Bitcoin puzzle transaction ~32 BTC prize to who solves it

[Friday4227]

This thread is officially full-on idiocracy mode.

So just in last 24h several guys broke secp256k1, ECC, new breaking discoveries on the algo side, the sqrt lower bound is basically ancient history due to linear searches and 128 GB of RAM, and now someone will break 135 before I finish my dinner? Is everyone gone mad? All of these are total delusions. Even the prefix guys had more brain then all of this non-sense...

Sorry I don't how to use this site. But there is things you don't know that you don't know. There are things that you can't even imagine so don't judge everything your knowledge. And in math ideas that somehow even slightly altered by emotions aren't useful or correct. So if you think it's not possible it's not possible for you.

[John_Ahmet]

Hello everyone, I wanted to ask you what speeds do you get in EC, and what search speeds do you have? In sequential BSGS, if anyone still does it that way.

Around 1 X per addition. Not sure what "sequential BSGS" means though.

I also don't really understand what you mean by saying 1x.
I am talking about sequential BSGS with jumps but linear in space.
No area between points is skipped, I am interested in what EC speeds you have.
What do you use, CPU or GPU?
And what actual search speed do you have in G/Ks in space?

I meant I get the speed of obtaining one X coordinate for every point addition. This is what you asked. In lack of any other context, this can get really really fast (> 25 G/s on a stock RTX 4090) but if you add context like "is X in some 100M list?" expect a 100x drop due to memory access latencies.

Still unclear of what's a "sequential BSGS", but rest assured, both a blazingly fast linear search ("does x equal target?") and BSGS are non-applicable for anything above ~80 bits key space.

I understand that you don't understand what sequential BSGS is. That doesn't surprise me, because many people – if not everyone – use public algorithms. I have my own. I worked on it for 8 months, writing it, testing it, fixing it, modifying it, optimizing it, and I'm still doing it. Every time I increase the speed. Currently I have 7 EK/s of sequential SECP curve search. Every searched point G=K. But that's not the end. I know perfectly well that for puzzle 135 it's still not enough, but there are certain tricks that shorten things. With this speed, searching the entire space of puzzle 70, I recover the key in 140 seconds. In my opinion that's a good result, but as I mentioned – not the final one! As you all know, the average block time is 10 minutes you understand what I mean, so it's not bad. I'm focused on puzzle 135 – that's my goal. Because I spent 8 months writing my own algorithm, I now wanted to compare my speed to others, as well as check what others have and what algorithms they use. I don't want anyone's codes, I simply want to know your speeds. Of course, they depend not only on the algorithm but also on the hardware. The simplest explanation is giving the time to solve one of the puzzles. I still have a few ideas for modifications and optimizations of the algorithm, but it's getting more difficult and time-consuming. I'm looking for something that doesn't exist – that's the key. Someone also asked earlier about RAM – I use 128 GB.

I've also started writing code. I'm undecided between Kangroo and BSGS. Since I don't have a GPU, I'm only interested in CPU solutions. Of course, I'm interested in the fastest one. With Claude Sonnet 4.6 and Opus 4.6, it's become easier to produce more successful solutions. If your method is really good, I'd like to contribute.

Sorry I don't how to use this site. But there is things you don't know that you don't know. There are things that you can't even imagine so don't judge everything your knowledge. And in math ideas that somehow even slightly altered by emotions aren't useful or correct. So if you think it's not possible it's not possible for you.

Why waste time trying to convince negative people? I know we can speed up these key-finding algorithms much more. In that sense, if you consider those who have ideas, you'll get more out of them.

[Friday4227]

Hello everyone, I wanted to ask you what speeds do you get in EC, and what search speeds do you have? In sequential BSGS, if anyone still does it that way.

Around 1 X per addition. Not sure what "sequential BSGS" means though.

I also don't really understand what you mean by saying 1x.
I am talking about sequential BSGS with jumps but linear in space.
No area between points is skipped, I am interested in what EC speeds you have.
What do you use, CPU or GPU?
And what actual search speed do you have in G/Ks in space?

