Hi, I’m Murch.
I’m working at Chaincode Labs on Bitcoin projects full-time. Among other thinks, this year I’ve:
  • reviewed and authored some PRs for Bitcoin Core mostly around coin selection, transaction building, fee bumping, and relay improvements
  • spent almost there weeks on a technical review of Mastering Bitcoin 3rd Edition
  • co-authored a weekly series on Mempool with Glozow in the Optech Newsletter
  • contributed to development discussion around Silent Payments and Cluster Mempool
  • given talks at the Mempool Summit in Nashville, TABConf, BTC Azores, and Bitcoin Research Day
  • co-hosted the Optech Recap almost every week
  • co-hosted the Socratic Seminar of the BitDevs NYC
  • recorded some ten or so podcasts
  • been avidly contributing to and moderating Bitcoin Stack Exchange
  • sometimes commented on the current mempool weather
I did another AMA in February.
Is it possible to create a valid transaction that nodes running bitcoin core can't relay? (i.e. no amount of configuring mempool settings will let them add it to their mempool)
reply
Yes, there are multiple policy rules that are not configurable. You would have to change the code and recompile to accept infringing transactions to your mempool.
A simple example would be a transaction with two OP_RETURN outputs
reply
Could you mimic the effects of a soft fork solely by tightening mempool policies?
reply
Not really. Making mempool policy stricter may eventually cause blocks to exclude transactions infringing on the new rules as hashrate updates. However, as long as just a small portion of the nodes (perhaps as little as 2–5%) run with the old rules and at least one miner has not upgraded, the transactions would still appear in blocks.
On the other hand, a softfork enforced by a majority of the hashrate would actually reorganize offending blocks out of the best chain since it would actually be a consensus rule rather than a policy rule.
reply
There are groups with a significant financial investment in bitcoin -- MicroStrategy is probably the most important one, given the size of their stack, and their prominence in the broader world.
Aside from just buying bitcoin and siphoning up supply, how could a company like MSTR best spend money to grow the wealth of their bitcoin holdings?
reply
I’d hope that they thoughtfully invest back into the ecosystem. When you hold enough bitcoins, you should recognize that contributing to development, education, and advocacy might have a greater payoff in wealth via what you already have than accumulating more instead.
I’m not intimately familiar with all they do, but I’ve seen some mention of education programs and advocacy for mining at least, so it seems to me that this is happening at least to some degree.
reply
I have a question from somebody else, that you may help. I didn't have enough practice and knowledge on this scenario so I take the advantage of this AMA.
How you could extract a proper xpub from a multi-sig wallet using 2 different hardware wallet devices? I know that each HW will have a different derivation path.
Let's say I want to setup that multisig, but using a watch-only xpub only for depositing into that wallet, without "opening" the multisig.
So how I could be sure I will deposit to the right address from that wallet without "open" it? For a single sig I know it's easy and simple, but how is going in a multi-sig scenario?
reply
My understanding is that a xpub only keeps track of a single chain of keys. Therefore, you would need keep track of multiple xpubs to reconstruct the addresses of a multisig wallet.
You should look into output descriptors, they can keep track of complex output scripts and a single output descriptor can express a full multisig wallet.
reply
What would scale your work by 10x? What would have to happen?
reply
I’m not great at prioritizing the most important task. I guess if that were different, my Bitcoin Core contributions would move a lot quicker, but on the other hand, I’d probably not have ~1900 posts on Bitcoin Stack Exchange and done over 8300 edits to posts there. I’d probably also not be posting on SN and Twitter as much, or co-host BitDevs. I’d probably be a more effective contributor if I were working more closely with a few people on one project rather than trying to do a dozen things at a time.
I like to think that some of the educational work I’ve done to help on-board people has led to more work being done on some projects overall than if I had worked on them directly.
Let me know if you have some good ideas how I could scale my impact, I’d be curious about an outsider perspective on that one.
reply
Hey Murch,
given we are near the end of the year, and since you listed a bunch of things you accomplished during the last 12 months, my 2 questions are:
  • What are your plans for the next year? any planned surprises or anything you are looking forward to do/release.
