08: Navigating the Vibe: AI, Collaboration, and the Future of Creativity

By No Solutions April 22, 2025 · Edited April 5, 2026

"With the shift towards this multi-agent collaboration and orchestration world, you need a neutral substrate that has money, identity, cryptography, and web-of-trust\ baked in, to make everything work." <a href="https://njump.me/@PABLOF7z">Pablo</a> & <a href="https://njump.me/@Gigi">Gigi</a> are getting high on glue nostr. \ Books and articles mentioned: <ul> <li><a href="https://www.goodreads.com/book/show/36656180-saving-beauty">Saving beauty</a> by Byung-Chul Han</li> <li><a href="https://wattenberger.com/thoughts/llms-as-a-tool-for-thought">LLMs as a tool for thought</a> by Amelia Wattenberger</li> </ul> In this dialogue: <ul> <li><a href="https://github.com/dergigi/vibeline">vibeline</a> & <a href="https://github.com/dergigi/vibeline-ui">vibeline-ui</a></li> <li>LLMs as tools, and how to use them</li> <li>Vervaeke: <a href="https://www.youtube.com/watch?v=A-_RdKiDbz4">AI thresholds & the path we must take</a></li> <li>Hallucinations and grounding in reality</li> <li>GPL, LLMs, and <a href="https://danb.me/blog/llms-licensing/">open-source licensing</a></li> <li>Pablo's multi-agent Roo setup</li> <li>Are we going to make programmers obsolete?</li> <li>"When it works it's amazing"</li> <li>Hiring & training agents</li> <li>Agents creating RAG databases of NIPs</li> <li>Different models and their context windows</li> <li>Generalists vs specialists</li> <li>"Write drunk, edit sober"</li> <li><a href="https://dvmcp.fun/">DVMCP.fun</a></li> <li>Recklessness and destruction of vibe-coding\ </li> <li>Sharing secrets with agents & LLMs\ </li> <li>The "no API key" advantage of nostr</li> <li>What data to trust? And how does nostr help?\ </li> <li>Identity, web of trust, and signing data\ </li> <li>How to fight <a href="https://nostr.band/?q=%23aislop">AI slop</a></li> <li>Marketplaces of code snippets\ </li> <li>Restricting agents with expert knowledge\ </li> <li>Trusted sources without a central repository\ </li> <li><a href="https://zapstore.dev/">Zapstore</a> as the prime example\ </li> <li>"How do you fight off re-inventing GitHub?"\ </li> <li>Using large context windows to help with refactoring\ </li> <li>Code snippets for <a href="https://olas.app/">Olas</a>, <a href="https://github.com/nostr-dev-kit/ndk">NDK</a>, NIP-60, and more</li> <li>Using MCP as the base\ </li> <li>Using nostr as the underlying substrate\ </li> <li>Nostr as the glue & the discovery layer\ </li> <li>Why is this important?\ </li> <li>Why is this exciting?\ </li> <li>"With the shift towards this multi-agent collaboration and orchestration world, you need a neutral substrate that has money/identity/cryptography and web-of-trust\ baked in, to make everything work."</li> <li>How to single-shot nostr applications\ </li> <li>"Go and create this app"\ </li> <li>The agent has money, because of <a href="https://nips.nostr.com/60">NIP-60</a>/<a href="https://nips.nostr.com/61">61</a></li> <li><a href="https://ppq.ai/">PayPerQ</a>\ </li> <li>Anthropic and the genius of <a href="https://modelcontextprotocol.io/docs/concepts/tools">mcp-tools</a>\ </li> <li>Agents zapping & giving SkyNet more money\ </li> <li>Are we going to run the mints?\ </li> <li>Are agents going to run the mints?\ </li> <li>How can we best explain this to our bubble?\ </li> <li>Let alone to people outside of our bubble?