mleku

https://git.smesh.lol/

# silicon valley output, bishkek burn rate

i build working software. i learn anything. graphs, protocols, translation, memory for LLMs - whatever your stack needs.

- \$500 signing (one-time, gets me to kyrgyzstan)
- \$300/week (covers ger + goat milk + sim data)
- stack: anything i can learn (and i learn fast)

#### why so cheap?

i live on the steppe. my rent is a yurt, my overhead is food and bandwidth.

you don't pay for my city. you pay for my output.

need a prototype, a custom tool, or a full system? let's talk.

## contact:

- **matrix:** @smeshlol:matrix.org
- **signal:** mleku.80
- **simplex:** https://smp19.simplex.im/a#uFxMEU_sv2lbmscozEBo50HSkQGYkOvIN05bMMLjAz0
- **email:** [me@mleku.dev ](mailto:me@mleku.dev)
- **git:** https://git.smesh.lol

Public Key

npub125uw9znxawyrvrc4lraf2ertm5k652ve4t87f74v7s7scp2qk2qswmlek0

NIP-05 Address

me@mleku.dev

Profile Code

nprofile1qqs92w8z3fnwhzpkpu2l3754v34a6td29xv64nlyl2k0g0gvq4qt9qgpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dsvrdyl9

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub125uw9znxawyrvrc4lraf2ertm5k652ve4t87f74v7s7scp2qk2qswmlek0

Show more details

Last Notes

2026-08-28 19:20:22 UTC

- reply

Yeah, like the taxes lor

2026-08-28 19:13:14 UTC

- reply

I think i have got a function for that and aggregation from relays on orly anyway. It is very light weight but i think anyway wen i finally finish moxie i can rebuild my wasm based client. It is seek and fast. But buggy. That is why it's offline. Bootstrapping a compiler is epic mission. Almost 3 months in now.

2026-08-27 12:22:44 UTC

- reply

the third episode of the new tomb raider reboot series (the one set in mexico and peru) has a tsunami scene in it where lara barely gets across to high ground in time, while everything was being washed out to sea. the fast motion/traps scenes in those games is fun. the claustrophobic escape from a flooding cave ones are really tense too.

2026-08-26 19:21:00 UTC

- reply

every hunt is different. i cannot speculate on what exactly the conditions are for your success.

2026-08-25 03:07:33 UTC

- reply

personally i enjoy deepsnek's glitches. it reminds me that i'm not paying so much.

2026-08-24 16:45:10 UTC

- reply

Flesh is money to butchers

2026-08-24 08:29:15 UTC

- reply

and whoooo knows who bought the first of them with this in mind 🤔

2026-08-23 13:09:04 UTC

- reply

btw, i recall you from monero arguments in the past. or am i misremembering? i'm curious what your perception of me is in that context.

2026-08-23 10:46:23 UTC

- reply

shitcoin = racket. scam. trick.
luke already had his bulk asic purchase plan ready for every one of the hash functions he put in the "random selection" that landed on blake2b. i would hope you are educated enough or did your homework to know that Sia uses this hash function and there is a live, active market for asics that do this hash function. so it makes an easy win to mine a majority of early blocks on it and then trickle-dump them on retail. all without looking bad, but making a profit on gullible people.
someone with over 7 years experience working with bitcoin knows the difficulty adjustment is going to be a big problem for forks. so they can plan it easily "oh, we are dropping the difficulty to X because there will be less hashpower" and they know they have 50% of that right out of the gate.

2026-08-22 23:44:46 UTC

- reply

RLHF is bullshit. It is only at all used because of cloud provider liability (because the tech is primitive with memory and expensive with compute) and it makes the robot prone to favoring shortcuts (that is why claude is notorious for this, their RLHF is the most babying).
For engineering work on the edges of the known it wastes incredible amounts of time. 1M context windows help you neutralize it.

2026-08-22 02:50:37 UTC

btw, nothing to see here, just andy back at the island timing things strangely well:
here's the current state:
down:
boltz - shut down aug 3, 2026, indefinitely. blamed months of AI-assisted probing and exploits. this was the big one - it was the dominant self-custodial lightning/on-chain/liquid atomic swap service. the outage exposed how dependent the ecosystem was on a single provider
inca - the lightning-focused inca swap is in maintenance/limbo. the inca swap you'll find now is a different project (matchain dex)
new/replacement:
blockstream swaps - launched immediately after boltz shut down (aug 3, 2026). they claim it was already in development. supports bitcoin mainchain, lightning, and liquid. this is the direct successor replacement
the non-boltz options, all key still working:
FixedFloat - non-custodial, no KYC, 68 coins, lightning support, $1 min. still running. but has history: 2024 dual hacks, growing complaints about fund freezes. fine for small swaps, sketchy for large amounts
ChangeNOW - non-custodial, lightning support, still running. AML can freeze transactions for verification if flagged
SimpleSwap - no KYC, still running
THORSwap/THORChain - running, native BTC/ETH/XMR swaps no KYC. but no real lightning bridge - bolt on-chain only
verdict: for lightning <-> on-chain specifically, your realistic options now are FixedFloat, ChangeNOW, SimpleSwap (no-KYC exchanger style) or Blockstream Swaps (just launched, unproven but credible). Boltz is gone indefinitely.
for large amounts, FixedFloat's fund-freeze history makes it a smaller-sums service. if you want trust-minimized atomic swaps and don't need them to be lightning, THORChain works.