I meant I get the speed of obtaining one X coordinate for every point addition. This is what you asked. In lack of any other context, this can get really really fast (> 25 G/s on a stock RTX 4090) but if you add context like "is X in some 100M list?" expect a 100x drop due to memory access latencies.

Still unclear of what's a "sequential BSGS", but rest assured, both a blazingly fast linear search ("does x equal target?") and BSGS are non-applicable for anything above ~80 bits key space.

I understand that you don't understand what sequential BSGS is. That doesn't surprise me, because many people – if not everyone – use public algorithms. I have my own. I worked on it for 8 months, writing it, testing it, fixing it, modifying it, optimizing it, and I'm still doing it. Every time I increase the speed. Currently I have 7 EK/s of sequential SECP curve search. Every searched point G=K. But that's not the end. I know perfectly well that for puzzle 135 it's still not enough, but there are certain tricks that shorten things. With this speed, searching the entire space of puzzle 70, I recover the key in 140 seconds. In my opinion that's a good result, but as I mentioned – not the final one! As you all know, the average block time is 10 minutes you understand what I mean, so it's not bad. I'm focused on puzzle 135 – that's my goal. Because I spent 8 months writing my own algorithm, I now wanted to compare my speed to others, as well as check what others have and what algorithms they use. I don't want anyone's codes, I simply want to know your speeds. Of course, they depend not only on the algorithm but also on the hardware. The simplest explanation is giving the time to solve one of the puzzles. I still have a few ideas for modifications and optimizations of the algorithm, but it's getting more difficult and time-consuming. I'm looking for something that doesn't exist – that's the key. Someone also asked earlier about RAM – I use 128 GB.

I've also started writing code. I'm undecided between Kangroo and BSGS. Since I don't have a GPU, I'm only interested in CPU solutions. Of course, I'm interested in the fastest one. With Claude Sonnet 4.6 and Opus 4.6, it's become easier to produce more successful solutions. If your method is really good, I'd like to contribute.

Sorry I don't how to use this site. But there is things you don't know that you don't know. There are things that you can't even imagine so don't judge everything your knowledge. And in math ideas that somehow even slightly altered by emotions aren't useful or correct. So if you think it's not possible it's not possible for you.

Why waste time trying to convince negative people? I know we can speed up these key-finding algorithms much more. In that sense, if you consider those who have ideas, you'll get more out of them.

I don't think it's about physical strengthening. It's about the creating a pattern not looking it. This has nothing to do anything else you knew or even familiar with the most close concept I can find is you're in 3 dimension world seeking answer to other big dimensions like 79d.
I'll consider you but we have to finish this puzzle in few hours or I'll die

[John_Ahmet]

Meanwhile, I solved puzzle 70 using only the CPU in a minimum of 7 minutes. On average, I definitely solve it in 12 minutes. I'm using the Kangroo algorithm, but I've never used BSGS. I'm not sure how long it would take me to crack it with that.

CPU: Intel(R) Core(TM) i7-7700 CPU @ 3.60GHz   3.60 GHz
RAM: 32GB

I don't think it's about physical strengthening. It's about the creating a pattern not looking it. This has nothing to do anything else you knew or even familiar with the most close concept I can find is you're in 3 dimension world seeking answer to other big dimensions like 79d.
I'll consider you but we have to finish this puzzle in few hours or I'll die

You're not providing any information, so I can't understand you.

[Niekko]

Meanwhile, I solved puzzle 70 using only the CPU in a minimum of 7 minutes. On average, I definitely solve it in 12 minutes. I'm using the Kangroo algorithm, but I've never used BSGS. I'm not sure how long it would take me to crack it with that.

CPU: Intel(R) Core(TM) i7-7700 CPU @ 3.60GHz   3.60 GHz
RAM: 32GB

I suggest you to look the PointsBuilder from kTimesG, it's the most fast solution for CPU.

