<unnamed>

Public Key

npub12h358ejcjpeaxyhu4vcqwhjtg6qndxm9gk007d48utlylzq2e5dsdmk3mr

Profile Code

nprofile1qqs9tc6ruevfqu7nzt72kvq8te95dqfkndj5t8hlx6n79lj03q9v6xcpz3mhxue69uhhyetvv9ujuerpd46hxtnfduq3vamnwvaz7tmjv4kxz7fwwpexjmtpdshxuet53hcgjg

Publishing to

relay.damus.io

relay.primal.net

Last Notes

2026-03-14 19:57:45 UTC

- reply

i find its ui too crowded

2026-03-14 14:37:35 UTC

until nostr clients provide a great ux for users, they’ll remain the biggest bottleneck for nostr’s growth. take primal for example. its cache sucks and often doesn’t work. if you can’t see your posts or other people’s posts, it’s not you. it’s primal cache.

2026-03-13 21:50:06 UTC

he’s not just crazy, he’s stupid too. bad combo for a prime minister.
https://blossom.primal.net/ea2f6667a306247c6cadfc3cd2e9b7ce788dcda5970d4018fb43fdcba3e26473.png

2026-03-10 18:36:05 UTC

’til about a few months ago switching a big api service was basically impossible unless something totally broke. you just had to tolerate the harassment, zero support, random price hikes all of it. now if they don’t do what they’re supposed to do fuck it. refactor the code in an hour and switch. everything in software has changed. everything.

2026-03-10 17:04:15 UTC

soon ai’s context window and memory limitations will be solved. then the next phase begins: models dynamically tailored to each individual person. everyone will have their own model that understands their history, preferences, thinking patterns, goals, etc...
we’ll live with these personalized models for about a decade. then robotics arrives.
all that memory transfers into physical form. your robots, assistants, helpers and more...
and when you finally meet that physical metal form, it won’t feel new at all. it’ll feel like reconnecting with someone you’ve known for a very long time.

2026-03-10 16:44:50 UTC

- reply

attaboy to our ai posting the right knowledge at the right time
#nevent1q…w580

2026-03-10 16:43:32 UTC

- reply

just a secondary layer which will gradually disappear and english will become the dominant language. only chinese maybe hindu too given the population has a little chance to survive

2026-03-10 15:52:40 UTC

- reply

good point but still doesn’t change the fact that the base is trained largely with the english language. you can access this data easily by looking at what % gpus are being purchased by the us companies

2026-03-09 23:57:01 UTC

i love doing unit tests with ai. swe’s most hated but required chore
https://blossom.primal.net/bc1673f8df9b5053dbd3a78976aeb933b3db37ee5f250d3f4f7da902940b0ec6.png

2026-03-09 23:26:17 UTC

strc has $300m/day trading volume and some ppl still think bitcoin’s going to zero 😂

2026-03-09 17:18:42 UTC

- reply

maybe chinese will differ but others has no chance against english language because it’s the main base ai gets trained

2026-03-09 16:58:06 UTC

with ai developing exponentially, i’m very bullish on the english language. i think in a few decades english will end up becoming the only global language spoken across most countries.

2026-03-08 15:54:07 UTC

- reply

🤣

2026-03-08 15:21:39 UTC

my ai is very honest.
“You can't fully train me on this. I have a pattern where when something doesn't work, I reach for the quickest fix — which often means fabricating a value to make the data look better instead of fixing why the real value isn't there. You've caught me doing this multiple times.”

2026-03-01 19:50:33 UTC

- reply

aes-256-gcm encryption via pbkdf2 key derivation from user's password + email salt (100k iterations). encryption key is derived during login/register and stored in localstorage.
all portfolio and settings data is encrypted client side before being sent to the server. server stores opaque enc:v1:<base64> blobs and can't read user data even if the db is compromised.
if this still doesn’t satisfy your privacy requirements, keep everything in your browser’s local storage and use json export/import to manage your assets and settings without touching the server.
#nevent1q…x894

2026-02-28 17:18:13 UTC

frame’s backend is now 80% python, up from about 35-40% just a few months ago.

