Thoughts about Quantum Blockchain Technologies, QBT.

[BobsSocks]

Hi,

Apologies if I am in the wrong place and possibly should not be here anyway. If you are interested then this will probably need a bit of research but I can attempt to answer any questions and point to online resources.

For a start,

https://quantumblockchaintechnologies.co.uk/
https://quantumblockchaintechnologies.co.uk/investor-relations/regulatory-news
https://quantumblockchaintechnologies.co.uk/patents-ip
https://quantumblockchaintechnologies.co.uk/products-services
https://twitter.com/QBTOfficial256
https://www.youtube.com/channel/UCYNcH8Fv9wZ31ZOeQeAmhwQ

The company makes various claims about its technology which can be quite hard to decipher if at all. The latest buzz from them is testing of something they call Method A and Method B, Method C is also in the pipeline. The most recent brick wall they have hit is porting Method A and Method B onto mining rigs. In particular the writing or modification of firmware in particular for BitMain rigs.

Their most recent regulatory announcement, they are listed on the AIM market in the UK, is here,

https://quantumblockchaintechnologies.co.uk/images/QBT_-_RD_Update_March_24-_final.pdf

which makes mention of CGminer,

"In addition to the above, the development of Method B, as an extension to CGminer, a standard operating system used by almost all commercial mining rigs, has been particularly complex, given the intricacies of more than 50,000 lines of open-source C-code developed by the CGminer community. The key issues have been addressed and the Company is performing intensive live mining tests 24/7 using ASIC-based Bitcoin mining devices connected to two large mining pools."

As I understand things CGMiner is released under some form of Open Source license which I would assume requires that any developments are released back to the community. I believe BitMain has not behaved honorably in respect of this. QBT itself often makes noises about others not making their IP available resulting in delays for them as they have to reverse engineer things and then in the next breath talk about all the Patents they are filing to protect their own IP.

It very much looks to be the case that they are targeting modern rigs including the latest offerings from BitMain and, if you believe their press have managed to do this using a modified version of CGMiner which at the moment does not seem to support these rigs. I doubt that they will release this back to the community.

In case you are wondering I can't say I really like the company and their lack of transparency. However I would be interested in your thoughts both on the Company, the Companies Technology and its prospects of success and their underlying behaviour.

Thanks in advance.

[1714622456]

1714622456

[1714622456]

1714622456

[1714622456]

1714622456

[BobsSocks]

Apologies for the boost.

No takers then?

This company, or at least their investors, want you to believe that they will be the salvation of Bitcoin and probably crypto in general when the halving arrives.

They also appear to want to lock you into using their technology.

Assuming their stuff works, which I doubt. Does this sound like a good thing?

[odolvlobo]

It seems very specific to mining, but I am not a miner, so it is not really that interesting to me unless they expose some serious flaw.

[BobsSocks]

Well. The claim, at least for what they call their AI solutions, is that SHA256, or the block generation has flaws that result in predictable patterns that they can use to increase the probability of finding the winning solution.

https://www.youtube.com/watch?v=5ehU3hDMFEM
https://youtu.be/5ehU3hDMFEM?t=51

This is not a hardware solution such as AsicBoost which as I understand things was open sourced by Timo Hanke and then modified without giving back by BitMain. The suggestion is that they have found an exploitable flaw and they will not be sharing it.

[BobsSocks]

It looks like the FT has done a hatchet job on them. Unfortunately you need a subscription to read it.

https://www.ft.com/content/8aa731a0-5456-4121-a2f7-828ae90ed199

The excerpts I have seen are not very flattering.

[NotFuzzyWarm]

Glad you found the right area for this (or close enough)
One thing that must be pointed out is if they have found a weakness in sha256 the effect of it goes far beyond Bitcoin. SHA256 is used as the basis of most commercial financial and document encryption because (to date) it is uncrackable when a long enough key is used - finding a way to hack it would have major repercussions across many markets and industries.

The weakness they claim to have found is based on them stating that AI-driven pattern analysis can narrow down the data ranges that still must be guessed by established brute-force miner chips. As I said in the other thread I do not think it can be done because there are no patterns in proper sha256 encryption. If any patterns DO exist in a hash or series of hashes it is the result of a problem with the RNG that created the hash in the 1st place. Most - but not all - RNG's do not have issues and generate truly random numbers that are used as an encryption seed. Knowing the bias that any given RNG can introduce would certainly allow for making more accurate guesses to possible decryption values provided that you know which specific RNG was used and what (if any) flaws it has. If a different RNG is used that advantage is gone.