\ </li> <li>Building pipelines of multiple agents\ </li> <li>LLM chains & piped Unix tools\ </li> <li>OpenAI vs Anthropic\ </li> <li>Genius models without tools vs midwit models with tools\ </li> <li>Re-thinking software development\ </li> <li>LLMs allow you to tackle bigger problems\ </li> <li>Increased speed is a paradigm shift\ </li> <li>Generalists vs specialists, left brain vs right brain\ </li> <li>Nostr as the home for specialists\ </li> <li>fiatjaf publishing snippets (reluctantly)\ </li> <li>fiatjaf's <a href="https://github.com/fiatjaf/blossom">blossom</a> implementation\ </li> <li>Thinking with LLMs</li> <li>The tension of specialization VS generalization\ </li> <li>How the publishing world changed</li> <li>Stupid faces on YouTube thumbnails\ </li> <li>Gaming the algorithm\ </li> <li>Will AI slop destroy the attention economy?\ </li> <li>Recency bias & hiding publication dates\ </li> <li>Undoing platform conditioning as a success metric\ </li> <li>Craving realness in a fake attention world\ </li> <li>The theater of the attention economy\ </li> <li>What TikTok got "right"\ </li> <li>Porn, FoodPorn, EarthPorn, etc.\ </li> <li>Porn vs Beauty\ </li> <li>Smoothness and awe\ </li> <li>"Beauty is an angel that could kill you in an instant (but decides not to)."</li> <li>The success of Joe Rogan & long-form conversations\ </li> <li>Smoothness fatigue & how our feeds numb us\ </li> <li>Nostr & touching grass\ </li> <li>How movement changes conversations\ </li> <li>LangChain & DVMs\ </li> <li>Central models vs marketplaces\ </li> <li>Going from assembly to high-level to conceptual\ </li> <li>Natural language VS programming languages\ </li> <li>Pablo's code snippets\ </li> <li>Writing documentation for LLMs</li> <li>Shared concepts, shared language, and forks\ </li> <li>Vibe-forking open-source software\ </li> <li>Spotting vibe-coded interfaces\ </li> <li>Visualizing nostr data in a 3D world\ </li> <li>Tweets, blog posts, and podcasts\ </li> <li>Vibe-producing blog posts from conversations\ </li> <li>Tweets are excellent for discovery\ </li> <li>Adding context to tweets (long-form posts, podcasts, etc)\ </li> <li>Removing the character limit was a mistake\ </li> <li>"Everyone's attention span is rekt"</li> <li>"There is no meaning without friction"\ </li> <li>"Nothing worth having ever comes easy"\ </li> <li>Being okay with doing the hard thing</li> <li>Growth hacks & engagement bait</li> <li>TikTok, theater, and showing faces and emotions</li> <li>The <a href="https://en.wikipedia.org/wiki/1%25_rule">1% rule</a>: 99% of internet users are Lurkers</li> <li>"We are socially malnourished"</li> <li><a href="https://vertexlab.io/">Web-of-trust</a> and zaps bring realness</li> <li>The semantic web does NOT fix this LLMs might</li> <li>"You can not model the world perfectly"</li> <li>Hallucination as a requirement for creativity</li> </ul> Recorded at <a href="https://mempool.space/block/890848">890,848</a>. <itunes:subtitle>#08: Navigating the Vibe w/ Pablo</itunes:subtitle> <itunes:duration>5728</itunes:duration> <link>https://njump.me/nevent1qvzqqqqqqypzpx77gg2frul26xkzr0gaq9n842u50axpcjhdsa3yeu388vrv5pftqys8wumn8ghj7mn0wd68ytn9d9h82mny0fmkzmn6d9njuumsv93k2tcppemhxue69uhkummn9ekx7mp0qqsdpepvgml06dcsfuuy0x0jhka9jtwpvqrh5ue6qxmtva2xje7n9aqs5852f</link> <pubDate>Wed, 04 Apr 2025 10:21:21 +0000</pubDate> <author>Gigi</author> <itunes:author>Gigi</itunes:author> <itunes:explicit>No</itunes:explicit> <itunes:keywords>nostr, development, sovereign