  • What do you think will be the most important things/events that will happen in 2024, related to the bitcoin ecosystem?
reply
I aim that 2024 will finally be the year that we get rid of the Knapsack coin selection algorithm in the Bitcoin Core codebase. I assume that there will be a few good parties around the halvening. I’m hoping to see Silent Payments, Cluster Mempool, and Package Relay (hopefully including Package RBF) get released. I don’t have any big surprises up my sleeve.
None of the above will be the most important things, though. I would guess that regulation of Bitcoin will be a focus in the next couple years which very well could have a huge impact on the overall direction of the project.
reply
thanks, looking forward to see what happens.
reply
Are you still climbing? If so, bouldering, trad or lead?
reply
The cadence has been a bit low lately, but it’s still bouldering for me.
reply
No questions, just want to say thank you for all those insightful answers on Bitcoin Stack Exchange
reply
Thanks, I’m glad you enjoy them
reply
How would you approach getting a soft fork merged from assessment to activation?
How would you you recruit reviewers for a new BIP?
reply
"If you want to build a ship, don't drum up the men to gather wood, divide the work, and give orders. Instead, teach them to yearn for the vast and endless sea."
It’s my observation that a softfork proposal only starts gaining momentum when there is a clear understanding of the benefits it provides and it builds a coalition that staunchly wants it to happen. Try to find passionate co-sponsors and hone the message explaining the benefits of the proposal.
reply
When are you going to buy a proper microphone?
reply
Oof. Thanks for the feedback. I guess I could use some advice on how to proceed.
For Twitter Spaces, I had tried using a microphone with my phone, but learned that I could not output sound to a bluetooth headset at the same time. I was relegated to using the speaker of the phone to listen. Since I sounded so poor when I tried to use the headset as input and output, I have now been using just the phone directly. I got some remarks that it sounded better and nobody had complained since, so I assumed that it was at least okay now. Since I’m not using my headset these days anyway, I guess I could try to use the microphone again. I’ll do some testing before the recap tomorrow. I had tried to use my computer for Twitter Spaces, but unfortunately the web interface doesn’t allow you to raise your hand or show other speaker’s raised hand and that made the interaction with guests significantly worse.
If you are referring to my sound quality in the recent Nodesignal podcast, that’s mostly my fault. I was going to borrow my colleagues high-quality microphone, but asked to late, so I ended up just using my laptop’s microphone.
Do you happen to know of a microphone setup that is compatible with an Android Phone, but would also allow me to listen to the Spaces with headphones? I’d be happy to throw some money to improve the Optech Recap, since I actually do that every week, and I’m sure any microphone I’d get for my phone would also work with my computer.
reply
Murch could you tell the stackers some fun facts about the mempool that most people don't know. You once told me some really good ones related to Satoshi at TabConf.
reply
Could you give me a hint? I don’t recall what you are referring to.
The only overlaps of Satoshi and mempool that come to mind right now are that the original purpose of the sequence field was not incentive compatible and that there used to be a portion of the block reserved for transactions with high coin-age priority. You could get a transaction into a block without paying a fee if the transaction had a high number of bitcoin-days-destroyed.
reply
ya that was it
reply
Great :)
reply
You've been actively involved with Bitcoin for several years now. I assume that you've build quite the knowledge graph. How do you maintain it? Do you have a note-taking process?
reply
I recently rediscovered the concept of a Zettelkasten and have been using a note taking program that caters to it for the past two years. I also realized that Bitcoin Stack Exchange is basically a multi-user Zettelkasten on Bitcoin knowledge. Given my strong involvement, this is where most of my knowledge graph resides. I also frequently search the Bitcoin Optech topics to look up at least cliffnotes on topics, and have become more comfortable just grepping the Bitcoin Core codebase to verify factoids I remember.
reply
That's awesome! While I'm relatively new to Bitcoin Core, I've been leveraging a Zettelkasten system since day one of my contribution journey to gather and store context around Bitcoin Core. This led me to envision the immense potential of having access not just to my knowledge graph but also to those of more established, experienced contributors. It's intriguing to consider the productivity boost and depth of insights such shared access could provide.
