justin

Bitcoin. Libertarian. Psychedelic therapy. ῥίζα πάντων τῶν κακῶν τὸ κίβδηλον νόμισμα
nostr: #npub18rg7xzr6ejxwg4lupcckxslw6ae4krfv7zudcs33drdqnpjlcklqsfall2

[justin @sharpone74]

While the concept of value is fundamentally intrinsic to our subjective experiences, no referent can truly possess value independent of a valuer. Paradoxically, certain referents emerge as so universally prized across human societies that they practically function as if they hold inherent worth.

Such is Bitcoin’s inevitable path. Substantial rewards accrue to those who deduce, recognize and adopt the former proposition before the masses.

[justin @sharpone74]

@CorySwan @illbackman It sometimes makes you wonder how someone like this became a billionaire in the first place. Seems he’s wrong more often than not

[justin @sharpone74]

@bensig The whole thing was a display of Dunning-Kruger which is why I posted so

[justin @sharpone74]

The OP was in response to comments Chamath made in response to a supposed AI threat on math. This thread took it to the quantum threat. It’s not the same thing. The AI on math threat has a significantly lower probability of being real than the quantum threat. The attack vectors are completely different.

[justin @sharpone74]

@SatoshiDropsHQ @BitcoinArchive Yeah the podcast was actually about threat to break cryptography based on AI based math discoveries. Not so much about quantum.

[justin @sharpone74]

@BitcoinArchive I tried https://t.co/RwSotdy67G

[justin @sharpone74]

@chamath "Break SHA256 and drain Satoshi's wallet"

You should adopt your proper name: Chamath Dunning-Kruger Palihapitiya.

1) There are many block reward P2PK addresses associated to Satoshi, not "a wallet".
2) It is ECDSA, a use of the ECDLP which protects 'wallets', not SHA256. SHA256 is used as the proof-of-work mechanism, to add new transactions to the timechain.
3) AI is no where close to "solving" the ECDLP. Go ahead an ask an LLM yourself why this is and why even an advanced 'math' model is unlikely to change it.
4) Shor's is the only "real" threat. It's 2026 and nothing above the number 15 has ever been factored with it. Quantum computers at such scale are at least 10 years away. https://t.co/vHr2w7IWcj

[justin @sharpone74]

@LorenHodl @peruvian_bull isn’t 2% roughly a holdover from the annualized increase in gold production?

[justin @sharpone74]

@bergealex4 Yes. You will NEVER get back the bitcoin you spend on this.

[Ben Sigman @bensig]

Quantum computing announcements love big numbers.

“40x fewer errors.”
“17x more efficient.”
“Record fidelity.”

All real numbers. All impressive. But what do they actually mean for the timeline to break Bitcoin’s cryptography?

Take the recent magic-state breakthrough. They demonstrated a technique on real hardware that reduces errors in the special quantum states needed for useful computation. That’s real progress.

But dig into the numbers.

Only about 8% of attempts succeeded. The error rates are still roughly a million times too high for a full ECDSA attack. And here’s the kicker: this technique was already assumed in the resource estimates published earlier this year.

So the experiment validates those estimates. It doesn’t suddenly accelerate the timeline by 40x.

Same story with the “17x more efficient” circuits.

AI optimized an elliptic-curve point-addition circuit and dramatically reduced the gate count. Great work.

But when you model an actual attack across different quantum architectures, the physical qubit savings can be closer to 0–30%.

Not 17x.

There’s a pattern here.

Quantum companies need to demonstrate progress. Investors want breakthroughs. Headlines love multipliers.

But improving one component of a quantum algorithm by 40x doesn’t make the entire machine 40x closer to breaking Bitcoin.

We’ve seen resource estimates for breaking ECDSA drop dramatically, from roughly 2,300 logical qubits to 1,175 and now potentially around 900.

That’s significant.

But logical qubits aren’t physical qubits. And a machine capable of sustaining the necessary computation at sufficiently low error rates is still a very different beast from what exists today.

The remaining gap is measured in orders of magnitude.

That’s why serious projections still tend to put cryptographically relevant quantum computing in the 2030s, not next year.

I’m not dismissing quantum progress. Quite the opposite. The engineering is fascinating and some of these advances are genuinely important.

But I think we’re getting a little too excited about multipliers applied to individual components of systems that don’t exist yet.

When I see a quantum breakthrough, I want to know one thing:

How much did this actually change the end-to-end cost of breaking Bitcoin’s ECDSA?

Not how much better one circuit got.

Not how many times fewer errors one experiment produced.

How much closer are we to a machine that can actually execute the attack?

That’s the number that matters.

[justin @sharpone74]

I'm as confident as can be. The incentives are aligned. People mistake a non-upgraded for no-action. Plenty of action is happening and will happen. This is a big upgrade and needs careful consideration, urgency from necessity will be the driver. Right now there is none. QC is unlikely to sneak up on us. As Ben points out, QC ventures need continued funding, so any advances are widely published and spun in the best way possible.

In 2001, Shor's was used to factor the number 15. It is 25 years later and no quantum computer has yet truly factored the number 21 (despite what you may have read). This is the canary to watch, not marginal error rate reduction reports.

https://t.co/vHr2w7IWcj

[justin @sharpone74]

@BitBallr @SoloSatoshi @heatbit_com Cancel while you can

[justin @sharpone74]

It's worse than that. I listened to the entire episode.

