Member Info for Phimx

Mickeydj cannot dispute what I say about the impossible Method C software being impossible so he is reduced to the brute force approach and has my posts deleted. If what I say is misleading or lies then Mickdj can explain to folks where I lie and what and what in my posts is misleading.

Go ahead Micky, explain how the ESP32-3 can filter at unbelievable speeds. The computer community is waiting with bated breath to hear about another QBT breakthrough!

.

Taking months to discover the analysis can't take place is a revelation and a ludicrous excuse.

RNS 5 December 2025

You have to marvel at the drivel Gardin comes out with

" Joint analysis of the hardware version of Method C AI Oracle had previously commenced with one of the three ASIC manufacturers, as announced on 8 April 2025. However, due to the need for the hardware version to be integrated into the ASIC architecture, a process which could take up to 18 months for the new chip to come to market, it has been mutually agreed to give operational priority to the new software version"

So, both QBT and the ASIC manufacturer began the analysis of QBT's magic Oracle but the the half wits realised after several months that is wasn't possible until it was integrated into an ASIC.

This paragraph is simply ludicrous drivel on a couple of counts. One, experienced logic designers would know immediately what would be required to analyse QBT's magic technology. Two, designs are analysed with simulators and FPGAs . QBT has, we are told, developed and carried out exhaustive testing on FPGAs. They have obtained vast amounts of data and obtained irrefutable evidence.

Just more turn-abouts by gobbledegook Gardin

Mickeydj: You post

" Phimx Remenber this TAT you posted and how wrong you were proved."

I'm not sure what had been proved ! Have deals been announced? Have contracts been signed? Is revenue pouring in ? Have clear unambiguous performance figures been produced?

What is the priority for sales and revenue this week, is it SaaS, ASICs, Method A, Method B, Method C, software, hardware, China, North America, After Market boards, small startups, big miners or just more delay and dreams ??????

Mickeydj: You wrote.

"Phimx: Calls the 10% gain "meaningless" and "Gardin garbage." Ignores lab context and RNS definition (efficiency above pool thresholds). Also mocks revenue promises from 2024—fair point on delays, but overlooks that tech evolution (e.g., software shift) explains pivots."

**************************************

You are keen to cite context in your posrings. So below is my original post the " meaningless 10%" is explained there. Anyone can Google what the pool minimum difficulty threshold is and consider what Gardin said about.

My post 24 November 2025

Consider what Gardin wrote in the latest RNS

"Over the past month, testing of the new software-only version in the Company's Milan laboratory has shown an average improvement in mining efficiency of over 10% - meaning, its capability of finding the winning hash at over 10% above the minimal difficulty threshold as set by most mining pools."

The 𝗺𝗶𝗻𝗶𝗺𝘂𝗺 difficulty threshold set by the pool is the lowest hashing rate a mining rig can have to join a particular pool. So what it appears to tell us, for the pool QBT has joined the QBT miner hashes 10% above X hash rate. In other pools the X hash rate may be different. The same rig could be 2% above in another pool. The statement says nothing about the improvement the software gives to the mining rig.

What is needed is the improved in hashrate of identical rigs but one hashing with the method software. That seems quite simple, why is it hidden behind obscure and meaningless statements from the company?

Mrhanji: You post

" Method A & B stalled at commercials.."

Such negativity!

Davemate90:

Does anyone honestly think that any monies received will find its way into the pockets of shareholders?

Mickeydj:

Unfortunately nowhere in QBT's published information does it mention that blocks of NONCES are being discarded.

From RNS 13 March 2024

" The fundamental feature of Method C is its ability to predict whether an input to SHA-256, the core algorithm for Bitcoin mining, is likely to generate a winning hash, or not. The underlying assumption for Method C is that an oracle decision is materially less computationally demanding than the SHA-256 calculation for the same input string."

We see that the advantage is created by replacing a SHA256 calculation by performing an Oracle decision on the input string. There are no plurals or batches. No mention of a single Oracle decision saving a multitude (millions) of SHA256 calculations, as it would if there was a decision on a batch as you suggest.

Nonces aren't batched. A start value is allocated and then an offset given for each SHA256. So if you had 1,000 channels of SHA256 then 4 Billion/1,000 would give a 1,000 NONCE start values separated by an offset of 4 million.

Mickeydj: You post

" It's predictive batch filtering upstream of parallel cores, not forcing serial per-hash."

Predictive batch filtering upstream of parallel cores, that looks very impressive but total fantasy. You are implying there is some form of batch filtering taking place before the miners microcontroller.

Where would this take place and why would an Oracle be loaded on the miner if the filtering had taken place upstream.

Tell everyone where QBT claims this description for Method C hardware or software. Note you've added the term parallel cores. Of what and where ?

Make believe son pure make believe.

Mickeydj: You post

" CONTEXT: Oracle predicts "likelihood" on inputs/nonce batches—skips irrelevant (30% Jan 2025; 10%+ average Nov 12 on S21 rigs). It's predictive batch filtering upstream of parallel cores, not forcing serial per-hash."

Cite anv QBT RNS or interview that mentions "batches" or any synonym.

What is wrong with my math? 1 TH per second is 1,000 Billion hashes in a second a truly staggering figure. If you don't feed the ASIC with that number of inputs per second you won't hash at that rate.

The only proof is what Gardin writes !

So Braiins randomly skip blocks of inputs and how does that help. Are you saying Braiins are already using a system that improves hashing rates ? If so why would they buy anything from QBT?

You also post

Nov 12 RNS: Software Method C on S21 rigs (234 TH/s) – 10%+ efficiency on live blocks. If serial bottleneck, it wouldn't work." 𝗲𝘅𝗮𝗰𝘁𝗹𝘆 𝗺𝘆 𝗽𝗼𝗶𝗻𝘁

"

Mickeydj: You post

" Yq3our "fairy tale" rants assume per-hash serial – Gardin never claimed that. It's ML batch prediction, proven in labs and on real hardware."

