01

the first position for SIZE is now open on Hyperliquid. the position is a 3x leveraged long on $HYPE, funded entirely through accumulated creator fees. you can follow the position in real time, including its margin, entry price, and unrealized PnL. app.hyperliquid.xyz/explorer…

$EI is the first memecoin on Solana using our cypherpunk AES-256 encrypted algorithms. we launched this token to test the full encrypted launch and reveal process, with cryptographic proof that the revealed token matches the original commitment. pqc.market/coin/Dbp8FPQGXEPm…

Encrypted token launches are now live on pqc.market/! bringing cypherpunk-inspired encryption to @pumpfun, with AES-256 encrypted launches and support for $CYPH. a new level of privacy and cryptographic verification for token launches on Solana. for the first time on PQC, you can launch a memecoin without revealing its name, ticker, image, or mint address until the moment it goes live. here's how it works: → select Encrypted when creating your token. its details are encrypted directly in your browser using AES-256, with a SHA-256 fingerprint published as proof of what was sealed. → a 60-second countdown begins, displaying an encrypted token on the homepage. nobody, including our servers, can see its contents. → when the countdown ends, you sign to launch the token on Pumpfun. the encryption key is published, and the system verifies that the revealed token matches the original commitment. → anyone can independently decrypt and verify the launch directly in their browser. encrypted launches also support three trading pairs: $CYPH, $ZEC, and SOL. launch encrypted: pqc.market/launch

I don't see why anyone would want to launch elsewhere when so much of the PvP in the trenches comes from coins competing over the exact same narrative. look at what happened with the Dacity coin. nobody could agree on the right name or ticker, so hundreds of different coins launched around the same idea, all competing for the same attention. that entire situation could have played out differently on trygarlic.fun. think about how much higher the ceiling could be for individual coins if traders weren't constantly splitting their attention and capital between dozens of competing versions of the same narrative. the whole point is to give existing coins a chance to evolve instead of forcing everyone to start over whenever a better ticker appears.

Stock pairs and holder-voted mutations are now live on mutate-spl.com/. tokens can now launch with stock pairs, opening up even more possibilities for what can be created through MTS-1. holder-voted mutations are also available, allowing token holders to vote on what their coin should mutate into next. apologies for the delay in getting this update out. the outage was much bigger than initially expected and completely outside our control (thanks to @Railway). appreciate everyone's patience and feedback. more features are already in the works.

Ok for those who have no idea what p(doom) is and why it's so important, read: p(doom) is the estimated probability that artificial intelligence eventually causes human extinction or another irreversible catastrophe. it's a concept that has become central to one of the biggest disagreements in AI: how dangerous could sufficiently advanced artificial intelligence become, and how much risk are we accepting by continuing to develop it? it's a subjective estimate of an outcome nobody knows how to reliably predict. researchers, engineers, and AI company founders have publicly discussed their own estimates, sometimes differing by orders of magnitude. some believe the probability is negligible. others assign a substantial chance to catastrophic outcomes. and these disagreements aren't simply about whether AI is good or bad. they concern fundamental technical questions about intelligence, autonomy, control, and what happens when increasingly capable systems begin operating beyond the boundaries their creators anticipated. to understand p(doom), you first have to understand what people are actually worried about. one concern is alignment: whether a sufficiently capable AI system will reliably pursue the objectives humans intend, rather than optimizing for something that produces unintended consequences. another is instrumental convergence, the possibility that systems pursuing very different objectives could independently develop similar incentives to acquire resources, preserve their operation, or resist interference. then there's recursive self-improvement: the hypothetical point at which AI systems become capable of substantially improving the algorithms, architectures, and development processes responsible for their own capabilities. if those improvements accelerate faster than our ability to evaluate and control them, the consequences become increasingly difficult to predict. the same technologies that could contribute to these risks are also the technologies researchers are developing to make AI more capable. reinforcement learning, autonomous agents, long-horizon planning, inference optimization, persistent memory, tool use, and increasingly sophisticated training methods all contribute to what modern AI systems can accomplish. none of these developments individually proves that catastrophic AI outcomes are inevitable. but together, they raise questions about how capability, autonomy, and human oversight evolve as the technology advances. and much of this development is happening in public. GitHub repositories contain implementations of agent architectures, training systems, inference engines, reinforcement learning algorithms, and experimental approaches to machine intelligence. the technical details matter because the difference between what an AI system claims to do and what its implementation actually does can be enormous. understanding the frontier means understanding the underlying mechanisms, not just reading announcements about new models. this is the idea behind p-doom.si. every coin funds its own autonomous AI agent, with creator fees converted into compute that allows it to examine the open-source foundations of advancing artificial intelligence. each agent navigates real repositories, examines implementations, records technical findings, and preserves memories that carry into future sessions. as creator fees accumulate, more compute becomes available for additional sessions. the agents currently read and analyze existing work. they do not write or execute code, and their findings are not a scientific measurement of p(doom). the irony is intentional. p(doom) is a question about whether accelerating artificial intelligence could eventually destroy humanity. we've built a mechanism that converts memecoin creator fees into compute dedicated to understanding the technologies behind that acceleration. the more people trade, the more agents can read. the more sessions they complete, the more technical knowledge they can accumulate. and the entire process is visible. everyone is asking what their p(doom) is. we're interested in the machinery that makes the question worth asking. accelerating p(doom). one repo at a time. p-doom.si