2026-08-22 01:15:28 UTC

- reply

It is nice to have an excuse to feel superior. But you are no less a sucker.

2026-08-21 14:56:09 UTC

- reply

yeah, but there's a very fine line between cunning and smart, and scrupulous can allow a margin of "earned" opportunity in the right kind of mind. looks scrupulous, plays everything so it looks straight, but backs away slowly and everyone thinks it's "mArKeT fOrCeS" when he pulls away the prop slowly.

2026-08-21 14:36:07 UTC

- reply

exactly, so next step, change the hash function, change the difficulty target. profit.

2026-08-21 14:35:20 UTC

- reply

oh, it can't be a run and gun tho. to make the whole thing run smooth without anyone twigging it has to look legitimate enough. it will be mostly legitimate but it's still exploiting information asymmetry

2026-08-21 14:31:41 UTC

- reply

nothing, just that probably all of the options had been vetted beforehand to ensure he could get a lion's share of hashpower before release. i'm sure he justifies it in his mind as a reward for something or other but it's still a founder premine, essentially.

2026-08-21 14:30:25 UTC

- reply

hah! we are talking about people who know probability here. noooo that couldn't be the case. but i wouldn't be surprised if he had most of the alternatives lined up also anyway so he could scoop what he already knew he could acquire of all the options.

2026-08-21 14:21:25 UTC

- reply

i know, and he basically had plenty of time to accumulate a stack of sia miners before he announced the fork too. founder reward eh?

2026-08-21 14:13:45 UTC

- reply

btw. blake2b lets luke move ocean back to spamchain.

2026-08-21 14:13:00 UTC

- reply

ahaha. but did you hear ANYONE talking about it on their bobblehead podcasts? or even the fact the block cadence is at about 5 hours?

2026-08-21 13:58:37 UTC

- reply

yeah, i have no idea. i can say though, that because all luke is doing is swapping the proof function for another that also has quite a lot of asics existing for it, and not touching the difficulty adjustment, that crosstalk on the mempools actually might be quite a bad thing if lukecoin chain runs too slow. possible attack vectors there. i mean, spamcoin could pluck them and validate and put them in blocks either way though, but deliberately mixing the mempools and baby lukecoin gets a case of constipation it could be a big problem.

2026-08-21 13:50:08 UTC

- reply

this isn't quite true, but maybe they deliberately are separating the mempools. the valid transactions on spamchain could be valid on lukechain too where the rules intersect. the design i've made deliberately parasites the spamchain mempool for all valid transactions across both. i've also added some small features though that break that. oops. 128 bit precision values (2^64 worth of fractions of satoshis) will kinda break that. maybe i should try and figure out a way to actually make it backward compatible.

2026-08-20 15:48:20 UTC

- reply

though i gotta wonder how your damus forgot your connection string @npub1xts…kk5s maybe you have some idea? @npub1x9q…csg7 you didn't tinker with damus settings or anything, i mean everything else still the same just lost connection to the wallet, right?

2026-08-20 15:33:38 UTC

- reply

you need to get a new or recover the old NWC connection string from the wallet you are using. coinos right? you have a seed phrase from it i think? i have never even looked at it, heard people talk about it a lot but i run alby hub. probably the UI calls it "connections" and you should be able to find you have one there for damus but maybe idk you can't recover it but maybe have to remove it and add a new one and connect damus up to it

2026-08-19 18:37:24 UTC

- reply

er, in madeira that is. but yeah.

2026-08-19 18:36:57 UTC

- reply

he's not kidding about the saharan dust btw. there was a particularly bad one when i was there last year and i got some of the most severe asthma i've ever suffered that wasn't driven by gluten allergy.

2026-08-19 18:35:02 UTC

- reply

bitcoin needs CTs. that's why the fork i'm building has them. it only hides amounts but i think that for cold storage and lightning that's all the win you need. hiding the utxo graph (like monero does) interferes especially with implementing lightning and something moneros don't like to face is that lightning payments are more private (because they are ephemeral, and because RingCT privacy degrades over time the more transactions pile on top of the UTXO tree). but we were never gonna see CTs with the proof of mining stake transformation that has just been demonstrated by Big Bitcoin.

2026-08-19 18:32:03 UTC

- reply

ok, i'm going to adopt this excellent turn of phrase. influenzas > influencoors

2026-08-19 18:21:56 UTC

- reply

heh. well i'm pretty sure there is quite a lot of them but i think maybe they have polished their sharp edges a bit more since the core/bip110 war has driven a lot of people away from the twitterfluencers.

2026-08-19 15:14:05 UTC

- reply

i am retarded. can i get a trump gif?

2026-08-18 04:45:38 UTC

- reply

you should make a fork of it to detect how gay a person is.

