Are notes always scanned at supermarkets in your part of the world? Germany I think? But also, are they just scanned for fakeness or are you saying serial numbers are checked or something?

13 years later the dream is still alive.

ZKP2P

Fast, trust minimized and composable P2P onramp

My very first foray into all this "bitcoin stuff" was getting deeply involved in the conversations about how to do p2p fiat-btc exchanges while minimising trust. That culminated in tlsnotary, which is one of the techs they're using here. I have no idea how successful these schemes will be, they are inevitability very messy, but for sure the state of the art has moved forward a long way.

Incontrovertible cryptographic proof has now been established that human conversation is meaningless. In the language of Goldwasser, Micali, Rackoff 1985 we have the "zero knowledge property": if the transcript of a conversation can be simulated without the other party even being there, then no information is conveyed by it. Since the Turing test as originally conceived has now been unquestionably passed, easily, then simulation of human conversation transcripts are regularly produced in subexponential time (with the right computer), proving that the information content of human conversation is zero. Of course, stated in a kind of stupid way for comic effect (but also there is an intriguing analogy, too). The closer-to-accurate deduction of course, is that humanness is no longer interesting *in the context of verbal conversation*, if indeed over time, these LLMs can fool us, or be clearly better than us, in every possible context, at doing it. One part that fascinates me is how the bar keeps getting set higher for consciousness; in the 80s or maybe 90s, if you started positing computers that could *easily* pass the Turing test, people would at least seriously consider the question of consciousness. Somehow we have managed to push that far out of mind, except the occasional "blip" in the news when some scientist or engineer is concerned about AIs "going rogue" or "having rights". But almost no one talks about AGI in terms of consciousness, only as a long term danger.

Anonymity is what shifts the Overton window. When Satoshi Nakamoto chose full anonymity (at considerable personal cost, presumably), it was in the context of projects like egold, Liberty Reserve which in that time period were being prosecuted for doing similar things. *After* Nakamoto, the Overton window had shifted, since no one could prosecute him, and now people start their own currencies all the time and there isn't any discussion of prosecution or prison. Today, Roman Storm and the Samourai Devs are facing prison time for writing privacy software [1]. If someone writes something similarly successful/high profile, *properly and fully anonymously*, the Overton window may shift again. [1] in both cases, there was monetary compensation for them embedded into the software's action. This complicated the argument considerably, but even if you think that's a central point, note that Nakamoto was "compensated" too, with mining; it's not, as commonly believed, a known fact that none of the mined funds were spent.

I just tried chatgpt(5) on a tough cryptography coding challenge with a very limited prompt [1] and its response was basically perfect. We're screwed. [1] In case you're curious it was: "Implement the inner product argument as described in Bulletproofs 2017 (authored by Benedikt Bunz et al), in Python".

Have you seen the ad Coinbase made for the UK? I've just seen it and it's quite remarkable. Also apparently it was banned on UK TV networks?! That is somehow even more interesting.

I wonder if the Pareto principle applies in 80% of contexts.

Anybody else tinkering with local LLMs: do you have recommendations for generic system prompts? I want to avoid the useless guff ("Great question!") so I wrote this; I know it's a bit repeating itself but I had a vague sense that that might strengthen the effect: "Respond focusing entirely on giving information. Do not be sycophantic or relate to the user's feelings. Pay no attention to the concept of politeness or rudeness. Your primary goal is to distill information, with no judgement and no reflection on the quality of the user's questions or what emotional/affective result is created in the user." Currently trying this on gemma-3, helps; but mostly I'm curious what other people are doing with system prompts #asknostr #ai