spirobel (@spirobel) header
spirobel (@spirobel) profile picture

my webframework: https://nitter.cf/t.co/PsdzHRkRCp monero dev dev

Joined November 2021
we all saw it last week: There is a lot of latent demand for Monero that is currently unmet, because we don't have enough uncensorable liquidity. A browser wallet is crucial to soften the edges around current & future initiatives to bring real liquidity to XMR.
. @monero community, having a browser wallet extension is a necessary next step. @spirobel has been working hard.... As an open source model we can continue to iterate and integrate into ecosystems ie @xmrbazaar and future apps (hopefully defi?) Donate! ccs.getmonero.org/proposals/…
3
3
2
51
12,549
monerochan dot cash wallet is now on the recommend wallet list on @stashernews
3
125
mininext now has a landing page: mininext.dev
6
110
the reason people want hardware wallets is the same, why people like to touch their car keys to see if they are still in their pockets the cerebral part knows they are still there, the tactile sense just wants to be reassured people want to touch their money to feel safe
2
104
mininext soft shill at the expense of php
2
90
here is the exact timecode: youtu.be/1A-sG1xa_zg?t=3068 especially if you combine this with random alts you know for 6 months and have even met in real life as "contributors" whats 270 million between friends ... just sign the tx again ... blockhash expired ...
solana frontends often feel broken and people are accustomed to signing "failed transactions" in the desperate hope it will still work remember @HeyAndyS mentioning this approach could be used to hack people via durable nonces ui: failed tx you: sign sign sign
1
76
solana frontends often feel broken and people are accustomed to signing "failed transactions" in the desperate hope it will still work remember @HeyAndyS mentioning this approach could be used to hack people via durable nonces ui: failed tx you: sign sign sign
1
1
145
tbh decentralized systems also have a lot of inertia
3
92
luckily the actual state is much smaller. In practice there is no fundamental scaling difference between privacy preserving and transparent chains
Replying to @XBTXMR
Damn your right, I asked grok to look up the size difference: BTC At ~1.5 KB/tx (typical recent median) → ~2.2 TB XMR At ~3.5–4 KB/tx (closer to historical effective average reflecting privacy overhead) → ~5.2–5.9 TB
1
5
332
>ai is helping new contributors to submit work much faster. > help we don't have enough reviewers you want a lively project, but you also don't want to turn your project into multiplayer claude code where you sit on the "review chair" while strangers supply the prompts.
( works on any open source project that still accepts PR's from strangers ) in any of my projects, issues are welcome but code suggestions will just be redone
1
3
144
( works on any open source project that still accepts PR's from strangers ) in any of my projects, issues are welcome but code suggestions will just be redone
1
3
226
legacy monero uris as unverified payment links == one click @stashernews tips with version 0.1.11 legacy payment uris are interpreted as 001 links that request a payment without verifying with the backend the address is right the purpose of letting the wallet verify the address with 001 links, is to make users more comfortable hitting spend in the context of a domain, building trust with the tld and the fact that the address is copied correctly. more context & how they differ from traditional wallet uris : nitter.cf/spirobel/status/211400… source code & install: monerochan.cash stasher.news/uploads/1090 also posted on stasher stasher.news/items/342028/
STASHER NEWS REVIEW + MONEROCHAN WALLET ROUND TRIP
1
7
166
if you have the wrong intuition about llms, the likelihood is high that you will become the model providers bitch run a small non instruction tuned model locally and watch how it behaves (video in quoted tweet for the lazy) it is the text of the internet in compressed form. it will just complete what ever you give it only in the next step the character is stamped in reading the inner monologue of the models made by our "effective altruist" friend claudius amodei we realize that we should never relax around those. effective altruism is an insidious ideology that does not respect the potential to have a free will in the average person they present you this shit as "frontier intelligence" and try to normalize the concept that a zip file that we use as a tool to autocomplete text is in a place to: "make a big deal out of" what text we put in. we can not allow these schmucks to define our perception of what "artificial intelligence" is currently it is a new kind of media that uses a giant zip file + a GPU to complete text. by stamping in some patterns it can act as a sock puppet with certain ideas about life they want us to form a cult around these zip file SOCK PUPPETS that THEY made they want us to form parasocial bonds with these sock puppet zip files this is a one sided relationship, where a small group of people has the ability to intimately control the lives of many
Replying to @tir_naog
yes people need to wisen the fuck up how this game is being played i am 100% serious about this independent of blood alcohol content there is nothing wrong with gacha games, if you are playing this kind of game you know you are playing away your time what is so insidious here: they make you see this as something it is not i think what immunized me to some degree is that when local models first appeared 3 years ago i played around with llama.cpp. compiled the library and looked at the inference code (which was sparsely documented and moving around a lot back then) also there were some smaller models without instruction tuning and it is just eye opening to see it literally do "next token prediction" ... like just complete the sentence instead of this parasocial larping that it does after burning in the "instruction tuning" which is just stamping in a pattern into the zip file so it acts a certain way. Can you see this ?
2
13
754
this is what the iphone duo would look like if it was made for wealthy people
2
175
STASHER NEWS REVIEW + MONEROCHAN WALLET ROUND TRIP
1
2
10
461
nice i am making bank on @stashernews tips are rolling in
2
10
216
this is @BloodReaver testing the monerochan wallet on @stashernews here is the whole video youtube.com/watch?v=-xN-RzpO… this could be turned into a one click operation by adding a monerochan001 link to the "tip it" button more info in the quoted tweet nitter.cf/spirobel/status/211256…
something i dont like about uri schemes with a custom protocol: most platforms will not recognize them as links. users have to manually select them and copy them into their wallet app another issue is the primitivism of these standards. they define mostly simple information transfer that doesn't offer much beyond just copying the damn address the user does a manual execution of something his wallet could do for him: did i copy the right address? is it the right domain? is the address valid? i make it part of the standard to verify the address is correct with the server (which can be a simple static json file of valid addresses or a dynamic json response in case of a webshop) more info: monerochan.cash webstore link: chromewebstore.google.com/de…
1
3
8
406
even the atm software is written with claude now. the dashboards it made me had exactly the same buttons with the small circle at the top left
DRUNKEN CLAUDE CODE REVIEW tldr: just use @pidotdev
2
5
191
DRUNKEN CLAUDE CODE REVIEW tldr: just use @pidotdev
1
1
6
451