Bitcoin Core is a dense and complicate subject with multiple overlapping knowledge bases: the codebase, GitHub, PR Review Club, Bitcoin Stack Exchange, Bitcoin Optech’s topics and newsletter, Bitcoin Optech’s topics and newsletter.
Efforts like bitcoinsearch make that easier, but as a user of a Zettelkasten system, do you see a potential for an enhanced, communal external memory akin to a Zettelkasten?
reply
I mean, Bitcoin Stack Exchange is an interlinked knowledge-base that literally collects answers of many collaborators to various questions. ;)
I do often wonder, though, how much more useful BSE could be if many of the questions that Bitcoin and Lightning users, developers and enthusiasts ask each other on various Slack channels, discords, personal messages, forums, and GitHub repositories every day bubbled up to BSE.
reply
How much is Bitcoin adoption limited by lack of awareness/knowledge in average people vs the user experience of wallets/software having a bunch of rough edges?
reply
20% to 80%? I’d say the majority of people have heard about Bitcoin at least a few times now, and people are using credit cards just fine without understanding how they work. We got a lot of rough edges to work out and Bitcoin has to solve some actual problem for people for them to start using it.
reply
Assuming Bitcoin adoption is at the bottom part of an S-curve, roughly how many decades do you think before adoption reaches the middle of the curve?
reply
Oof, no idea. Another 25 years?
reply
Thanks. As I learn more about the state of Bitcoin software/protocols, my estimates have grown too.
reply
Thanks for your work on stackexchange
reply
My pleasure
reply
What podcasts are you on?
reply
I co-host the Chaincode Podcast and the Bitcoin Optech Recap. Besides I have at least been on Blocktrainer and Nodesignal (both German) this year.
reply
This has been a very busy year for you, I suggest next year will be grand
reply
Hah, thanks
reply
What are the chances Udi is financially incentivise by public miners to push Ordinal/ BRC-20 narrative so they make mining more profitable (through drastic increase of the tx fees)?
How much time you could sustain pumping jpegs with $7.5M funding?
reply
I don't know
Edit: I don’t have any special insights into Udi’s motivation and have no idea how long I could. Honestly, I don’t want to spend time on pondering how I’d even go about it.
reply
I love this simple answer, especially from a brilliant mind like you. Intelligent people know to say "I don't know" when they really don't know.
reply
Oops, I just saw your answer after editing. Thanks, though
reply
Anyways, I love your answer.
reply
What is your programming experience? What is your favorite programming language?
Is Rust overhyped?
reply
I had little programming experience before starting to study computer science. I started working as a part-time programmer during college and have been working in developer rolls since graduating. So at this point I have been getting paid for programming at least part-time for ~14 years, although only a few of those years that was my main job activity.
I used to really like Scala, but I haven’t used it in years. I mostly use C++ these days, but don’t particularly like it. Rust sounds very attractive, but I haven't gotten around to playing around with it much. There is a best tool for any task, but often the best tool is what people already know, so I find the "everything must be rewritten in Rust" stance a bit exasperating.
reply
over the next four years, will the sum of bitcoin transaction fees be larger than the sum of block subsidies?
reply
I sincerely doubt it. Even after the upcoming halving the subsidy will be 3.125 ₿ per block. To achieve this, we’d need blocks to remain full for the next four years and transactions to bid an average feerate of more than 312.5 ṩ/vB. My expectation would be that the inscription and rune hype will eventually abate and while those uses may continue to buy any "left-over" blockspace, I doubt that they will be pushing feerates as high as we’ve seen last week continuously.
reply
Who would win if we can chance to organize a fight between bull and bear?
reply
Let's see, a (Brown) Bear is a Challenge Rating 1 Large Beast, and a Bull is a Challenge Rating 1/2 Large Beast. I haven’t rolled the dice, but I guess that the Bear would win.
reply
But we know the truth.