engineering, shipping, memes, web, internet, conversation, dialog</itunes:keywords> <itunes:summary><![CDATA["With the shift towards this multi-agent collaboration and orchestration world, you need a neutral substrate that has money, identity, cryptography, and web-of-trust\ baked in, to make everything work." <a href="https://njump.me/@PABLOF7z">Pablo</a> & <a href="https://njump.me/@Gigi">Gigi</a> are getting high on glue nostr. \ Books and articles mentioned: <ul> <li><a href="https://www.goodreads.com/book/show/36656180-saving-beauty">Saving beauty</a> by Byung-Chul Han</li> <li><a href="https://wattenberger.com/thoughts/llms-as-a-tool-for-thought">LLMs as a tool for thought</a> by Amelia Wattenberger</li> </ul> In this dialogue: <ul> <li><a href="https://github.com/dergigi/vibeline">vibeline</a> & <a href="https://github.com/dergigi/vibeline-ui">vibeline-ui</a></li> <li>LLMs as tools, and how to use them</li> <li>Vervaeke: <a href="https://www.youtube.com/watch?v=A-_RdKiDbz4">AI thresholds & the path we must take</a></li> <li>Hallucinations and grounding in reality</li> <li>GPL, LLMs, and <a href="https://danb.me/blog/llms-licensing/">open-source licensing</a></li> <li>Pablo's multi-agent Roo setup</li> <li>Are we going to make programmers obsolete?</li> <li>"When it works it's amazing"</li> <li>Hiring & training agents</li> <li>Agents creating RAG databases of NIPs</li> <li>Different models and their context windows</li> <li>Generalists vs specialists</li> <li>"Write drunk, edit sober"</li> <li><a href="https://dvmcp.fun/">DVMCP.fun</a></li> <li>Recklessness and destruction of vibe-coding\ </li> <li>Sharing secrets with agents & LLMs\ </li> <li>The "no API key" advantage of nostr</li> <li>What data to trust? And how does nostr help?\ </li> <li>Identity, web of trust, and signing data\ </li> <li>How to fight <a href="https://nostr.band/?q=%23aislop">AI slop</a></li> <li>Marketplaces of code snippets\ </li> <li>Restricting agents with expert knowledge\ </li> <li>Trusted sources without a central repository\ </li> <li><a href="https://zapstore.dev/">Zapstore</a> as the prime example\ </li> <li>"How do you fight off re-inventing GitHub?"\ </li> <li>Using large context windows to help with refactoring\ </li> <li>Code snippets for <a href="https://olas.app/">Olas</a>, <a href="https://github.com/nostr-dev-kit/ndk">NDK</a>, NIP-60, and more</li> <li>Using MCP as the base\ </li> <li>Using nostr as the underlying substrate\ </li> <li>Nostr as the glue & the discovery layer\ </li> <li>Why is this important?\ </li> <li>Why is this exciting?\ </li> <li>"With the shift towards this multi-agent collaboration and orchestration world, you need a neutral substrate that has money/identity/cryptography and web-of-trust\ baked in, to make everything work."</li> <li>How to single-shot nostr applications\ </li> <li>"Go and create this app"\ </li> <li>The agent has money, because of <a href="https://nips.nostr.com/60">NIP-60</a>/<a href="https://nips.nostr.com/61">61</a></li> <li><a href="https://ppq.ai/">PayPerQ</a>\ </li> <li>Anthropic and the genius of <a href="https://modelcontextprotocol.io/docs/concepts/tools">mcp-tools</a>\ </li> <li>Agents zapping & giving SkyNet more money\ </li> <li>Are we going to run the mints?\ </li> <li>Are agents going to run the mints?\ </li> <li>How can we best explain this to our bubble?\ </li> <li>Let alone to people outside of our bubble?