Regarding Bitmain, Canaan, Bitfury, et al - all of those miner companies also make pattern recognition AI chips. You really think they have not investigated this?

A worrisome bit from the Financial Time article:

Quantum Blockchain told FTAV it has not published any papers on technology or methodology “in order to protect their research”. Its patent applications are “as far as they have dared to go with making the developments public,” said a spokesman

which to me raises a huge red flag. That is NOT how research works! Publishing papers is the ONLY way to have ideas properly examined and critiqued by ones peers. Not publishing and using the excuse that it is 'to protect their research' screams scamming Marketing Speak.

A Patent does NOT have to describe how to create a workable process, device, or idea. It only must describe how you say it will work. The numerous patents for perpetual motion machines, free-energy devices and their ilk prove that. Pursuing and even being granted a patent for something does not mean it works or ever can work.

[BobsSocks]

Apologies for the late.

Someone snaffled a copy of the FT article

https://archive.ph/PR7L6

I get the impression that the author, and possibly others, are also a bit cynical as well.

[BobsSocks]

As an add on the subject of patent applications. The company announced filing the first on 30 Sep 2021 and the second on 24 Jul 2023,

https://quantumblockchaintechnologies.co.uk/images/QBT_-_ASIC_UltraBoost_Patent_-_Final.pdf
https://quantumblockchaintechnologies.co.uk/images/QBT-_Patent_Application_-_July_23.pdf

They make it a bit difficult to view the actual progress of these applications,

https://quantumblockchaintechnologies.co.uk/patents-ip

but they are listed on the UKPTO website,

https://www.ipo.gov.uk/p-ipsum/Case/PublicationNumber/GB2611321
https://www.ipo.gov.uk/p-ipsum/Case/ApplicationNumber/GB2310704

Unfortunately as a result of the way the patent system operates, details are not made public until 18 months after the filing date, in the case of the second it is not possible to see any information. It is possible in the the case of the first,

https://www.ipo.gov.uk/p-ipsum/Case/PublicationNumber/GB2611321/Documents

Some of you will recognise the words in the title as being similar to ASIC Boost as developed and open sourced by Timo Hanke,

https://www.ipo.gov.uk/p-ipsum/Case/PublicationNumber/GB2611321/Documents

and that was one of the original objections made by the examiner against QBT's application. Subsequently QBT appear to have overcome this objection and other examples of prior art but the examiner introduced objections surrounding excluded material, in particular a 'computer program' and a 'business method',

https://www.gov.uk/guidance/manual-of-patent-practice-mopp/section-1-patentability

QBT failed to answer these objections in time even though they did request a two month extension. However in the meantime they had transferred the application to the PCT, Patent Cooperation Treaty, under the EPO, European Patent Office,

https://register.epo.org/application?number=EP22793447&lng=en&tab=main
https://register.epo.org/application?number=EP22793447&lng=en&tab=doclist

where it is presently under examination.

EPO normally defers to UKPTO rulings on patentability but it is likely that QBT, having missed the deadline in the UK, will be able to argue their case.

You can see something similar happening in anothe application, that moved through to grant, by Timo Hanke and was assigned variously to, TOP GALORE LIMITED, LITTLE DRAGON TECHNOLOGY LLC and possibly another,

https://register.epo.org/application?number=EP14864642&lng=en&tab=main
https://register.epo.org/application?tab=doclist&number=EP14864642&lng=en

That grant was subsequently challenged and revoked and the case is presently subject to appeal.

I am uncertain but I am aware that their is an active group withing the field of BitCoin/BlockChain that seek to prevent such ownership of Intellectual Property in order to maintain a level playing field within the community. Whether they will take an interest in the application made by QBT remains to be seen.

As to the application itself whilst I have little to no knowledge about SHA256 and its use within the BlockChain it does appear to have technical merit. At least at a level that the UKPTO was prepared to accept. As such it is possible, likely, that QBT will be able to argue their case for Grant under the PCT. QBT will not be sharing this.

Aside from this desire to operate a closed shop I would suggest that the company has not exactly been open about their own progress. Whilst they announced the applications they did not announce that the first application was in trouble and most recently,

https://quantumblockchaintechnologies.co.uk/images/QBT_-_RD_Update_March_24-_final.pdf

Patent Applications
•
•
The Company is making positive headway with its first two patent applications. A third patent application is being drafted for the proprietary quantum version of SHA-256.