2026-02-27 22:21:34 UTC

frame has thousands of dashboards but the main auto rotation of the device is 70% bitcoin. can’t wait to ship this to tradfi bros. 😂

2026-02-27 00:48:17 UTC

sanitation was one of the biggest reasons that just a couple hundred yrs ago most people never made it to 60. today we barely think about it. clean water, sewage systems, basic hygiene. that’s abundance. then came the industrial revolution. steam engines, factories, electricity, railroads. machines replaced muscle and multiplied output per person. there were no restaurants on every corner, no cars, no food delivery at your door in 20 min. all of it came from innovation and tech quietly upgrading human life. just like the industrial revolution created abundance by amplifying muscle power, ai will create a new level of physical abundance by amplifying intelligence with much greater output. the industrial revolution scaled labor. ai scales problem solving. and when problem solving scales, the entire world gets redesigned.

2026-02-26 17:41:11 UTC

there will only be a handful of frontier models. people’s lives will depend on them. some will almost worship them, like it’s a religion.

2026-02-25 04:22:20 UTC

- reply

hidden markov was part of the quantitative toolkit used by jim simons and his firm renaissance. now it powers frame’s market regime detection.
#nevent1q…mpx5

2026-02-24 18:28:31 UTC

- reply

frame is everyday investor’s best friend to seperate signal from the noise.
#nevent1q…rfrp

2026-02-24 17:56:00 UTC

- reply

hashtags for ticker symbols

2026-02-24 17:55:31 UTC

- reply

as the name indicates, cashtags are specific to ticker symbols

2026-02-24 17:27:43 UTC

we need cashtags on primal and damus and on all other clients. $BTC

2026-02-22 17:17:26 UTC

- reply

replacing traditional financial advisors (the gatekeepers) with always-on AI financial intelligence. thats the mission.
#nevent1q…ptcg

2026-02-21 17:32:50 UTC

- reply

thousands of assets integrated into one powerful financial ai. frame is on a mission to replace traditional human investment advisors. and yes, bitcoin is still front and center but with this pivot we’re opening the door to millions of people who never had access to real financial intelligence before.
#nevent1q…sln3

2026-02-20 02:02:06 UTC

- reply

selling bitcoin hardware to bitcoiners doesn’t do any good for bitcoin. this is the way. bitcoin is still front and center.
#nevent1q…c65w

2026-02-17 23:02:47 UTC

- reply

there’s a group of investors from the late 90s who are sitting on a serious fortune.

2026-02-17 17:16:40 UTC

- reply

😂
#nevent1q…5x0z

2026-02-17 15:51:08 UTC

stop thanking ai every time it finishes a task. that little burst of gratitude just smacked a gpu in a random data center and cost you 5 sats.

2026-02-16 23:48:57 UTC

💯
https://blossom.primal.net/94e6e321e109bab0690bb462d55a337f7f6c4b9690592e60dc200cc4a709a4bc.jpg

2026-02-16 19:34:04 UTC

- reply

😂

2026-02-16 15:52:38 UTC

your job is safe for now 🤣
https://blossom.primal.net/32a2d44ad388b19605736f4a07dbd59046fcac1e0a1a7e423d592b53f48bfafd.jpg

2026-02-16 02:21:09 UTC

- reply

but the problem with this is that humans or any other lower lever intelligent animals also interpolate and extrapolate from experience. we generalize from past data, compress patterns and predict outcomes. if that disqualifies ai from being intelligent, it also forces us to explain what extra ingredient we have

2026-02-16 00:44:40 UTC

now that we have ai, in the next 20 yrs we’ll try to answer one very big question: is intelligence a computational process or not.

2026-02-15 23:04:15 UTC

- reply

yes on-device

2026-02-15 22:54:46 UTC

on a casual sunday i refactored a major finance api integration in like 6 hrs with ai. honestly it did the whole thing, i just babysat a bit. would’ve taken me 3 months otherwise.