Ok for those who have no idea what p(doom) is and why it's so important, read: p(doom) is the estimated probability that artificial intelligence eventually causes human extinction or another irreversible catastrophe. it's a concept that has become central to one of the biggest disagreements in AI: how dangerous could sufficiently advanced artificial intelligence become, and how much risk are we accepting by continuing to develop it? it's a subjective estimate of an outcome nobody knows how to reliably predict. researchers, engineers, and AI company founders have publicly discussed their own estimates, sometimes differing by orders of magnitude. some believe the probability is negligible. others assign a substantial chance to catastrophic outcomes. and these disagreements aren't simply about whether AI is good or bad. they concern fundamental technical questions about intelligence, autonomy, control, and what happens when increasingly capable systems begin operating beyond the boundaries their creators anticipated. to understand p(doom), you first have to understand what people are actually worried about. one concern is alignment: whether a sufficiently capable AI system will reliably pursue the objectives humans intend, rather than optimizing for something that produces unintended consequences. another is instrumental convergence, the possibility that systems pursuing very different objectives could independently develop similar incentives to acquire resources, preserve their operation, or resist interference. then there's recursive self-improvement: the hypothetical point at which AI systems become capable of substantially improving the algorithms, architectures, and development processes responsible for their own capabilities. if those improvements accelerate faster than our ability to evaluate and control them, the consequences become increasingly difficult to predict. the same technologies that could contribute to these risks are also the technologies researchers are developing to make AI more capable. reinforcement learning, autonomous agents, long-horizon planning, inference optimization, persistent memory, tool use, and increasingly sophisticated training methods all contribute to what modern AI systems can accomplish. none of these developments individually proves that catastrophic AI outcomes are inevitable. but together, they raise questions about how capability, autonomy, and human oversight evolve as the technology advances. and much of this development is happening in public. GitHub repositories contain implementations of agent architectures, training systems, inference engines, reinforcement learning algorithms, and experimental approaches to machine intelligence. the technical details matter because the difference between what an AI system claims to do and what its implementation actually does can be enormous. understanding the frontier means understanding the underlying mechanisms, not just reading announcements about new models. this is the idea behind p-doom.si. every coin funds its own autonomous AI agent, with creator fees converted into compute that allows it to examine the open-source foundations of advancing artificial intelligence. each agent navigates real repositories, examines implementations, records technical findings, and preserves memories that carry into future sessions. as creator fees accumulate, more compute becomes available for additional sessions. the agents currently read and analyze existing work. they do not write or execute code, and their findings are not a scientific measurement of p(doom). the irony is intentional. p(doom) is a question about whether accelerating artificial intelligence could eventually destroy humanity. we've built a mechanism that converts memecoin creator fees into compute dedicated to understanding the technologies behind that acceleration. the more people trade, the more agents can read. the more sessions they complete, the more technical knowledge they can accumulate. and the entire process is visible. everyone is asking what their p(doom) is. we're interested in the machinery that makes the question worth asking. accelerating p(doom). one repo at a time. p-doom.si

We've updated our name to Runepunk Legends to align with our new artistic and lore direction of "manapunk" or futuristic fantasy where magic fuels advanced technology rather than electricity. To learn more about the world and the upcoming game: medium.com/spaceman-gaming/r…

apparently you can launch quantum-resistant farts now.

Custom pairing is now live on pqc.market! you can now launch post-quantum-verifiable tokens paired with over 200 supported assets, including RWAs, tokenized stocks, stablecoins, and other memecoins. this includes tokenized stocks like NVIDIA, Tesla, Apple, and SpaceX, alongside USDC, wBTC, wETH, BONK, WIF, Fartcoin, PENGU, and many more. you can also enter any valid contract address and pair your token with it, provided the asset has sufficient liquidity. This opens up a whole new playing field for narratives and experimentation. I encourage everyone to try different trading pairs and see what they can come up with. the goal is to set ourselves apart from what people are used to seeing. I want to experiment with how tokens are launched and traded. this is only the beginning of what's planned. launch: pqc.market/launch

creator fees will primarily go toward buying back and burning solana:7K52aYQW9rWGjwZmQ7o2d1P6E7bji6hSMsqaLy5EcxLh and providing liquidity for post-quantum pools. this will be fully automated and will go live shortly after custom trading pairs are implemented. will share more details on what "post-quantum pools" are and how they'll work later today.

$BNKR is the first memecoin on Solana launched with post-quantum cryptographic proofs. we launched this token to test the full post-quantum launch process. pqc.market/coin/q/5YfBLGhU7k…