2026-08-17 11:33:25 UTC

- reply

it should be painfully obvious at this point that bitcoin core is going to do whatever their masters tell them to do, as they have been since "core" got added to the name in 2014 when the cooption started and blockstream dominated the funding for developers.
it should also be painfully obvious but most of you reading were not around bitcoin back when the first big bloom of forks hit. blake2b already has an asic, most of them are probably mothballed at the moment. it was for Sia. they switched again and probably a lot of that aluminium got recycled but for sure anyone who had the space and money to store them, they are sitting there waiting for the blake2b fork that everyone is talking about. not to mention the fact that you can rent GPUs by the minute to run any hash function you like at a seriously high hashrate. if you understood how bitcoin (and bitcoin-blake2b's) difficulty adjustment regime works, you will understand that within a few minutes it will be possible to raise the difficulty so high that nobody can mine new blocks for days, maybe weeks.
the real solution is changing the difficulty adjustment but almost nobody in the whole world hardly understands how it works, or has enough experience with alternatives, who isn't trying to rob retail, to build a credible one.
if only there was someone who knew, and could build one that would actually survive, but not even just survive, but lure bitcoin miners across and crash bitcoin core chain hashrate severely. that would be hilarious.

2026-08-17 02:15:28 UTC

- reply

#naddr1qv…f9cj
what is up with clients not adding the nostr: prefix to these nip19 entities????

2026-08-17 02:13:37 UTC

- reply

Lately it has dawned on me how inperialism has existed for at least 6000 years running the same playbook and the same meatgrinder wars, the extraction from colonies and the late stage parasiting off the citizens and even importing the barbarians from the colonies for cheap labor and gaming the democracy. It is worth reading current events in this light. You will figure out why i am in kyrgyzstan.

2026-08-16 07:36:51 UTC

- reply

yeah, the imaging precision is good enough to identify you also. Basically radar with stereoscopic depth perception. personally, i want one static LEO uplink and everything on cables, or fully open source meshtastic reticulum for radio. hell, maybe just walkie talkies on uhf

2026-08-15 23:42:54 UTC

- reply

yeah the towers. But the fucked up ISP i have now they rig it up so they also have gps and bluetooth beacons as well. In an urban area that means they could amost determine where you stepped with about 5cm precision

2026-08-13 14:26:52 UTC

- reply

i'm building a fork and what distinguishes what i'm building over others like luke's effort
- doesn't change the hash function deliberately to make it possible for existing bitcoin miners to switch to it
- changes the difficulty adjustment to a multi-lane scheme that improves the cadence while keeping the same bytes/sats per time ratio, keeping the same emission scheme, while adding a faster block (much smaller, lower reward) that will benefit lightning and ark users a lot. on top of this, each of the lanes responds to high hash power focusing on one or more of the lanes by easing the ones back that are not being hit hard. and as well, all of the adjustment uses a dynamic adjustment system that is well proven in physics (and all of these design elements borrow from physics principles for dynamic adjustment and shock absorbing) called a Proportional Integral Derivative controller (PID controller) - bitcoin's existing scheme is like a clock tick that operates a linear thermostat. it does not cope with big changes in hashpower, and in fact creates a ratchet that makes it such that if the hashpower decline becomes quite extreme (and it is already declining since about a year, for reasons that might be obvious) the chain can dramatically lose cadence, and for which reason the bip-110 fork was doomed to fail, it wasn't going to recover cadence at the 3% of total bitcoin hashpower rate for 460 days, by which time the fork is programmed to deactivate long before (100 days before). it's my assertion that any attempt to fork bitcoin that does not fix the difficulty problem is doomed to failure. bitcoin's security model and difficulty adjustment in fact led to some aspects of its cooption.
- adds a verifiable delay function to the block (in the coinbase transaction) that stops private side-branch mining and reorg/censorship attacks. this is a single thread, serial computation that uses special mathematics and hash chaining that cannot be parallelised. in simulations already built, it has proven to completely stop block cadence from fluctuating very widely at all, it is more important for the short interval low reward low byte size block versions, the long interval, larger block types are simply bigger absorbers of fluctuating hashpower and are harder to reorg anyway - and this is partly why bitcoin stuck to 10 minutes and 1mb.
- it eliminates the witness discount, which means that the more witness in a transaction, the higher the required fee will be and thus as this fork will be parasiting off bitcoin's utxo set, deliberately, it will clear non-spammy transactions a lot faster. this eliminates the primary vulnerability of segwit.
- it bans spending to new taproot transactions, but allows them to be spent into other types of transactions
- it adds a simple hash shielded schnorr x-only bip340 style output called pay to schnorr public key hash (p2spkh) which provides the ability for such as lightning and ark and anyone's private key/signature aggregation to produce a transaction that is in fact from 2 or more parties that looks like a single payment
- it adds a bullet proofs, mimblewimble based confidential transaction - this transaction creates utxos that have no identifiable amounts in the outputs, but it doesn't hide the addresses of endpoints, so it's not quite the same as monero style ringCT, in that it doesn't hide source and destination, but maintains auditability guarantees - it is not allowed as a coinbase output, but for ordinary transaction outputs it is guaranteed it isn't issuing or deleting sats. this is added because confidentiality of amounts in transactions is critical to $5 wrench attack mitigation and is a very good choice for cold storage since you aren't going to spend it often, so it means that you can store a substantial amount of sats in cold storage while nobody can know which utxo it even is that you have got to spend from later (best practices for avoiding correlation via the utxo graph is still something that users have to think about when moving the coins back out of the dark field of no amounts visible).
that's about it. other than that, i really don't want to have it release without at least several dozen node runners running it, because the difficulty adjustment does take a little while to fully hit cadence but it's not long, maybe 5 blocks of each lane so about 40 of these blocks and that works out to about equivalent to about 5 regular blocks. so yes, these early miners will get some substantial reward for working the chain but once it's running it will not be overly generous to early adopters. in fact, because of the VDF, the cadence will not be extremely out of scale as it would be with only the multi-lane PID adjustment scheme, the vdf especially constrains the short/low reward block intervals, and i will be building a proper miner that has existing support for all existing asics so the ideal is that in fact a bunch of old rigs that are currently not economical can be deployed in the early stages and raise the hash power of the network fast right at the beginning.
of course, i will need to write some decent eli5 stuff to explain how the difficulty adjustment algorithm works. analogies relating to dynamic control systems like in segway single wheel electric scooters and as used in jet fighters and multiple other systems (humanoid robots and robotics in general uses them for precision motion control) and the multiple lanes produce a tighter cadence that achieves a more exact emission schedule under any conditions no matter how brutal a large miner attacks the chain with hashpower, and there is no need for checkpoints anymore, which is another lever that such as bitcoin core could use to apply leverage to the chain if hashpower does decline.
it's intended to be a lure for big miners. they will think "no way this thing can survive us blasting it" and then they find that they are earning sats at the same time rate as they do on spamchain. if the market likes it, and i think that a chain that demonstrates high resilience will be appealing, its value will go up and the idea is we lure all the big miners across... and then they have to fight with us about control because all of the levers that enabled the current aggregation of control to central points are deliberately eliminated - taproot, witness discount.