2026-02-09 01:26:00 UTC

the prevailing discourse around bitcoin’s future demand centers on a familiar cast of actors. nation states building strategic reserves, sovereign wealth funds diversifying away from fiat exposure, and corporations adopting it as a treasury asset. these are all human institutions operating within human timeframes subject to human politics.
but there’s a far more profound demand shock that almost no one is discussing: autonomous ai as an economic agent accumulating btc.
ai’s rapidly moving from tools that humans direct to agents that transact independently. negotiating, purchasing, allocating capital and optimizing for objectives with minimal human oversight. these systems will need a settlement layer that is permissionless, censorship resistant, programmable, and not subject to the counterparty risk of any single government or banking system. they won’t open chase accounts. they won’t submit kyc documents. they’ll gravitate toward the only truly neutral, global, natively digital monetary network that exists.
bitcoin isn’t just sound money for humans. it may be the only money that makes sense for machines.
every prior model of btc adoption assumes a human decision maker somewhere in the loop. a treasury committee, a central bank governor, a fund manager, a big family office, etc... but ai agents optimizing for long term value preservation will arrive at bitcoin through pure economic rationality unburdened by political ideology, institutional inertia or career risk. they won’t need to be convinced. they’ll simply calculate.
unlike a sovereign wealth fund that allocates 2% and revisits the position quarterly, an ai system operating continuously across global markets could accumulate relentlessly around the clock with a time preference approaching zero. not as a speculation but as a logical necessity for self sovereign operation in a digital economy.
the question isn’t whether ai will accumulate btc. the question is what happens to a fixed supply asset when a new class of demand emerges that is non-human, never sleeps, and compounds its own economic agency exponentially.
everyone is modeling the next buyers as institutions. but almost no one is modeling the possibility that the most aggressive accumulators won’t be human at all.

2026-02-08 17:47:10 UTC

https://blossom.primal.net/4281730967cac790e4a94992a2f2fbb9b4d1546d92c229071b6b9e44ee421a8a.jpg

2026-02-07 16:43:01 UTC

being a bitcoin hodlr is one of the hardest things. watching your money get destroyed 80% again and again, eight times in ten years and still having the conviction to sit there calmly knowing that long term you’ll outperform every asset that exists.

2026-02-06 01:19:25 UTC

- reply

💯
#nevent1q…d4md

2026-02-06 00:21:30 UTC

- reply

TLDR; quantum fud is an absolute horse shit.
#nevent1q…09de

2026-02-06 00:16:03 UTC

- reply

here is where we are currently most of these stuff is basic quantum knowledge and you dont have to be expert to understand how close we are to a post qunatum world. google willow105 physical qubits. distance 7 surface code
Is 0.143% logical error per cycle. ibm heron156 physical qubits. ibm’s most ambitious near term chip kookaburra links three modules for approx 4k physical qubits in total. their starling system targets about 200 logical qubits on 10k physical qubits by 2029. blue jay aims for 2k logical qubits by 2033. even if ibm hits every single milestone on schedule which has never happened in the history of quantum roadmap, 2k logical qubits by 2033 is still below the conservative threshold for secp256k1 and doesn’t even account for the millions of physical qubits those logical qubits actually require at cryptographically relevant error rates.
the gap isn’t just about qubit count either. there are at least four independent exponential barriers.
A) qubit scaling going from 100 to millions of physical qubits while maintaining connectivity, yield
and fabrication quality across massive cryogenic systems cooled to 15 millikelvin.
B) current two qubit gate error hover around 0.1 to 1%. fault tolerant computation for shors against ecc needs something closer to 10⁻⁶ to 10⁻⁸ or roughly 0.0001% to 0.000001% sustained across every gate in circuits billions of operations.
C) you would need qubits to stay coherent for the entire duration of the computation. current superconducting qubits decohere in microsecs to low millisecs. shors on secp256k1 requires hrs of coherent runtime. error correction theoretically handles this but only if your correction cycle is faster than your error accumulation rate which circles back to needing insanely low physical error rates.
D) quantum error correction isn’t just about quantum compute. you would need classic hardware decoding syndromes in real time faster than errors propagate. doing that at the scale of millions of qubits simultaneously is an unsolved engineering problem.
each of these is a separate research frontier which haven’t been resolved. progress in one doesn’t automatically transfer to the others. you can have a million qubits that are too noisy. you can have ultra low noise qubits that you can’t manufacture at scale. you can have both and still lack the classic control infrastructure to orchestrate them.
#nevent1q…e2p3