With a simple modifications AVX2 for h160, you can use it for 71 too

[Friday4227]

Meanwhile, I solved puzzle 70 using only the CPU in a minimum of 7 minutes. On average, I definitely solve it in 12 minutes. I'm using the Kangroo algorithm, but I've never used BSGS. I'm not sure how long it would take me to crack it with that.

CPU: Intel(R) Core(TM) i7-7700 CPU @ 3.60GHz   3.60 GHz
RAM: 32GB

I don't think it's about physical strengthening. It's about the creating a pattern not looking it. This has nothing to do anything else you knew or even familiar with the most close concept I can find is you're in 3 dimension world seeking answer to other big dimensions like 79d.
I'll consider you but we have to finish this puzzle in few hours or I'll die

You're not providing any information, so I can't understand you.

Why would I wanna share anything here. Everybody is watching.. find a way. I can show you something

[Grzegorz2022]

Meanwhile, I solved puzzle 70 using only the CPU in a minimum of 7 minutes. On average, I definitely solve it in 12 minutes. I'm using the Kangroo algorithm, but I've never used BSGS. I'm not sure how long it would take me to crack it with that.

CPU: Intel(R) Core(TM) i7-7700 CPU @ 3.60GHz   3.60 GHz
RAM: 32GB

I suggest you to look the PointsBuilder from kTimesG, it's the most fast solution for CPU.

With a simple modifications AVX2 for h160, you can use it for 71 too

Can you share what CPU and the time to find a key for a puzzle? Thanks

This thread is officially full-on idiocracy mode.

So just in last 24h several guys broke secp256k1, ECC, new breaking discoveries on the algo side, the sqrt lower bound is basically ancient history due to linear searches and 128 GB of RAM, and now someone will break 135 before I finish my dinner? Is everyone gone mad? All of these are total delusions. Even the prefix guys had more brain then all of this non-sense...

Could you tell me the time it takes to find the key for one of the puzzles with your PointsBuilder program, and what CPU you're using?

[John_Ahmet]

I don't think it's about physical strengthening. It's about the creating a pattern not looking it. This has nothing to do anything else you knew or even familiar with the most close concept I can find is you're in 3 dimension world seeking answer to other big dimensions like 79d.
I'll consider you but we have to finish this puzzle in few hours or I'll die

This repository is only used to find and store X coordinates. It's a tool to help find keys, but it seems insufficient on its own. I compiled and ran the project on my machine. I have 10MK/s speed, but it's not very useful for finding keys!

What parameters should I provide to solve puzzle 70? We can start by measuring that.

[cryptouser1001]

Friday4227, you just might have the answers we're all looking for here.

A 79D being could:

i.) Perceive the entire key space as a geometric shape, locating the private key via topological minimization.

ii.) Use hyper-dimensional algebra to collapse the ECDLP search space by exploiting symmetries invisible in 3D.

iii.) Take advantage of quantum-gravitational computation across dimensions to brute-force keys in real time.

[Scanuserfast]

I scanned last night 79D and is not there. So have fun!

[brainless]

About mara
https://www.coindesk.com/markets/2026/04/27/mara-holdings-targets-bitcoin-quantum-threat-and-network-resilience-with-new-foundation

4. Eliminate puzzle 71

Do you mind to elaborate on this? How exactly can the creator (user saatoshi_rising claimed to be it) eliminate puzzle #71 other than being able to mine a block with a non-public mempool? We already know that it's key to not broadcast the spending transaction of puzzle #71 in the public.

If you have no clue what you're talking about, how about simply not talking and writing nonsense and polluting this thread with noise?


BTW, I wrote an email to slipstream@mara.com asking when Slipstream service will be available again and let's see if and what they answer.

Already asked mara, 4 times in 10 days, no reply to me still after 10 days, I think support also shutdown about these subject
If you get any answer, sure waiting your post about it

Have you received any reply from mara regarding their slipstream service back ?