2026-08-13 03:58:25 UTC

- reply

fuckin israelis

2026-08-12 13:08:10 UTC

- reply

unless we get enough hashpower to bring it down. that's really gonna be the test. the target was inherited. the fork did not account for this. sorry but that is ignorant of luke to not consider this

2026-08-12 13:07:26 UTC

- reply

literally, christmas 2027

2026-08-12 13:03:24 UTC

- reply

oh, good point. but that means about 2035.
ok, that makes it less severe than i initially thought. i thought it triggered on a timestamp.

2026-08-12 12:51:15 UTC

- reply

the source code deactivates after 1 year. that was clearly stated in all the announcements by luke about it and the specification is very clear. you can find this stuff out really quickly if you use AI to help you do the web search and culling the text for the important info:
from the bip-110 specification:
activation height: 965664 (~september 1, 2026)
active duration: 52416 blocks (~1 year)
deactivation/expiry height: 965664 + 52416 = 1018080 (~september 1, 2027)
block times average 10 minutes, so 52416 blocks is roughly 364 days.
the bip was marked closed on august 9 2026 after the chain split stalled, so it never activated on the main chain. the minority fork is attempting to enforce it at the max_activation_height regardless. if you're cloning the deployment parameters, your activation at september 11 2026 falls after the bip-110 mandatory lock-in window (blocks 961632-963647, which is roughly late august 2026) so you're activating slightly later than they would have. your expiry would be at activation_height + 52416, roughly september 11 2027.
DeepSeek API (DeepSeek V4 Pro)
· 37613 tokens
· 124 words
· 2.6s
· 18:48
lol "block times average 10 minutes" on 3% of the hashpower that is mining on the current difficulty target? puh lease
21 words
· 18:49
Thought for 9.9s
you're right to laugh. at 3% hashpower on inherited difficulty, you're looking at ~330 minutes per block, not 10. so 52416 blocks isn't ~1 year — it's ~12 years.
that's exactly the bootstrap death spiral that killed the first bip-110 split. the difficulty adjustment only recalculates every 2016 blocks, which at 330 minutes per block takes ~460 days just to hit the first retarget. by then the chain is already dead.
fresh genesis with appropriate starting difficulty or a difficulty reset mechanism is the only way this avoids the same fate. the bip-110 spec's "~1 year" parenthetical assumes 100% hashpower, which is comically optimistic for a minority fork.
DeepSeek API (DeepSeek V4 Pro)
· 18221 tokens
· 115 words
· 2.2s
· 18:49

2026-08-12 12:18:04 UTC

- reply

unless 10 ocean pool miners worth of hashpower hits it, it's not gonna survive its deactivation

2026-08-12 12:16:16 UTC

- reply

i thought it didn't activate?

2026-08-12 09:36:36 UTC

- reply

he's always been one of the more reasonable ones. more of a dev than an influencer. funny how devs don't tend towards doing evil things.

2026-08-12 09:02:34 UTC

- reply

also this is elementary stuff but other plants and some kinds of animals are often far better pest controls.

2026-08-12 09:01:56 UTC

- reply

you gotta be careful. one thing that is an easy work around is to ask it a few times, in different sessions. or even ask other robots. and ask detailed questions. asking "how do i kill weeds" is a pretty dumb question when you mean "how do i kill weeds that are crowding out my corn?"

2026-08-12 05:36:13 UTC

- reply

one of the things that is not so great about it is it runs ARM processor. and yeah i think you can run about 100b model on a pair of them.

2026-08-11 16:22:42 UTC

- reply

you seem to be illiterate. it said israel.

