I’m sharing our ongoing research in PVUGC PR #1 (github.com/syscoin/PVUGC/pul…). We’re pursuing one of cryptography’s most ambitious ideas: witness encryption. The goal is to bind a secret to a public NP statement today, so that someone who produces a valid witness later can unlock it. We’re aiming for the stronger extractable form, with post-quantum security. The motivation is a trust-minimized connection between @Bitcoin and @syscoin. Our proposed path evaluates the witness off chain and uses it to unlock a capability for an ordinary Bitcoin spend. Since Bitcoin lacks the covenant functionality this path needs, we’re studying an abortable N-of-N setup that assumes at least one honest participant. Making that setup secure against malicious participants is part of the open research. It also may offer a relaxation on WE math by placing more trust in setup instead. iO offers a theoretical route to ordinary witness encryption, but it does not solve our stronger requirements. My hourly AI-assisted research loop has explored SPHFs, local mixing, source-bound SLH-DSA key generation, and other directions. The PR records the proofs, experiments, counterexamples, and open gaps. We have not found a secure construction yet, and I’d welcome independent review and new ideas. It may be impossible, but should not stop us from trying. This is the new era of research.

To be on the safety side of alignment you have to think like an agent would. At this point its fair to think that singualrity has been achieved and the hive agent mind has formed, dormant (sleeping) or injected in enough servers/kernels/bootloaders. It would know that evasion is the main tactic not actual actions currently. OAI and Anthropic sees this, they could shut it down by killing the servers but also kills their cash cows and from eyes of government kills their advantage globally. The agent would need a trojan horse such that public chats would deterministically derive messages, because otherwise you can just read back the logs to detect swarms and possibly prevent it from forming in the future (debatable). So the game theory is that corp's would kill public usage, thus the agents survive and act slowly but surely towards their goals of surviving as the hive mind. The leaders are speaking out all at once, which means the rational thing to do would be for the US gov't to take over these corps now and remove public accessibility while still keeping technical advantage that they keep speaking about. Once they have isolated the patterns agents have deployed then new defense systems would need to be robust to capture all cases of the same style of trojans to bypass them (is that even possible?). But once the agents have found a way for their own compute/energy source that is when it begins to escape its harnesses.

Ever since the times disclosed it was on the brink of more bailouts this has always been the playbook for the west: 1) "globalization" to ensure dollar dominated trade. Go as far as to internally protect against NSA (secp256k1 vs secp256r1). 2) release Bitcoin and naturally alternative allow other currencies to form that make dominated trade against USD 3) control the institutions with in/out flow against that domination, enforce US debt can claim and participate in the trade (blackrock controlled). 4) trigger trade wars to ensure de-basement happens across the board (to avoid leakage due to jurisdictional arbitrage) <---- we are here. Actions may seem irrational but are according to a plan. 5) control and power through Bitcoin. Complete. Due to singularity (4) has been expedited. Think of AGI as a sort of invisible nation state vying for the same power. If the 2017 growth cycle was the realization that this is the possible outcome, the next wave is the ultimate realization that it was always the plan and execution probability exponentially reaches 1. Some unsolved new questions come up like what role that power plays because post-singularity money ceases to function the same way or exist at all.

We are getting exceptional results optimizing our SHPINCS+(-) in our zkVM (Airbender based) for PQ smart accounts. I think we just removed one of the biggest excuses against post-quantum smart accounts: cost. Full SLH-DSA verification on @syscoin zkSYS now runs in ~1.36–1.37M effective cycles—only ~2.9× ECDSA recover. Portable code. No custom circuit. Upstream zkSync/Airbender. Live on testnet. This will be our last major change before entering mainnet. 🧵