Can you point to anything Gardin has stated that claims "batches" of NONCES are being discarded. Since the hash results are entirely random for each NONCE used by individual SHA256 then each needs to be processed.

Evidence for the serial approach is clearly stated by Gardin in the RNS Oct 2024

" Method C's AI Oracle is able to predict the likelihood of an input to SHA-256 (the core algorithm to mine Bitcoins) to produce a winning Hash. Should the AI Oracle calculate that the current SHA-256 will not be successful in finding the winning Hash, it skips that calculation and moves on to the next input. "

Note the phrase " moves on to the next input" one after the other normally abbreviated to serially.

It is you writing fantasy. Gardin has made a big error with his software method story line

Mickeydj: You post

" Now you're now continuing to offer me investment advice."

I see no investment advice given! I recommended you take heed of what you've learned of AIM. I've learned my hard lessons as I suspect many others have and have deserted the markets. Labour are enticing fresh meat to buy shares I note. More naive punters enticed to pour their savings into get-rich-quick schemes!

Mickeydj:

There is one certainty in the QBT fairy tale and that is, Method C cannot be run fast enough on a BitAxe ESP32 Microcontroller to filter 1 TH/s inputs to the SHA256 in real time mining.

Hashing could be done with pencil and paper. You might manage 1 Hash per 10 mins and theoretically you could hash the winner but it's not a commercial proposition.

Mickehdj: You post

" The actual facts remain: Software Method C on S21 rigs (10%+ gain) there are 3 makers testing now"

As you say Mickydj, no evidence only the word of one man. Now, you are no fool yet you support Gardin's claim that somehow a slow microcontroller can perform an astronomic 1,000 Billion sha256 input filter operations each and every second. You prefer Gardin's claim to mine, very obviously then a few cheap microcontrollers should be able to outhash an ASIC --- why is the Bitcoin even using ASICs with such stupendous speed available from an ESP32-3?

Your knowledge in this area is lacking I would seek third party arbitration or ask Gardin himself. In my opinion you know its impossible but you are clinging vainly to protecting your excessive investment here but it's turned out to be just another AIM fairy story.

You are learning an expensive lesson here ! Be intelligent and pay heed to the lesson.

Mickeydj:

Gardin has stated that all the big players are aware of QBT. It is credible that any of them could be talking negotiating with QBT, in fact that's what Professor Impossible what you to believe.

Oh and an NDA does not necessarily prevent disclosure of potential partners names. It, we assume, is there to protect disclosure of QBT's secrets. The contents are whatever the parties agree on and sign up to.

Mickeydj: You post

"QBT's software overlays existing rigs, no custom chip needed (as per the software pivot, Dec 5 RNS)."

Are you still peddling Gardin's preposterous and impossible claim that a 240 MHz microcontroller can perform 1,000 billion filter operations per second on a BitAxe miner? This claim is the most bizzare statement Gardin has made.. Gardin also claims it can be loaded up on any miner and presumably he's talking commercial miners.

A Bitmain S21 Pro hashes at 200 TH/s and higher, i.e. 200 times faster than the hobbyist BitAxe therefore the software has to filter 200,000 Billion SHA256 inputs per second, an unbelievable feat for any microproccesor or microcontroller ever built

You need to research what Gardin is claiming for method C software, in fact ask him, I'll bet you don't get an answer.

Any electronics hobbyist will tell you his claim is impossible. Even a Google of microprocessor versus microcontroller will show you microcontrollers are slower than microprocessors.

What his claims also tell us is there is there can be no world class team backing him up otherwise they would be correcting his schoolboy errors. He must be desperate to be making such claims.

.

Usersteve:

Surely Bitdeer know they only have to run QBT's method C software and their problems are solved !

It runs on all miners! Tell Gardin to give them a call and for a few million he can get them out of their pickle !

Mickeydj:

And still no justification for the " batch pre-filtering upstream" fantasy. Your attempt at justifying Gardins fantasy ASIC input filter running on a ESP32 240 MHz microcontroller and filtering 1 TH per second of inputs. Truly the greatest miracle of computer technology this century. And of course this unbelievable technology will run on any miner, presumably those that can hash at 500 TH/s !!!!

Perhaps it was this post of his he wanted removed. It showed irrefutable he created a false and very misleading excerpt from a nonexisting RNS. Below is part of my reply to his ficticious post

"No you haven't, you posted

" "Method C doesn't "keep up" per-hash—it's batch pre-filtering upstream"

This is misleading drivel you made up but claimed it came from an RNS.

𝗕𝗨𝗧 𝘆𝗼𝘂 𝗰𝗮𝗻𝘁/𝘄𝗼𝗻𝘁/ 𝗱𝗼𝗻'𝘁 𝗰𝗶𝘁𝗲 𝘁𝗵𝗲 𝗥𝗡𝗦 𝗮𝗻𝗱 𝘆𝗼𝘂 𝗰𝗮𝗻'𝘁 𝘁𝗲𝗹𝗹 𝘂𝘀 𝘄𝗵𝗮𝘁 𝗼𝗿 𝘄𝗵𝗲𝗿𝗲 𝗯𝗮𝘁𝗰𝗵 𝘂𝗽𝘀𝘁𝗿𝗲𝗮𝗺 𝗳𝗶𝗹𝘁𝗲𝗿𝗶𝗻𝗴 𝘁𝗮𝗸𝗲𝘀 𝗽𝗹𝗮𝗰𝗲.

Absolutely Allforprofit.

Mickeydj wants to hide the truth, he can't contest it so he has it removed.