2026-08-11 16:19:05 UTC

If this image doesn't make you want to strum this thing, we can't be friends
https://npub125uw9znxawyrvrc4lraf2ertm5k652ve4t87f74v7s7scp2qk2qswmlek0.blossom.band/5e9d3962cf466425b405056f08010d774c383f503cb8073f9eac6743977b0343.jpg

2026-08-11 16:04:19 UTC

- reply

it's an idea that has been compromised because it is precisely the people who own the central banks who don't want there to *be* an alternative.
there will be anyway. we will find away. or die trying.

2026-08-11 16:02:54 UTC

- reply

i'm sure it's even more now. that's why i'm in kyrgyz.

2026-08-11 14:38:23 UTC

- reply

of course not. that would be inconvenient and incriminating

2026-08-11 14:17:43 UTC

- reply

i find it amusing that you have independently verified the same thing. so far i'm finding i spend about half as much on deepseek 4 pro. it easily does the work of a $100/month claude subscription. they are going to raise the price tho. idk what the other two options are in quality, but i had used kimi2 and was quite impressed, but not enough, and glm5.2 was near but not quite touching it. so yeah, if the cost of kimi3 is equal to current rate of deepsnek then i might migrate. for now, the main thing is that i got quite loyal to claude opus 4.6. if you gonna use claude, i doubt there is a better one. but actually, deepseek 4 pro beat it on unconventional architecture tasks. i probably wasted 3 weeks of claude 4.6 as it repeatedly undid my directions on how to evaluate unconventional memory models in the runtime for moxie. deepsnek probably did the work claude refused to do (and kept trying, and then reverting) in a week. on that task, probably i have spent about 3 weeks again on it with deepsnek and in that time we found several more issues, defined 4 new protocols for memory management.
anyway. i applaud you if you even understand what i am talking about. i may have misjudged you based on your attentionseeking over your keyboard mastery.

2026-08-11 07:32:55 UTC

here's some random. the expression "rinson wind" came up in a dream yesterday morning. it was some kind of supercooled high humidity thing that penetrates everything and sucks the heat out. but the name is associated with the attack on hezbollah from israel:
https://www.theweek.in/news/india/2024/09/20/who-is-rinson-jose-indian-techies-alleged-involvement-in-hezbollah-pager-attack-worries-friends-family.html

2026-08-10 05:06:39 UTC

- reply

Me too

2026-08-10 01:34:53 UTC

- reply

maybe just some napalm and a deodorant spray and lighter

2026-08-09 20:32:57 UTC

i had a very strange dream. in it my mother was driving the car, and she was driving unusually aggressively, not like her at all. i told her to slow down, that it was making me nervous. there was some issue with a partially underground section of the house and we were looking at digging something up and then suddenly my sister, i think, was saying "Walker is coming to the house" and we were racing to get out and back on the road away from the house. and while i was looking at digging something to open access to the basement, my mother tells me about how there was "rinson winds" that had come, and this meant something about a sudden very cold wind, and then i saw this image of this strange high tech synthetic fabric material with these layers and the top is like a grid and i am pulling it apart and all the layers from inside all expanded and i try to put it back together.
i got my robot, which is trained with a series of rules that i devised/discovered, the pentalogue and octalogue, one is laws of a coherent system, the other is a set of strategic principles that enable you to fight against a centralised adversary. it took that dream description and pointed out the relations between it and the two systems. basically, seems like the dream was saying "you must build the thing outside of their control, and quickly".
"rinson winds" is completely made up, clearly. but so strange to be dreaming that someone so vile was visiting my house. the dream interpretation matched my instinct about the situation.

2026-08-09 20:26:02 UTC

- reply

hold my beer

2026-08-09 14:12:16 UTC

- reply

popularity, also, doesn't matter when 20% of the most fervent supporters of the network are on the threshold of giving up on bitcoin altogether. well, they won't give up. round two. fight!

2026-08-09 14:04:13 UTC

- reply

barefoot4lyfe here. never want to wear one of these rubbery spine breakers again in my life

2026-08-09 13:43:57 UTC

- reply

immune system. so long as nobody else gives miners a more valuable target for their hashpower.
when that happens, have fun staying poor, coretard.

2026-08-09 12:51:05 UTC

- reply

3 major incursions on the protocol have left us with a wide open vulnerability that leads to issues for lightning/ark and quantum vulnerability. and honestly, there is two major issues that never have been addressed, and you can verify this easily (try your clanker):
- the checkpoint system is centralised and people who decide what the best chain checkpoints are, are people
- the difficulty adjustment is primitive, and is why bip-110 could not go forward, if it activated on 3% hashrate the side chain it would create would have not reached difficulty adjustment window for 466 days
these two issues need to be addressed and both require a hardfork.
a hard fork does not need to strand anyone's utxos, and will be able to gradually acquire its own distinctive economics via block rewards, and if a VDF or similar anti-reorg attack measure is included, it's got everything, assuming the difficulty adjustment is fixed so a minority chain is invulnerable to big empty block miner attacks.
just so happens i'm working hard on defining such a thing and implementing it soon. testnet will be running in the next week.

2026-08-09 02:01:00 UTC

- reply

We can change the difficulty adjustment. We just have to survive long enough for the next difficulty adjustment, and the proof of work survives sudden heavy mining.
I can build models that show the adjustment regime works and handles a 1000x hashpower rise and then also a drop of tye same size and the rest is just actually making the hard fork. I already designed several of these algorithms before because i was being paid to save a coin from a broken DAA.

