Deficits, Curves, and Bitcoin: what a 2 trillion dollar hole, an AI solving a 30 year cryptography mystery, and Bitcoin holding near 81K mean for people starting from zero

The 2 trillion dollar hole

The US Treasury dropped a number on Friday that should get your attention. The federal deficit for fiscal year 2026 came in at 1.992 trillion dollars. That is up 12 percent from 1.775 trillion the year before. It is the third largest deficit in American history, behind only the two pandemic years. This is now the 25th straight year the government has spent more than it collects. The last surplus was in 2001.

Here is the detail that matters. Receipts did not fall. They hit a record 5.40 trillion dollars. Outlays just grew faster, to a record 7.40 trillion dollars. Net interest on the debt rose to roughly 1.36 trillion dollars. That means the government now spends more on interest than on most of what it actually does. Social Security cost nearly 1.7 trillion. Medicare came in near 1.1 trillion. Bond traders sent 10 year and 30 year Treasury yields to 24 year highs. Total public debt is past 40 trillion dollars.

What does a 2 trillion dollar deficit mean for someone starting from zero?

It means the biggest institutions on earth cannot balance a budget, and you still have to balance yours. There is no rescue coming from Washington. The machine runs on debt, and the people running it are fine with that. Your job is to not copy them. Debt is a tool that makes rich people richer and breaks everyone else. Learn the difference between leverage and drowning. Build skills that pay regardless of what the Treasury does.

Can you name one person whose life got better because the deficit grew?

Claude solves a 30 year cryptography mystery in 30 minutes

Now the story I am most excited about this week. On October 16, Anthropic published research showing that Claude solved one of the longest running mysteries in cryptography. In under 30 minutes.

Here is the background. Almost every secure connection on the internet runs on something called elliptic curve cryptography. When you log into your bank or open Substack, the lock in your browser comes from math published by NIST, the National Institute of Standards and Technology, back in the late 1990s. Those published standards include a set of curves called the P curves: P-192, P-224, P-256, P-384, and P-521. Each one has a seed value printed in the standard. For nearly 30 years, nobody could explain where those seeds came from. They did not match any published generation method. Some people suspected a backdoor. The NSA had a hand in the process, and the paranoia never fully went away.

Anthropic gave Claude to its cryptography team. Claude wrote a scanner that slid along each published seed one bit at a time and tested whether the bits were consecutive outputs of an old random number generator. It tried 34 combinations. Then it spotted a pattern: the top bits were all zero across four of the seeds, something that would happen by chance with probability 1 in 5810. That is the fingerprint of a 1970s C library function called lrand48, which only returns the top 31 bits of a 48 bit state. From there, Claude used a classic lattice algorithm called LLL to recover the generator's internal state. Every recovered state regenerated its seed exactly. And every one of those generators was seeded with a plain Unix timestamp. The P-256 seed, which protects a huge share of the world's encrypted traffic, was generated on August 16, 1997, at 1:52 AM, after 1842 tries.

No backdoor. Just a weak random number generator and a clock. The mystery that fueled 30 years of conspiracy theories died in 30 minutes of agent work.

The people who win from here are the ones who ask clear questions and let machines do the patient work.

What should you learn from this?

This is what AI agents are actually good at: patient, systematic search work that no human would finish. Claude did not have a flash of genius. It wrote a scanner, ran 34 combinations, noticed a statistical anomaly, and followed it. The technology that matters is the agent loop: the model writes code, runs it, looks at the result, and tries again. That loop is now good enough to answer a 30 year open question in cryptography before lunch.

What mystery in your field could a patient machine solve in 30 minutes?

One thing to do this weekend

Here is one concrete thing you can do this weekend. Pick one small open source project you care about. Give a coding agent one clear, narrow task: find every place the code calls a random number generator and tell me whether it is cryptographically secure. You will learn two skills at once: how agents actually work, and how real security audits start. Both are billable skills. No course required.

Bitcoin holds near 81K

Now the markets. Bitcoin is trading at 81,454 dollars, down about half a percent in the last 24 hours. Ethereum is at 2,423.94 dollars, down about 1.4 percent. The total crypto market cap is about 2.75 trillion dollars, and Bitcoin holds 59.4 percent of it.

Read that Bitcoin price against the deficit story above. 81,454 dollars for an asset with a fixed supply of 21 million coins, while the government burns 1.992 trillion in a single year and pays 1.36 trillion in interest. Nobody voted on the 40 trillion dollar debt. Nobody can vote on it. But anybody with a phone can hold an asset no committee can print. That contrast is the entire thesis. Prices move day to day. The thesis does not.

The uncomfortable truth

Nobody is coming to fix your money for you. The government cannot fix its own. The mystery got solved by a machine because a human asked a clear question and let it work. The people who win from here are the ones who ask clear questions and build real skills: how money works, how cryptography works, how agents work. Everything else is noise.

Start this weekend. One small project. One agent task. One skill you can prove. Do that 50 times and you are not starting from zero anymore.

---

Join my free Skool community: https://www.skool.com/henry-jenkins-5224

My website: https://henryljenkins.com

LightofLife.global

Twitter @Nevaquit

YouTube @HenryJenkins3