[kTimesG]

A 79D being could:

i.) Perceive the entire key space as a geometric shape, locating the private key via topological minimization.

ii.) Use hyper-dimensional algebra to collapse the ECDLP search space by exploiting symmetries invisible in 3D.

iii.) Take advantage of quantum-gravitational computation across dimensions to brute-force keys in real time.

Hope you were joking There are no higher-dimensional symmetries whatsoever, waiting to be discovered, which means there aren't any topological relations which might exist.

This is derived from the curve equation, so the only relations between points are the symmetry and the endomorphism. Nothing else CAN exist. Whoever seeks them, will get slapped hard by the math.

As for higher-dim BSGS: every time you add a new dimension, you are increasing the difficulty, approaching a linear search.

For example, using a "cube" for searching means you need to first compute an entire cube face, to be able to detect the collision with the line, so going from sqrt(n) to a cuberoot(n)**2 complexity (which is higher). Increase dimensions = increase complexity.

[Friday4227]

A 79D being could:

i.) Perceive the entire key space as a geometric shape, locating the private key via topological minimization.

ii.) Use hyper-dimensional algebra to collapse the ECDLP search space by exploiting symmetries invisible in 3D.

iii.) Take advantage of quantum-gravitational computation across dimensions to brute-force keys in real time.

Hope you were joking There are no higher-dimensional symmetries whatsoever, waiting to be discovered, which means there aren't any topological relations which might exist.

This is derived from the curve equation, so the only relations between points are the symmetry and the endomorphism. Nothing else CAN exist. Whoever seeks them, will get slapped hard by the math.

As for higher-dim BSGS: every time you add a new dimension, you are increasing the difficulty, approaching a linear search.

For example, using a "cube" for searching means you need to first compute an entire cube face, to be able to detect the collision with the line, so going from sqrt(n) to a cuberoot(n)**2 complexity (which is higher). Increase dimensions = increase complexity.

 If you lift the curve there is leakage of the piraty of key. ??

[NUCLEAR7.1]

Regarding the brute-force penetration test used to measure the security of Bitcoin keys, 83 bits should be the maximum within the framework of the puzzle. Even with this generation, it's impossible to reach and break 83 bits, and quantum computers won't be of any use. Therefore, the creator should be described as having been stupid for 9 years and +2 years.  

The puzzle creator could improve the puzzle's utility without bringing in any extra funds from outside - just spend 84-160 across to the unsolved portion 71-83
This allows the creator to maintain his creative idea of making the reward number resemble the puzzle number, so that the reward for puzzle 71 is 71 BTC after applying this change. However, the problem the creator will face is that he must find a secure way to transfer funds from 84-160, as the public keys will be revealed. Therefore, he must make a transfer secretly and securely. If he does not, 84-130 will be vulnerable to theft.  

if the creator wanted to increase the reward slightly.  
The creator must abandon their creative idea of making the reward number resemble a puzzle number. They could change the reward from 136-160 to 71-135, distributing 370 Bitcoin across these wallets and increasing the reward percentage to approach or exceed the cost. However, this would negate their creative idea of making the reward number resemble a puzzle number.

We all know about the large scale of the problem. Therefore, the cost will always outweigh the reward and remain economically unviable even with these increases in reward. If the goal is simply to create incentives and excitement, then so be it. As for quantum computers, I don't think they will do anything.

If the creator implements my idea, this will increase the popularity of these puzzles among the new generation. 

[Jorge54PT]

Without Mara's slipstream, it's no longer advisable to work with wallets up to 100 bits. Nothing guarantees that you'll be able to receive payments without bot attacks. That's why I've already abandoned the 71 bit and switched to the 135 bit

[NUCLEAR7.1]

Regarding the brute-force penetration test used to measure the security of Bitcoin keys, 83 bits should be the maximum within the framework of the puzzle. Even with this generation, it's impossible to reach and break 83 bits, and quantum computers won't be of any use. Therefore, the creator should be described as having been stupid for 9 years and +2 years.  