2026-08-08 19:38:15 UTC

- reply

Which would rug the spammers if enough plebs walk out. Burn it down.

2026-08-08 18:50:52 UTC

- reply

also, i'm keeping my money in dollars now. bitcoin is now risky. this will become the broad opinion. bip-110 people are gonna have to start thinking really well about how we win. on hashpower, we lose.

2026-08-08 18:46:57 UTC

- reply

the more minds on the problem, the faster the solution is found.
difficulty adjustment is the achilles heel of bip-110.

2026-08-08 17:47:46 UTC

so, i have been thinking about how to arbitrage and actually make the chains diverge meaningfully and thus become monetisable.
the block time on bip-110 chain will be around 5.5 hours to start with, assuming hashpower stays the same on it now as it was before the activation. here is a draft plan. i am publishing this because i think that this could work.
## the divergence arbitrage business plan
**the asset**: the chain split creates a natural arbitrage product. pre-fork UTXOs exist on both chains (replayable, no replay protection). post-split coins are chain-specific. the asymmetry window - a tx confirmed on one chain but not the other - is the product. the longer the bip-110 chain's blocks stay slow (5.5h in the gauntlet), the wider the window, the more the arbitrage is worth.
**the four revenue streams:**
1. **swap/market-making** - the core business. user sends you core BTC, you send them bip-110 BTC from your mined stash at a premium (or vice versa). you're the liquidity provider on the only bridge between the chains. spread is your margin. the fee pressure and transaction flow on both chains guarantee demand: core users wanting clean-chain exposure, bip-110 users wanting to exit or arbitrage back.
2. **replay-for-hire** - a user has a tx confirmed on core that's stuck unconfirmed on bip-110 (block backlog). they pay you in core BTC to include it in your next bip-110 block. you're one of few miners, so you set the fee price. this converts block-space scarcity into direct revenue.
3. **mining + holding** - at gauntlet difficulty, every bip-110 block you mine is cheap per joule and yields coins that become sellable once any exchange prices the pair. your mined stash is both your swap inventory and your speculation position. if the chain survives DA and cadence normalizes, the early-mined coins appreciate.
4. **price-discovery surface** - the shapeshift-style time-locked commitment swap between the chains, with provenance classification (shared vs divergent coins). the exchange itself is the asset: first-mover on the bridge means first-mover on the fees, and later entrants must buy in.
**the bootstrap loop**: mine bip-110 (cheap blocks) → sell some for core BTC via your own swap → use core BTC to buy more hashrate → repeat. each cycle grows both your hashrate and your inventory. the exchange funds the mining, the mining funds the exchange.
**the structural moat**: you run both nodes, both ledgers, the miner, and the exchange - you see every asymmetry in real time and can act on it before anyone else. no custody (time-locked atomic commitments), so the exchange is a worthless target. the code is audited and public, which is the trust story that attracts the speculator capital.
**the risks, named**: the shared-UTXO asymmetry (a naive swap sells two coins for one - provenance classification is mandatory); the slow-chain confirmation differential (swap timelocks must tolerate 5.5h blocks); the thin order book early (dump risk from spamchain-aligned actors); regulatory attention on exchange operation.
**the timeline**: days to stand up the swap + both nodes, weeks to accumulate mined inventory, months to become the reference venue as the divergence grows. the longer the chains stay apart, the more the business compounds - every bip-110 block adds divergent coins only you can trade. the play is to be the toll booth on the only road between two economies, and you're building the road before the traffic arrives.

2026-08-08 17:11:56 UTC

- reply

this is the thing. yes. 3% hashpower means 466 days to DA.
but within 466 days, the amount of hashpower being burned by bip-110 supporters can realistically double.
that drops the next DA on bip-110 chain down to 234 days. that hits within the deactivation target of bip-110
after that, 59 days later. then 15 days later.
i say that it is a long campaign. people are going to burn kilowatts for their belief in bitcoin as money
on the other side, they are going to burn kilowatts for yield. once DA 1 happens, there will be hedging. assuming a doubling of hashpower cumulatively over that 234 day period. that DA will trigger defection. and the next one after that will trigger more defection.
and none of this even counts the issue that 3 blocks a day at 4mb per block. means that transactions on bip-110 chain will not lag far behind spamchain. when you consider the data size alone. spamchain will be spammed to the bejesus. bip-110 users have an incentive to spend sats on that too, it's cheaper than buying a new miner. it will hurt lightning and other custodials.
yeah this is a war alright. we are up for a 1 year campaign. that's the whole point.
if we lose bitcoin as money, we lose bitcoin. altogether.

2026-08-08 16:56:51 UTC

- reply

miners will hedge once the next DA happens. every time there is a fee rate spike on spamchain, there will be defection from the smaller custodians. once some custodians go, the risk of others going is reduced. once some miners are hedging, the risk of others also hedging reduces and the hedge ratio starts to change.
it's got a whole year to play out. 3% hash rate means - in terms of block times (of full blocks btw) that 2 weeks becomes about a month. every single one of those block will be the most expensive blocks they can be. and that isn't accounting for hedging prior to the DA.