Now more realistic question. Let's imagine we are waking up and Bitcoin is the protocol of money every country use. How do you think how many sats an average male/female will have in their saving accounts?
300k sats will be unreal to have, right?
reply
Given an upper bound on the money supply of 2.1 quadrillion satoshis and a global population of 8.1 billion people, the average amount per person would come out as 259k sats. However, money tends to be distributed rather unequally, so presumably 300'000 sats would put you among at least the top 10% bitcoin hodlers. Economics is not really my expertise, though.
reply
Anyway thanks for responses and for your work to make Bitcoin better
reply
Do you know an approximate ETA for channel factories? Could be next year?
reply
I am not aware of anyone working on them explicitly, so it’s hard to give a timeline. I’m not the most plugged into Lightning Dev these days, though.
reply
Hey Murch! What do you think of the current debacle with Ocean, and the OP_RETURN debate? Seems like there are a lot of heated discussions about which datacarriersize size is "correct", what constitutes as censorship, and what are it's implications regarding Bitcoins future. Are you on any side of the argument?
reply
It’s been 80ish bytes in Bitcoin Core since 2014, which Luke at least acknowledged—even if it had been 40ish bytes in Knots then and has been ever since. Given that more than 99% of listening nodes are running Bitcoin Core, I don’t think it’s much of a debate that the current broadly accepted datacarriersize value would be Bitcoin Core’s default value. I don’t think it’s a good use of our time to engage with Luke’s unique perspective on this one.
reply
Can I earn the same mount of money in bitcoin open source development than in another "normal" industry like software developer at a car manufacturer?
reply
I would say that many of us earn less than we could make working in a normal industry or even at Bitcoin companies. Many of the grants come in around the amount that entry-level engineers make in NYC or Silicon Valley. On the other hand, we get to work on something we are passionate about with a huge portion of agency which is also worth a lot. I have heard of a few positions that make competitive amounts.
reply
Thanks for the insight.
reply
If you were not an engineer, what would you do to support bitcoin?
reply
I’d probably spend more time on advocacy and informing regulators
reply
I see the Optech posts but they too technical for me to understand
reply
I’m sorry to hear that. The Optech Newsletter aims to inform people working in our industry, so it naturally tends to be pretty technical. We do try to provide a lot of background information via e.g. the topic pages, and you can find more information on many different questions on Bitcoin Stack Exchange, but I could see how it might be a bit hard to get far enough into the topic to get much out of our newsletter. If you are just starting to dive into the technical aspects of Bitcoin, maybe a book like Grokking Bitcoin or Mastering Bitcoin 3rd Edition would be more accessible.
reply
What did you have for breakfast today?
reply
Yogurt and a cup of coffee
reply
What is the maximum price you expect of Bitcoin when the last satoshi is mined?
reply
I have no idea
reply
What do you think of the recent OP_TXHASH PR? Do you think there's a reasonable chance at it getting merged?
reply
I’m still affording myself to abstain from the whole introspection debate—it’s not my wheelhouse and while the people that are really excited about it are still bashing in each other’s heads to agree what we are even looking for, I am not sure it’s worth having an opinion. I see the point of vaults and eltoo as applications of introspection, but most of the other things that have been floating around so far don’t strike me as ways to advance the technology and protocol to support bitcoin the currency.
reply
Why don't mining pools have a backup block template filled with lower paying transactions ready to counter a stale block?
reply
I assume you mean an empty block.
The biggest challenge with that would be that the miners would still need to verify that none of the transactions from their staged template already got included. It’s easier to just shoot of an empty block template immediately as a new work instruction and then to start following up a fraction of a second later with some actual templates. Altogether, empty blocks have become pretty infrequent, so I’m not sure another approach would be worth the investment.
reply
would you rather fight 100 duck-sized horses or one horse-sized duck?
reply
If I'm allowed to prepare and equip myself one horse-sized duck. If it's spontaneous as I am, the 100 duck-sized horses.
reply
got any work for svelte developer?
reply
Sorry, I do not
reply