\ </li> <li>Building pipelines of multiple agents\ </li> <li>LLM chains & piped Unix tools\ </li> <li>OpenAI vs Anthropic\ </li> <li>Genius models without tools vs midwit models with tools\ </li> <li>Re-thinking software development\ </li> <li>LLMs allow you to tackle bigger problems\ </li> <li>Increased speed is a paradigm shift\ </li> <li>Generalists vs specialists, left brain vs right brain\ </li> <li>Nostr as the home for specialists\ </li> <li>fiatjaf publishing snippets (reluctantly)\ </li> <li>fiatjaf's <a href="https://github.com/fiatjaf/blossom">blossom</a> implementation\ </li> <li>Thinking with LLMs</li> <li>The tension of specialization VS generalization\ </li> <li>How the publishing world changed</li> <li>Stupid faces on YouTube thumbnails\ </li> <li>Gaming the algorithm\ </li> <li>Will AI slop destroy the attention economy?\ </li> <li>Recency bias & hiding publication dates\ </li> <li>Undoing platform conditioning as a success metric\ </li> <li>Craving realness in a fake attention world\ </li> <li>The theater of the attention economy\ </li> <li>What TikTok got "right"\ </li> <li>Porn, FoodPorn, EarthPorn, etc.\ </li> <li>Porn vs Beauty\ </li> <li>Smoothness and awe\ </li> <li>"Beauty is an angel that could kill you in an instant (but decides not to)."</li> <li>The success of Joe Rogan & long-form conversations\ </li> <li>Smoothness fatigue & how our feeds numb us\ </li> <li>Nostr & touching grass\ </li> <li>How movement changes conversations\ </li> <li>LangChain & DVMs\ </li> <li>Central models vs marketplaces\ </li> <li>Going from assembly to high-level to conceptual\ </li> <li>Natural language VS programming languages\ </li> <li>Pablo's code snippets\ </li> <li>Writing documentation for LLMs</li> <li>Shared concepts, shared language, and forks\ </li> <li>Vibe-forking open-source software\ </li> <li>Spotting vibe-coded interfaces\ </li> <li>Visualizing nostr data in a 3D world\ </li> <li>Tweets, blog posts, and podcasts\ </li> <li>Vibe-producing blog posts from conversations\ </li> <li>Tweets are excellent for discovery\ </li> <li>Adding context to tweets (long-form posts, podcasts, etc)\ </li> <li>Removing the character limit was a mistake\ </li> <li>"Everyone's attention span is rekt"</li> <li>"There is no meaning without friction"\ </li> <li>"Nothing worth having ever comes easy"\ </li> <li>Being okay with doing the hard thing</li> <li>Growth hacks & engagement bait</li> <li>TikTok, theater, and showing faces and emotions</li> <li>The <a href="https://en.wikipedia.org/wiki/1%25_rule">1% rule</a>: 99% of internet users are Lurkers</li> <li>"We are socially malnourished"</li> <li><a href="https://vertexlab.io/">Web-of-trust</a> and zaps bring realness</li> <li>The semantic web does NOT fix this LLMs might</li> <li>"You can not model the world perfectly"</li> <li>Hallucination as a requirement for creativity</li> </ul>

#07: Zig Multiplatform w/ Justin

\“The age of the idea guys has begun.\” <a href="https://njump.me/@justinmoon">Justin</a> &<a href="https://njump.me/@Gigi"> Gigi</a> vibe about Zig. Articles mentioned: <ul> <li><a href="https://wattenberger.com/thoughts/llms-as-a-tool-for-thought">LLMs as a tool for thought</a> by Amelia Wattenberger</li> <li><a href="https://nakamotoinstitute.org/library/micropayments-and-mental-transaction-costs/">Micropayments and Mental Transaction Costs</a> by Nick Szabo</li> <li><a href="https://wattenberger.com/thoughts/our-interfaces-have-lost-their-senses">How our interfaces have lost their senses</a> by