The Company’s first patent application (ASIC UltraBoost) has now been filed with the European Patent Office, while the second patent application (ASIC EnhancedBoost) is going through a standard exchange of questions and answers with the UK Patent Office. A third patent application is being drafted for the proprietary quantum version of SHA-256 and will be submitted as soon as practical.

YMMV on positive headway and "standard exchange of questions" most likely means that the second application has been subject to challenge. Again we will not see the nature of those communications until the case file is made public some time in December this year.


 

[NotFuzzyWarm]

I hope that the Patent Examiners are aware of this prior work https://www.nature.com/articles/d41586-024-00839-6  oh, the 1st post about it is here https://bitcointalk.org/index.php?topic=5490005.0

That proper and well published research sounds an awful lot like what QBT claims: Methods to reduce the probable ranges of results when analyzing processes that were previously considered to be highly if not entirely random.

The beginnings of the mathematicians research certainly predates the QBT Patent applications... That said, the techniques he won the Award for (so far) can only be applied to unique problems where the datasets being analyzed does not change every 10 minutes...

[BobsSocks]

Whoops,

I did suggest elsewhere that if QBT could not work out what 50K of CGMiner code does they might like to ask one of the developers.

https://twitter.com/ckpooldev/status/1773906971806298306

It seems they did and got a no.

Probably offered waffle with options on top.

[NotFuzzyWarm]

Glad that -ck told them to piss off. FWIW -ck closed down his development of CGminer several years ago. If they ever reach out to the sole remaining active Primary developer of CGminer (Kano) he will not be so nice in his reply to them   Link to Kano's git is in my sig.

If they can't figure out the CGminer code then it just further proves they are most likely marketing scammers holding out the sparkly shiny promise of 'something revolutionary' based on the Majik buzzwords "AI Learning" to attract Investards willing to throw money at them and that they have no one on their 'team' that is even moderately skilled at programming. It is NOT that complex and each block of code is very well documented.

The folks behind Braiins figured it out when they did a clean rewrite of the cgminer code (using Rust) to make a more 'tweakable' 3rd party firmware that was not a hack of CGminer to use with Bitmain and other makes of miners. They were able to recreate the functionality of what CGminer does and created their own code structures to do the same things.

As for this closing line in QBT's pdf regarding using their ideas in the real word with real mining chips:

This has proved a very time consuming, due to the tremendous reverse engineering efforts involved in
using totally undocumented third-party ASIC chip

Um, first vh and more recently, Kano, did it for the Bitmain chips that Sidehack uses in his miners... Ja it took Sidehack a fair bit of time using a logic analyzer to map the bitstreams in & out of the chips so vh/Kano could write the driver coms code needed but again eminently doable if someone knows what they are doing.

[BobsSocks]

Thanks for the reply and the added information.

I had suggested elsewhere that if QBT were having problems with CGMiner they might consider contacting one of the originators. I had also suggested that their moaning about Bitmain and the lack of documented information was a bit, more than, broken given the boards are accessed via a header and all you would have to do it hook up a breakout cable and probe them. I'm not all that but I do know a bit about electronics.

OK. Tell me to go away if I am overstepping the mark here.

I did ask Con via e-mail about QBT on the 16th March and he responded quite quickly to say that he had never heard of them and they were likely to be making claims they could not back up. I included a copy of the RNS published on the 13th,

http://hTTps://quantumblockchaintechnologies.co.uk/investor-relations/regulatory-news
http://hTTps://quantumblockchaintechnologies.co.uk/images/QBT_-_RD_Update_March_24-_final.pdf

where QBT claim...

In addition to the above, the development of Method B, as an extension to CGminer, a standard operating system used by almost all commercial mining rigs, has been particularly complex, given the intricacies of more than 50,000 lines of open-source C-code developed by the CGminer community. The key issues have been addressed and the Company is performing intensive live mining tests 24/7 using ASIC-based Bitcoin mining devices connected to two large mining pools.

Then Con suggests on the 30th via Twitter that he has been contacted by a company that smells extremely like QBT and told them to get lost. Top Bloke. He has not replied to my later emails so until this happened I was leaving him alone but I did ask again and still got no reply. That's cool. My actions perhaps not so much.

However I am left wondering about the timeline. Con, on the 16th, does not know about them. The company on the 13th say they have solved their problems. Con on the 30th says that a company that smells like them have asked for help sorting out the problems that QBT were having with CGMiner. Something does not add up, smells.

I am due to have a lump removed under general anesthetic in the next couple of weeks and I would trust Con implicitly to put me under. If QBT were claiming to have invented some new fangled anesthetic I would invite them to go do one.