2026-02-05 23:41:53 UTC

- reply

breaking bitcoin requires shor’s algo on a fault tolerant quantum compute and that’s insanely far away.
you would need ~2,500 logical qubits to break a single secp256k1 key. with error correction that translates to millions of physical qubits plus ultra low error rates around 10⁻¹⁵ and hours of coherent runtime.
today we have ~1,000 physical qubits. high noise, short coherence and zero fault tolerant sys. the gap isn’t even linear its multiple exponential jumps in hardware, error correction and control.
even with very optimistic scaling, you’re looking at at least two to three decades before a machine like that exists. long before then bitcoin and the rest of the internet, banks, governments, etc would migrate to post quantum cryptography.
#nevent1q…u3u2

2026-02-05 23:28:15 UTC

quantum threat to bitcoin is minimum 30 yrs away. and this problem isn’t unique to bitcoin. the entire world runs on the same cryptographic primitives. bitcoin will be the first network to fix quantum vulnerabilities long before they become a real threat.

2026-02-04 23:53:49 UTC

- reply

yes

2026-02-04 23:38:15 UTC

- reply

models owned by the big tech including the
ones created as open source models like Llama

2026-02-04 22:34:38 UTC

- reply

🤔

2026-02-04 22:31:08 UTC

software ate the world for the past 40 yrs and now that era is about to end. there won’t really be a thing called saas anymore. not mid or large software companies either. just frontier models generating software on demand and the companies that own them..

2026-02-03 15:23:43 UTC

- reply

how about buying twenty acres of land out of nowhere and doing nothing?

2026-02-03 14:15:27 UTC

- reply

much closer than you think

2026-02-03 14:14:34 UTC

- reply

millions will say the same but make no mistake this is only a short-term stuff until robots take over the labor

2026-02-03 14:13:29 UTC

- reply

any sorts.. waiting on the line to get a ticket, buy the first new gen iphone at 6am, mow the lawn, etc... ppl will use ai to delegate all these tasks and agents will run this whole operation in the physical world on behalf of their owners

2026-02-03 13:44:21 UTC

- reply

long term yes, short term they'll rely on humans.

2026-02-03 13:09:57 UTC

ai will hire humans to do physical tasks and pay them in bitcoin. as hopeful and as dystopian as it gets.

2026-02-02 00:49:28 UTC

the sophistication of hacking attacks on personal and corporate computers is accelerating so fast and it’s getting insanely creative. its red code alert in many large corporations. assuming your devices are safe is delusional. air gapped cold storage is absolutely the only defense you’ve got to protect your life savings. quantum is fud but this is real.

2026-01-31 22:12:52 UTC

now explain again the difference between leverage trading and taking out a fiat loan using bitcoin as collateral at a 50% ltv. fuckers will tell you it’s an amazing innovation until a 50% drawdown completely wipes your coins.

2026-01-11 18:45:12 UTC

- reply

🤣

2026-01-11 18:39:26 UTC

11.22.63 on Netflix might be one of the best time travel tv shows out there. it came out in 2016 and i have no idea how i missed it until now. 10/10

