Good morning everybody.
It’s Labor Day, which apparently I’m celebrating by doing more labor.
Bitcoin is still stuck below $80,000, and at this point we need a catalyst strong enough to get us through that level and actually keep us there. But Bitcoin isn’t really the biggest story this morning. Over the weekend, we got a surprisingly strong U.S. jobs report, another escalation between the United States and Iran, oil pushing toward $100, the release of GPT-6 Astra, and another reminder that quantum computing is moving forward whether crypto is ready for it or not.
There’s a lot happening at once, and I’m trying to figure out how these pieces eventually fit together.
Jobs, Iran and $97 Oil Complicate the Bitcoin Picture
The United States added 162,000 jobs in August, while unemployment remained at 4.1%, according to the Bureau of Labor Statistics. That’s substantially stronger than the roughly 55,000 jobs economists had expected.
One of the strangest details in the report was the gender split. Women accounted for approximately 158,000 of the net employment gain, compared with only about 4,000 for men. That’s an extraordinary one-month number, although economists have cautioned that these monthly gender figures can be noisy and shouldn’t automatically be treated as evidence of a permanent structural shift.
Still, I want to understand what’s happening there.
What industries are creating these jobs? Are they concentrated in traditionally female-dominated sectors? Are male-dominated industries shedding workers at the same time? One month isn’t enough to answer those questions, but the disparity is large enough that it’s worth investigating.
The bigger immediate issue for Bitcoin is what this means for rates.
A 4.1% unemployment rate is low. Combine that with inflation that remains above target and oil prices rising again, and the Fed doesn’t have an obvious reason to rush toward easier monetary policy. Markets are now putting the probability of a September rate increase around 58%.
Then we have Iran.
There was a lot of conflicting information circulating this weekend, so here’s what we can actually verify. U.S. Central Command says Iran launched ballistic missiles toward a U.S. aircraft carrier and guided-missile destroyer. According to CENTCOM, the ships successfully evaded those attacks and were not hit. The United States subsequently struck three Iranian crude carriers, permanently disabling two and destroying a third.
That’s important because some reports and social-media posts claimed Iran successfully hit the American warships. I mentioned that possibility on the podcast, but the available U.S. account says otherwise, and independent reporting also describes the Iranian missile attack as unsuccessful.
The economic consequence is easier to see.
Brent crude is now around $97 a barrel, while WTI is around $92 as markets price additional risk around Middle Eastern oil supplies and the Strait of Hormuz.
That’s another inflation problem.
If oil breaks $100 and stays there, the Fed’s job gets more complicated. And if markets respond by pricing tighter monetary policy for longer, that’s not necessarily the catalyst Bitcoin needs to finally establish $80K as support.
GPT-6 Astra Changes What an AI Agent Can Actually Do
Then we got what I think may ultimately be the much bigger story.
OpenAI officially released GPT-6 Astra on September 3, and this isn’t simply another model that writes slightly better emails.
Astra is built around computer use, browsing, software engineering, cybersecurity, science and complex professional work. OpenAI specifically describes it as capable of handling demanding computer and browser tasks, while its release notes emphasize coding, research, computer use and multi-step workflows.
That’s the difference I care about.
Previously, a lot of automation depended on knowing exactly where something was. The system needed to know where the login box was, where the button was and what sequence of actions to perform.
Increasingly, the model can understand what it’s looking at and operate software more like a person does.
For crypto, start thinking through what that eventually means.
An AI agent could interact with a wallet.
It could manage a trading account.
It could move stablecoins.
It could interact with DeFi positions.
It could participate in DAO governance.
It could deploy and monitor smart contracts.
And because the model can navigate changing interfaces rather than relying exclusively on a predetermined sequence of clicks, the number of financial systems an agent can potentially interact with gets much larger.
I’m not saying you should hand Astra your private keys tomorrow morning.
I’m saying we’re getting closer to a world where autonomous software can actually operate financial infrastructure rather than simply tell you what button to press.
And there is another side of this.
OpenAI says Astra is its first model to reach the company’s Critical cybersecurity capability threshold. According to OpenAI, with appropriate tools and access, Astra can find previously unknown security vulnerabilities and develop new exploitation methods against well-protected systems without requiring a human to guide every individual step.
That’s extremely relevant to crypto.
We have enormous amounts of money sitting behind smart contracts, bridges, exchanges and wallets. If offensive AI agents become dramatically better, defensive AI agents are going to have to become dramatically better too.