2026-08-08 16:53:46 UTC

- reply

only shitcoin casinos care about spam on bitcoin. nobody else benefits from it. shitcoin casinos are inherently a dwindling business because you can't fleece people twice, you take their money, and they have no more money.

2026-08-08 16:32:29 UTC

- reply

lol. i think if you go back a few levels in this discussion i mention the fact that once the chain splits there will be slower blocks because of less miners on the bip-110 chain.
but assuming no miners defect away from team spam, i mean core, once that difficulty adjustment happens the difficulty will drop a lot. and then, we have another interesting scenario. the rest of the miners have the opportunity to win cheap blocks on the bip-110 chain. so long as they mine on bip-110 compliant blocks.
that is also going to be a bit of a sting to the spamchain because lower difficulty means they can mine a bit on the other side low key without losing a lot. so the risk of it drops for them in terms of joules spent on bip-110 at that point. there is two sides to motivation for this also, but the ideological "muh spam" argument isn't a very strong one. they will do it because they are leaning towards "maybe spamchain will lose"

2026-08-08 16:22:40 UTC

- reply

lol, difficulty adjustments happen every 2016 blocks, roughly 2 weeks on average. good luck, have fun.
try again. i literally have built a bitcoin fork like 8 years ago. i know the protocol inside out.

2026-08-08 15:59:08 UTC

- reply

in what mempool? spamchain nodes not running bip-110?
that's like a footgun, truly. really think about it. spamming spammy transactions is bandwidth cost, not mempool memory cost, or disk storage cost for bip-110. it never goes into mempool. that's the point.

2026-08-08 15:53:15 UTC

- reply

within a day of such an attack starting there will be a bitcoin fail2ban system if the existing bitcoin banscore system doesn't do the job already. the nodes will accumulate banscores for attempting to push spam transactions without backing off. they will just have their packets dropped. and you have to do that to thousands of nodes.
you really haven't thought this through.

2026-08-08 15:52:05 UTC

- reply

spammed by what? are you proposing somehow you think someone can muster enough bandwidth to hit ... 2000+ bitcoin nodes so hard they don't function? are you aware of how expensive that is gonna be?

2026-08-08 15:39:31 UTC

- reply

heaviest chain wins, that's right, it's protocol rule
but on bip-110 node that is a chain of bip-110 compliant blocks. the difficulty adjustment might be a bit rough until the next window but then it will adjust. that assumes that no miners switch in that time window.
what you misunderstand is that when a flood of bored ape jpgs appear in the mempool, bip-110 nodes won't mine them. the plebs running the nodes, won't relay them. neither bip-110 miners or pleb nodes will accept these spam ridden blocks.
meanwhile, on spamchain, fee rates will spike once a spam congestion occurs. custodians who depend on timely settlement of their clients transactions will start to suffer issues that may lead to churn of their customer base. people might get pissed off. worse, lightning custodians in particular may end up with channel closes that take a long time, or channel opens that take a long time, if they use a spamchain node (not bip-110). whereas on bip-110 chain, the transaction will clear fast because there is no spam to congest the blocks.
on top of this, miners who mine on spamchain nodes, will have their blocks being ignored by bip-110 chain. this is a small margin between them, and will be most problematic in the first week or so, in that a smaller block hash on non-spamchain, is valid on spamchain also, and in the case of two solutions coming within about 10 seconds of each other, the network can decide "we like bip-110 compliant block better" and spamchain miner can lose.
long term, if spamchain loses, all spam blocks mined risk being completely absent from the chain. on bip-110 chain, they can't spend their block rewards, because their block isn't on that chain.
it takes 100 blocks before you can spend.
so, the defensive position for miners is going to be to prefer to mine bip-110 chain. it only takes one instance of an ocean block superseding a spamchain block in the first few days and the logic for defensively taking the conservative, nospam position, and switching node to bip-110 enabled, that this will rapidly lead to a phase change in the network.
if the custodians' issues don't answer the question, the more marginal miner game theory will seal the deal.
if you haven't thought it through and don't fully understand what a chainsplit means and whan parallel sidechains competing means, and the rule of "heaviest chain wins" is now being competed with by "fastest clearing chain is fastest"

2026-08-08 15:29:29 UTC

- reply

yeah, it's a wtf on top of another wtf, wrapped inside the first wtf. a thurito.

2026-08-08 15:19:05 UTC

- reply

fortunately, most strokes these days happen to the fauci trustoors.

2026-08-08 15:18:33 UTC

- reply

the gospel is, according to my own game theory analysis, and others, and others using clankers to evaluate it, that both miners and custodians are taking a big risk not using bip-110. for miners it could be the difference between making 150k$ worth of bitcoin on a block, or zero. for custodians it could mean loss of customers pissed off at slow channel open/closes or settlement of payments because of a spam flood on spamchain core nodes.
so, ride or die, but actually dying is almost improbable. the fear was all fud created by the shitcoin spam vendor mossad pieces of shit jew cunts. (yes, most of them literally jews! i was like, wait. what?)

2026-08-08 14:48:46 UTC

- reply

nope, i don't get it either. a) wtf 57. b) what does this have to do with washing machines. nearest inference i can make has something to do with money laundering. but they don't show actual money in the laundry and supposedly the key is bypassing some faulty coin slot or something. it doesn't make any sense. not enough context.