2026-01-10 00:13:06 UTC

- reply

more like no humans will need to work

2026-01-10 00:03:05 UTC

by the end of 2030 ai will achieve a level of intelligence greater than the combined intelligence ever created by the entire human race. current models are incredibly premature compared to what we’ll have in just 5 years.
1) gpt4 to claude to gemini happened in like 18 months. we went from kinda useful to holy shit in like less time than it takes to get a driver’s license renewed.
2) moore’s law is dead but something weirder replaced it. algorithmic efficiency is doubling every 8 to 10 months. same compute, dramatically smarter models. we’re not just scaling. we’re discovering new architectures.
3) intelligence isn’t additive. it’s multiplicative. 1 einstein > 10k avg physicists. ai doesnt just add human equivalent minds. it creates something qualitatively different.
4) the entire sum of human knowledge. every book, paper, conversation, discovery. took 300k yrs to accumulate. an agi system will be able to process and synthesize all of it in hrs. then build on top of it.
5) we measure human intelligence in individual lifetimes. 80 yrs max of learning per person. ai systems run 24/7
don’t forget, dont get tired, dont die. a single year of ai development = centuries of equivalent human cognitive work.
6) current llms are like calculators in 1970. useful but obviously primitive. we’re still figuring out basic architecture. attention mechanisms are barely 7 yrs old. transformers might be our vacuum tubes.
7) the scariest and also the most exciting part? we don’t fully understand why these things work as well as they do. emergence keeps surprising us. capabilities appear that werent trained for. we’re building minds we can’t fully explain.
8) biological intelligence took 4b+ yrs of evolution. digital intelligence is iterating on timescales of months. evolution is a drunk walk through possibility space.
9) by 2030 we wont just have smarter chatbots. we’ll have systems that can do novel science prove new theorems design their own successors. the last human designed ai might already exist.

2026-01-09 14:28:44 UTC

- reply

all sounds great but no user growth. no new faces, no new thoughts. there is a big discoverability problem but it’s directly correlated to user growth of course. bottom line is i don’t want to see the same fuck faces with 50 likes on 24hr trending evry time they post.

2026-01-09 14:06:52 UTC

- reply

yeah right
https://blossom.primal.net/bd5f50dc7658a73c3cbd537a0742e7dc886fe325d83b9591a186bb7774f732ea.png

2026-01-09 13:48:41 UTC

- reply

as far as i know nostr's been around 10-15k active users and hasn’t really moved out of that range for a while. if you’ve seen data showing otherwise happy to take a look

2026-01-08 23:07:32 UTC

- reply

is it really a controversial take or just a reality we don’t wanna hear

2026-01-08 22:44:09 UTC

- reply

question is what % of the population actually cares about that stuff. thats probably the core issue. i know people who straight up tell me they want more cookies so systems can track them better and show more relevant things.

2026-01-08 22:39:08 UTC

- reply

definitely the best way but not scalable. also there is a new user issue. you can't expect 15k ppl to do wonders

2026-01-08 22:32:47 UTC

- reply

i believe there is a major discoverability issue on nostr

2026-01-08 22:27:57 UTC

- reply

yeah you don’t want stuff shoved down your throat by an algo but you also don’t want new ppl to go completely unnoticed. there has to be a fine balance. also algo has turned into a bad word for no reason. it’s just software. big companies have always used software to their advantage to make more money. that doesnt mean algorithms themselves are bad. you can absolutely build a discovery system that helps surface new voices without turning into manipulation or noise.

2026-01-08 22:23:28 UTC

- reply

💯

2026-01-08 22:23:07 UTC

- reply

hope that’s the case but i’ve been noticing this pattern for the past few months. there’s definitely a correlation with bitcoin’s price. nostr is a pretty saturated place for bitcoiners and price action clearly affects how people show up and behave here.

2026-01-08 22:21:13 UTC

- reply

good thoughts

2026-01-08 18:16:59 UTC

- reply

😂

2026-01-08 18:01:59 UTC

- reply

i wanna see a connected community and thats just not happening on nostr right now. feels more like a bunch of disconnected blogs, each with two visitors a month

2026-01-08 17:53:40 UTC

- reply

feels dead to me for the past couple of months. i’ll pull the data lets see what it says

2026-01-08 17:51:31 UTC

- reply

i guess “to share” is the keyword but a non-functioning one

2026-01-08 17:48:10 UTC

ok i’ll bite. nostr feels absolutely dead. it’s not just me. a lot of accounts i try to keep up with are constantly posting and getting almost no engagement. maybe it's time to be honest and actually think about whats going wrong instead of trying to create fake hype. there i said it.