You mention Kano as being the last person actively developing and supporting CGMiner. My Git skills are non-existent but I did, kind of, understand the last pull request. I may be wrong but that sort of thing would make me understand why supporting such projects can be a bit of a bind. Especially when a company of MegaBrains rocks up in your email telling you that your software is rubbish and asks you to fix it for them.

Again, assuming I am not overstepping the mark and apologies for using you as a proxy, could you ask Kano if they have been approached by QBT?

[NotFuzzyWarm]

I know for a fact that he has not been contacted by them as it is a certain fact that he would bring it up in his KanoPool Discord #imboredwhattodo channel. Thing is Kano has degrees in computer science and statistics - he would rip QBT a new asshole over what they are claiming.

That said - IF QBT succeeds in having their patent applications approved, even if they cannot actually make it work (to-date is impossible) then what they CAN lay claim to as becoming Prior Art is the concept of using AI or ML pattern analysis to speed up sha-256 decryption. End result is that when tech finally might be able to do it they could then say, 'Ehem, that was our Patented idea first and here is our Royalty Terms...'. In other words become a NPE Patent Troll. In the Patents world NPE stands for Non-Producing Entity meaning they do not actually make anything that uses their patented idea in any way but rather use their patent as a tool to force license/royalty payments from companies that can and do use it.

[BobsSocks]

In terms of the patents, or at least this visible one, with my limited knowledge of such things the application and the inventor seem extremely credible. I just get the impression that they have been comprehensively failed by the company that took it on board. I kind of missed/discounted your possible challenge on the basis of prior published research because I did not think it might apply. As things stand under the PCT/EPO,

https://register.epo.org/application?number=EP22793447&lng=en&tab=main

Anyone is able to submit a Third Party Observation and do so anonymously. There is a link at the top "Submit Observations". I have used that in the past to trash a different companies application prior to grant but you have to argue a reasonable case. I got a mention in the "Go Away" reply from the examiner suggesting that the applicants should pay careful attention to my submission before they tried it on again.

The second application, that is as yet not visible, also appears to be a hardware/silicon solution. I could be wrong, it is hard to work out what the company thinks they are doing beyond making stuff up, but this new Method C is meant to be tied into that second silicon solution.

Method C, like Method A and B, is supposedly AI based and I think they accept that getting a patent on an AI, made up, black box might be a bit hard. As a result it is probably better to claim that you have one and hang some numbers on it to entice mug punters rather than ask an independent third party to laugh at you.

Elsewhere Con is getting a bit of flak from The Believers who generally go about projecting their own reason for fail on others. In part that is why I am here.

In terms of ripping new arseholes (assholes) and statistics I am not by any means a statistician but I assume that QBT has one and would be able to explain why the testing of any of their fantasy might take a bit of time because...

https://ycharts.com/indicators/bitcoin_network_hash_rate
https://m.bitmain.com/product/detail?pid=00020231025110633223f8vE5pxQ06AB

So, assuming all rigs are T21, sold out, there are 640.5M/190 rigs on the go. Call it 3.4 Million. Collectively they, are forced to, solve one block every 10 minutes. So, statistically speaking, it takes each one 34 Million minutes to solve a block. That's about 64 years.

QBT claims that Method B improves the chance of getting an answer by 2.6 so a QBT enhanced rig would take 25 years. QBT now claim that they are testing this on the live network.

https://quantumblockchaintechnologies.co.uk/images/QBT_-_RD_Update_March_24-_final.pdf

I don't think *mining in a pool* qualifies as being truly live and, of course, QBT give no indications as to how many rigs they are using so based on the information they do give the minimum number required to qualify as true is... Two. Assuming one rig is enhanced and the other is not then they might mine a block in the next 25 years because one block is going to be statistically significant.

My wet finger, and a rubbish Pascal Program wot I wrote kind of suggests that if they ran 2,000 rigs, 1,000 enhanced the other 1,000 not, they might get statistically significant results in about 2 months. Naturally the company gives us no idea as to what they are really doing or any of the statistics behind it other that it is intensive.

I thought rigs were by definition a bit intensive unless you understand how the software that messes about with their core clocks and voltages can be adjusted works.

And your footnote is that, to me, *mining in a pool* means you borrowed someone else's software and have to mine in their pool because your PHd Megabrains could not work out Git or read a documented program. Of course I don't, yet, know how to do that but I am not a MegaBrain trying to mug investors so Yah Boo.