I think we’re heading toward adaptive security fighting adaptive attacks.
Agent versus agent.
IBM’s Quantum Roadmap Is Real, But There’s an Important Correction
The other technological development I talked about today is quantum computing.
And I want to make one correction to what I said on the podcast.
IBM’s current roadmap does not say Starling will have 256 logical qubits. IBM says it plans to deliver 200 logical qubits capable of running 100 million quantum gates in 2029. Before then, its roadmap calls for systems scaling to as many as 1,080 physical qubits as it develops the architecture required for fault tolerance.
That’s still a big deal.
IBM also recently announced that it successfully connected and cooled two modular cryogenic systems, an engineering step toward linking the hundreds of quantum chips that a future fault-tolerant system will require.
But there’s another important correction.
Having 200 logical qubits does not mean the machine can simply represent every atom in the observable universe and therefore instantly brute-force a Bitcoin private key.
A quantum system with n qubits mathematically has a state space involving 2^n amplitudes, but that does not mean you can independently read out all 2^n possibilities or automatically solve arbitrary computational problems in one operation.
Quantum computers gain advantages through specific algorithms that manipulate those states in useful ways.
That’s why the actual cryptographic threat is more complicated than saying “256 qubits equals 2^256 possibilities, therefore Bitcoin is broken.”
But my underlying point remains the same.
Crypto should prepare before quantum computers become a practical threat.
Bitcoin doesn’t have to be breakable today for developers to start thinking about migration paths, vulnerable public keys and post-quantum cryptography. If IBM actually delivers a fault-tolerant 200-logical-qubit machine in 2029 and continues toward the 2,000-logical-qubit Blue Jay system it has planned beyond that, this stops being something the crypto industry can dismiss as science fiction.
Technology often feels like hype right up until the moment it works.
Are We Already at AGI?
This is where I’m going to separate documented fact from my opinion.
OpenAI calls Astra its most capable model and reports extraordinary performance across computer use, coding, cybersecurity and science.
My opinion is that we’re already entering what I would call AGI.
I understand that people use different definitions of AGI. Under some definitions, Astra won’t qualify. Under others, systems capable of reasoning across domains, writing software, conducting research, navigating computers and completing complicated professional workflows are getting awfully close to what people spent decades describing.
I also think the models available publicly may not tell us everything about the capabilities companies have internally.
That’s an inference, not something I can prove.
But what we can see publicly is already enough to change the conversation.
And this is where AI, quantum computing and crypto start running into each other.
AI agents are becoming capable of operating computers and financial systems.
AI is becoming powerful enough that cybersecurity itself has to become more adaptive.
Quantum computing is moving toward fault-tolerant machines.
Crypto is trying to build a global financial system based heavily on cryptography, open-source software and autonomous smart contracts.
Then layer on war, $97 oil, inflation, interest rates and geopolitical instability.
That’s why I woke up this weekend looking at all of this and wondering how the hell we’re supposed to piece it together.
I don’t have a clean answer yet.
But I don’t think you can understand where crypto is going anymore by looking exclusively at a Bitcoin chart.
Crypto Prices
Bitcoin is sitting around $79,392, up approximately 2% over 24 hours.
Ethereum is approximately $2,500, roughly flat.
BNB is around $745.
XRP is approximately $1.40, down around 1% over 24 hours but up roughly 3% over seven days.
Solana is around $105, down approximately 1.4% over 24 hours and up around 3% over seven days.
TRON is approximately $0.335, essentially flat.
Hyperliquid is around $87.61, down roughly 2% over 24 hours but up approximately 8.5% over seven days.
Zcash continues its run at approximately $1,182, roughly flat over 24 hours but up an incredible 45% over seven days.
Litecoin is around $57.62, up approximately 6% over 24 hours, 20% over seven days and 27% over the past month.
Total Crypto Market Cap: approximately $2.7 trillion.
Fear & Greed: 74, Greed.
Bitcoin still hasn’t solved $80,000.
But today I care more about what’s happening around Bitcoin.
A strong labor market and $97 oil are complicating the Fed’s decision. The United States and Iran are escalating again around one of the most important oil corridors in the world. AI systems are moving from answering questions toward actually operating computers and completing work. And IBM is publicly targeting a fault-tolerant quantum computer by 2029.
I don’t know exactly how all of those pieces fit together yet.
That’s kind of the point.
These technologies and economic systems are starting to collide faster than anybody can neatly put them into separate boxes. Crypto is sitting right in the middle of that collision.