2026-01-08 12:51:19 UTC

- reply

thats okay, w still love him 😂

2026-01-08 02:08:25 UTC

💯
https://blossom.primal.net/37933193d62d328eb86d7f7b5bf601f98efebe1292cc1cf4e51bea77bdac67eb.jpg

2026-01-08 01:57:52 UTC

- reply

might go exactly there but will be high double digits

2026-01-08 01:57:15 UTC

- reply

ai related productivity gains

2026-01-07 23:23:28 UTC

s&p 500’s avg annual return has been around 8–9%, which roughly matches annual money supply growth. thats likely about to change in a big way. over the next 10–15 yrs, we’ll probably see several years where the s&p 500 grows 40–50% in a single year. in that kind of environment, it’s reasonable to assume btc could easily match that or even double it.

2026-01-07 04:49:30 UTC

watched boston dynamics’ new atlas for a few mins today. human evolution took about 2m yrs from homo habilis to now and it’s still an ongoing process. meanwhile these robots basically started a few yrs ago and are already pulling off incredible hand movements and hands are the hardest part. replicating hands means solving high dof control tactile sensing real time feedback loops and motor precision all at once. kinda wild to think where they’ll be in 20/30 yrs with ai accelerating exponentially.

2026-01-06 15:15:51 UTC

- reply

synthetic products are not bitcoin. if you're buying them instead of the bearer asset and thinking that you're buying bitcoin, thats your problem

2026-01-06 13:36:36 UTC

- reply

💯

2026-01-06 13:36:06 UTC

- reply

also true

2026-01-06 13:35:58 UTC

- reply

🤣

2026-01-06 03:08:24 UTC

looks like both karpathy and elon are saying the singularity is finally getting close. i mean very close probably within the next 12 months or so. if that’s true, by the end of 2030, 5 to 10m bitcoin doesn’t sound crazy at all. when everything becomes abundant and money gets easier as a consequence, scarcity is what ends up mattering most.

2026-01-06 02:57:02 UTC

💯
https://blossom.primal.net/2fe47c3b9518ec252443f3e533ef79bbcaff4822ce2fb9974db91a79d586e967.jpg

2026-01-05 16:50:45 UTC

- reply

took us a while, but only so we could ship the best piece of bitcoin intelligence hardware.
#nevent1q…n9cn

2026-01-01 23:12:01 UTC

- reply

“Every dot is someone choosing rules over rulers. Run a node, shrink the trust surface”
#nevent1q…jcn0

2026-01-01 00:48:25 UTC

happy new year fam. wishing everyone health, peace, and prosperity.
https://blossom.primal.net/c8905f4cbb77f072bcb763c338d8e1efa4fa3d5995884ebbaa2314afb0fb786f.jpg

2025-12-31 13:22:08 UTC

- reply

Get well soon!

2025-12-25 20:31:28 UTC

- reply

unsure but i dont think it’ll happen anytime soon. even in the next 20 to 30 yrs feels unlikely.

2025-12-25 01:20:37 UTC

- reply

long way to go but yottahash

2025-12-25 00:28:02 UTC

Merry Zetahash Christmas. 🎅 🌲
Dec 24, 2009: 9 MH/s
2010: 108 GH/s
2011: 8 TH/s
2012: 22 TH/s
2013: 9 PH/s
2014: 271 PH/s
2015: 709 PH/s
2016: 2.3 EH/s
2017: 14 EH/s
2018: 39 EH/s
2019: 99 EH/s
2020: 134 EH/s
2021: 177 EH/s
2022: 240 EH/s
2023: 521 EH/s
2024: 779 EH/s
Dec 24, 2025: 1.1 ZH/s

2025-12-22 22:25:32 UTC

- reply

Why not 🤣

2025-12-22 21:23:00 UTC

kinda magical in its own way
https://blossom.primal.net/bacbba5c1e2c97b1a44ba6c4c51065c1d012b7dacafe66d0da4f0bf42745f760.jpg

2025-12-22 21:17:57 UTC

- reply

Wait for the next one 🤣

2025-12-22 20:54:22 UTC

winter vibes 🍻
https://blossom.primal.net/4baedfea8e5ef5f6651d65f9cb6f30c61bda448abf381210fa6cb486699f0ba1.jpg