13:41
  SPEAKER_00: The mathematicians 20 years, the data center getting bombed, the chips running out, the job market shifting in real time.
• 13:50
  SPEAKER_00: The only question that matters is the same one I ask myself every morning.
• 13:53
  SPEAKER_00: When I'm in the gym at zero dark thirty, am I doing the work to be ready for what's next, or am I hoping this works out?
• 14:00
  SPEAKER_00: Hope is not a strategy, especially right now.