First they spent about half of the episode saying that currency as we know it will be obsolete and only tokenized energy will matter. Then proceeded to throw out options for this; all of which would require an oracle (e.g tokenized compute) while ignoring that obvious fact: bitcoin is already tied directly energy sans oracle. Bitcoin = naturally tokenized energy.

Then near the end they indicated that LLM's were going to break the math behind cryptography anyway and "crypto" was doomed. @chamath literally said bitcoin would be hacked and then go to zero. ROFL. Quantum was hardly mentioned.

More here: https://t.co/K8OnYhGGo0

[justin @sharpone74]

@chamath "Break SHA256 and drain Satoshi's wallet"

You should adopt your proper name: Chamath Dunning-Kruger Palihapitiya.

1) There are many block reward P2PK addresses associated to Satoshi, not "a wallet".
2) It is ECDSA, a use of the ECDLP which protects 'wallets', not SHA256. SHA256 is used as the proof-of-work mechanism, to add new transactions to the timechain.
3) AI is no where close to "solving" the ECDLP. Go ahead an ask an LLM yourself why this is and why even an advanced 'math' model is unlikely to change it.
4) Shor's is the only "real" threat. It's 2026 and nothing above the number 15 has ever been factored with it. Quantum computers at such scale are at least 10 years away. https://t.co/vHr2w7IWcj

[justin @sharpone74]

@vitorfidm @dotkrueger A bond requires repayment of the principal at terms end, an equity (which STRC is) does not. This is difference #1

[justin @sharpone74]

@schulzzy Imagine: Shor's capable quantum computers and mercury -> gold machines released in the same year, 10 years from now.

[justin @sharpone74]

@matteopelleg @cluborange @chamath https://t.co/7YrubFQpOB

[justin @sharpone74]

@chamath "Break SHA256 and drain Satoshi's wallet"

You should adopt your proper name: Chamath Dunning-Kruger Palihapitiya.

1) There are many block reward P2PK addresses associated to Satoshi, not "a wallet".
2) It is ECDSA, a use of the ECDLP which protects 'wallets', not SHA256. SHA256 is used as the proof-of-work mechanism, to add new transactions to the timechain.
3) AI is no where close to "solving" the ECDLP. Go ahead an ask an LLM yourself why this is and why even an advanced 'math' model is unlikely to change it.
4) Shor's is the only "real" threat. It's 2026 and nothing above the number 15 has ever been factored with it. Quantum computers at such scale are at least 10 years away. https://t.co/vHr2w7IWcj

[justin @sharpone74]

@alexwg The same Dunning-Kruger label applies to your take. A PhD in computer science should know the difference between recovering 1990s public parameters and breaking modern cryptography.

Anthropic's ECDSA announcement shows Claude reverse engineered the public seeds for the NIST ECDSA curves. They were the output of a weak C library PRNG lrand48 called with contemporary unix timestamps. That’s pattern-matching on published numbers plus lattice reduction on a known weak generator. It was not an 'inversion attack'.

This in no way was an attack on ECDSA itself, and doesn't even hint at a crack in the hardness of the ECDLP. Anthropic states this explicitly in their own results.

Calling this evidence that frontier models will eventually break cryptography (e.g. bitcoin) is flat wrong. It's FUD.

[justin @sharpone74]

@alexwg The same Dunning-Kruger label applies to your take. A PhD in computer science should know the difference between recovering 1990s public parameters and breaking modern cryptography.

Anthropic's ECDSA announcement shows Claude reverse engineered the public seeds for the NIST ECDSA curves. They were the output of a weak C library PRNG lrand48 called with contemporary unix timestamps. That’s pattern-matching on published numbers plus lattice reduction on a known weak generator. It was not an 'inversion attack'.

This in no way was an attack on ECDSA itself, and doesn't even hint at a crack in the hardness of the ECDLP. Anthropic states this explicitly in their own results.

Calling this evidence that frontier models will eventually break cryptography (e.g. bitcoin) is flat wrong. It's FUD.

[justin @sharpone74]

@chamath "Break SHA256 and drain Satoshi's wallet"

You should adopt your proper name: Chamath Dunning-Kruger Palihapitiya.

1) There are many block reward P2PK addresses associated to Satoshi, not "a wallet".
2) It is ECDSA, a use of the ECDLP which protects 'wallets', not SHA256. SHA256 is used as the proof-of-work mechanism, to add new transactions to the timechain.
3) AI is no where close to "solving" the ECDLP. Go ahead an ask an LLM yourself why this is and why even an advanced 'math' model is unlikely to change it.
4) Shor's is the only "real" threat. It's 2026 and nothing above the number 15 has ever been factored with it. Quantum computers at such scale are at least 10 years away. https://t.co/vHr2w7IWcj

[justin @sharpone74]

@CriptoZorro Buyer beware. Terrible support.

[justin @sharpone74]

@SoloSatoshi I won't buy another product from @heatbit_com ever again. Which sucks because this product looks cool, but their support has been absolutely horrendous.