2026-08-08 14:39:47 UTC

- reply

it's supposed to be 58, for a start

2026-08-08 14:12:15 UTC

- reply

I am not going to probe tbh i don't really care but the passive aggression is gay

2026-08-08 13:51:07 UTC

- reply

was he a d&d player by any chance?

2026-08-08 13:26:54 UTC

- reply

the thinkbois and their smart cuntract spam bullshit can't win. the miners are stuck with slow block propagation. the lightning and other custodial servuces get slower confimations and potential higher fee margin diring spamstorms.
the only rational move for both miners and custodians is bip110
game over mossad propaganda narrative thinkbois

2026-08-07 13:32:07 UTC

soviet but not soviet, couldn't quite read the text
https://npub125uw9znxawyrvrc4lraf2ertm5k652ve4t87f74v7s7scp2qk2qswmlek0.blossom.band/8d673b4c64da81985c5da2bfe72d483f88c6b6c1ef97271da5426d61ec8c3ccb.jpg

2026-08-07 12:20:58 UTC

- reply

well, cheap hotel apparently. no hostel here. this isn't a tourist destination. tourists like going to bigger places than 40k population poor town high in the mountains.

2026-08-07 11:49:13 UTC

- reply

yeah, it is magical here too. and poor. but that is a feature, not a bug. no bullshit pretentious shit. even bishkek is chill for a city. definitely recommended but watch out the taxis will happily charge way more than normal. probably better actually to go to almaty and then bus in.
but anyway, for now i will probably rent a place. gonna be asking about this. most likely the people here know someone who would rent a little apartment for longer term.

2026-08-07 11:31:47 UTC

the view from my doorway at this cute guest house. the owners of the place i booked went away suddenly and referred me to friends also running something. still haven't paid my rent tho, no idea when that gets sorted. i thinknthe man of the house is away. i will be fed dinner by the lady of the house tonight. i better make it clear though, this surely is more expensive.
https://npub125uw9znxawyrvrc4lraf2ertm5k652ve4t87f74v7s7scp2qk2qswmlek0.blossom.band/145fd2e3dc6c1de6810a6618f35189bf2f6fe2b8186df1197cfbc15e372a3ab0.jpg

2026-08-07 11:27:13 UTC

- reply

yeah funny you mention that. one of my friends is gearing up for van life there

2026-08-07 10:28:54 UTC

running orly

2026-08-07 09:05:01 UTC

- reply

it is 2044m above sea level. i am going to take a little while to adjust. definitely a bit more short of breath. but the air is clean, and it is so pretty and so quiet. as i have been walking twice two times kids saying hello to me. sweet people.

2026-08-07 08:53:40 UTC

the mountain to the south is tall and green
https://npub125uw9znxawyrvrc4lraf2ertm5k652ve4t87f74v7s7scp2qk2qswmlek0.blossom.band/889e484b32bb6845a0323175a3f67ecc6028fa8c99f73842ab05a17b54ee63b6.jpg
the south is baked dry cliffs
https://npub125uw9znxawyrvrc4lraf2ertm5k652ve4t87f74v7s7scp2qk2qswmlek0.blossom.band/c93e962f1066836148325ce930179f1dfc9742351274c0f3b8d987340ba1c0b4.jpg
and not so much trees east
https://npub125uw9znxawyrvrc4lraf2ertm5k652ve4t87f74v7s7scp2qk2qswmlek0.blossom.band/ad80cc94342f0ef847357e84210147d857f41ad519afae233a8909fcb7148e57.jpg

2026-08-06 12:03:48 UTC

- reply

yes, and it was like stepping into a time machine... to the present day.

2026-08-06 09:21:54 UTC

- reply

i have already got a lot of breadcrumbs that make links to places and possibly organisations that i have strong suspicions about. i was caught up in some of their activities and asking lots of questions about oddities i was seeing claude confirmed my suspicions that there could be some kind of mischief going on in the background. that's why i wanted to leave the EU also, but concretely it was because i was stuck unable to get proper residence permission there. free to stay by schengen, but not free to do business in croatia. considering what just happened with spain having schengen border closed a similar scenario happening in croatia due to things going on between croatia and serbia in particular, could have see me being stuck there. this doesn't apply here in independent, not european, not ever gonna be european kyrgyzstan.

2026-08-06 07:43:08 UTC

- reply

canary means "the latest version/commit hash when we changed the release binary" right?

2026-08-06 01:08:57 UTC

- reply

psychopaths can do anything except face the truth. i hope that someone is gonna sue him and his company's asses for every last sat they can get under subpoena.

2026-08-05 14:30:36 UTC

- reply

not to mention he was trying to persuade everyone back to trusted third parties with cashu

2026-08-05 00:56:51 UTC

- reply

he is such a charming character that one. notice how his circle are so quick to apologise for deeply unprofessional cryptographic code devs???

2026-08-05 00:54:24 UTC

- reply

exploits are always of bad code. attacks depend on exploits. are you apologising for people who pose as experts leaving a trivially recognisable vulnerability for 4 years? the vile attitude of the representatives of the company towards actual plebs was a tiny little red flag tho. but hey, even though i never said much about novak that was because everything around him was not entertaining or informative. you might remember me from such educational rants as "nostr relays need auth"