The puzzle creator could improve the puzzle's utility without bringing in any extra funds from outside - just spend 84-160 across to the unsolved portion 71-83
This allows the creator to maintain his creative idea of making the reward number resemble the puzzle number, so that the reward for puzzle 71 is 71 BTC after applying this change. However, the problem the creator will face is that he must find a secure way to transfer funds from 84-160, as the public keys will be revealed. Therefore, he must make a transfer secretly and securely. If he does not, 84-130 will be vulnerable to theft.  

if the creator wanted to increase the reward slightly.  
The creator must abandon their creative idea of making the reward number resemble a puzzle number. They could change the reward from 136-160 to 71-135, distributing 370 Bitcoin across these wallets and increasing the reward percentage to approach or exceed the cost. However, this would negate their creative idea of making the reward number resemble a puzzle number.

We all know about the large scale of the problem. Therefore, the cost will always outweigh the reward and remain economically unviable even with these increases in reward. If the goal is simply to create incentives and excitement, then so be it. As for quantum computers, I don't think they will do anything.

If the creator implements my idea, this will increase the popularity of these puzzles among the new generation. 

I think he has enough power to cooperate with a miner privately (marapool) and secretly so that the public key is not stolen and does not appear via mempool

[rt07]

Yes, with AI I did a summary.

Quantum computers → threaten old Bitcoin wallets → ~5.6M BTC are vulnerable → this creates two simultaneous crises: technical (BIP-361 proposes freezing the coins) and market-related (any freeze = panic and the worst day in Bitcoin's history).
MARA, as the largest public miner with 38,000+ BTC on its balance sheet, can't ignore this — and responds in four ways at once: launches the MARA Foundation for quantum research, pivots into AI/HPC data centers (hedging against Bitcoin instability), sold 15,133 BTC ($1.1B) to pay off debt, and in the process shuts down niche products like Slipstream.

[NUCLEAR7.1]

Yes, with AI I did a summary.

Quantum computers → threaten old Bitcoin wallets → ~5.6M BTC are vulnerable → this creates two simultaneous crises: technical (BIP-361 proposes freezing the coins) and market-related (any freeze = panic and the worst day in Bitcoin's history).
MARA, as the largest public miner with 38,000+ BTC on its balance sheet, can't ignore this — and responds in four ways at once: launches the MARA Foundation for quantum research, pivots into AI/HPC data centers (hedging against Bitcoin instability), sold 15,133 BTC ($1.1B) to pay off debt, and in the process shuts down niche products like Slipstream.

Quantum computers will not threaten unspent wallets

[rt07]

Yes, with AI I did a summary.

Quantum computers → threaten old Bitcoin wallets → ~5.6M BTC are vulnerable → this creates two simultaneous crises: technical (BIP-361 proposes freezing the coins) and market-related (any freeze = panic and the worst day in Bitcoin's history).
MARA, as the largest public miner with 38,000+ BTC on its balance sheet, can't ignore this — and responds in four ways at once: launches the MARA Foundation for quantum research, pivots into AI/HPC data centers (hedging against Bitcoin instability), sold 15,133 BTC ($1.1B) to pay off debt, and in the process shuts down niche products like Slipstream.

Quantum computers will not threaten unspent keys

Maybe but MARA panicked and already closed Slipstream.

[NUCLEAR7.1]

Yes, with AI I did a summary.

Quantum computers → threaten old Bitcoin wallets → ~5.6M BTC are vulnerable → this creates two simultaneous crises: technical (BIP-361 proposes freezing the coins) and market-related (any freeze = panic and the worst day in Bitcoin's history).
MARA, as the largest public miner with 38,000+ BTC on its balance sheet, can't ignore this — and responds in four ways at once: launches the MARA Foundation for quantum research, pivots into AI/HPC data centers (hedging against Bitcoin instability), sold 15,133 BTC ($1.1B) to pay off debt, and in the process shuts down niche products like Slipstream.

Quantum computers will not threaten unspent keys

Maybe but MARA panicked and already closed Slipstream.

Man, I think there is a better alternative to Slipstream, and there will certainly be one in the future.