Amelia Wattenberger</li> </ul> Talks mentioned: <ul> <li><a href="https://youtu.be/PBAuHv5cPl8">The Art of Bitcoin Rhetoric</a> by Bitstein</li> </ul> Books mentioned: <ul> <li><a href="https://bitcoin-resources.com/books/human-action">Human Action</a> by Ludwig von Mises</li> <li><a href="https://bitcoin-resources.com/books/working-in-public">Working in Public</a> by Nadia Eghbal</li> </ul> In this dialogue: <ul> <li><a href="https://github.com/fiatjaf/nak">nak</a>\ </li> <li>Files</li> <li><a href="https://syncthing.net/">SyncThing</a> (and how it BitTorrent Sync became Resilio Sync)\ </li> <li>Convention over configuration\ </li> <li>Changes & speciation\ </li> <li>File systems as sources of truth\ </li> <li>Vibe-coding shower thoughts\ </li> <li>Inspiration and The Muse\ </li> <li>Justin's LLM setup\ </li> <li>Tony's setup (o1-pro as the architect)\ </li> <li>Being okay with paying for LLMs\ </li> <li>Anthropomorphising LLMs</li> <li>Dialog, <a href="https://rubberduckdebugging.com/">rubber-duck debugging</a>, and the process of thinking\ </li> <li>Being nice and mean to LLMs</li> <li><a href="https://njump.me/nevent1qqsgmjffd7n382dj4e2l9dtctnrx98lw0vnt86rfggrgfq6vfpnysmszyqgmn2y5qndlxq6w0cvgdw5acnrdxahj8xs3sfcm6tk9v75gnpgvu6htgn9">Battlebots</a> & Gladiators</li> <li>Hedging your bets by being nice to Skynet</li> <li>Pascal's Wager for AI</li> <li>Thinking models vs non-thinking faster models</li> <li>Sandwich-style LLM prompting, again (waterfall stuff, <a href="https://gist.github.com/dskvr/b34f836b68ff2ca51bcc8fe7763eceec">HLDD / LLDD</a>)\ </li> <li>Cursor rules & Paul's <a href="https://github.com/futurepaul/prompt-buddy">Prompt Buddy</a>\ </li> <li>Giving lots of context vs giving specific context\ </li> <li>The benefit of LLMs figuring out obscure bugs in minutes (instead of days)\ </li> <li>The phase change of fast iteration and vibe coding\ </li> <li>Idea level vs coding level\ </li> <li>High-level vs low-level languages\ </li> <li>Gigi's "vibeline"\ </li> <li><a href="https://duckduckgo.com/?q=peterson+logos&iar=videos&iax=videos&ia=videos">Peterson</a>'s Logos vs <a href="https://duckduckgo.com/?q=vervaeke+dialogos&iax=videos&ia=videos">Vervaeke</a>'s Dia-Logos\ </li> <li>Entering into a conversation with technology\ </li> <li>Introducing MCPs into your workflow</li> <li>How does Claude think?\ </li> <li>How does it create a rhyme?</li> <li>How does thinking work?\ </li> <li>And how does it relate to dialogue?</li> <li>Gzuuus' <a href="https://github.com/gzuuus/dvmcp">DVMCP</a> & using nostr as an AI substrate</li> <li>Language Server Protocols (LSPs)</li> <li>VAAS: Vibe-coding as a service</li> <li>Open models vs proprietary models</li> <li>What Cursor got right</li> <li>What ChatGPT got right\ </li> <li>What Google got right</li> <li>Tight integration of tools & remaining in a flow state</li> <li>LLMs as conversational partners</li> <li>The cost of context switching\ </li> <li>Conversational flow & how to stay in it\ </li> <li>Prompts VS diary entries\ </li> <li>Solving technical vs philosophical models\ </li> <li>Buying GPUs & training your own models\ </li> <li>Training LLMs to understand Zig\ </li> <li>Preventing entryism by writing <a href="https://ziglang.org/documentation/master/">no documentation</a>\ </li> <li>Thin layers & alignment layers\ </li> <li><a href="https://bitcoin-resources.com/books/working-in-public">Working in public</a> & thinking in public\ </li> <li>Building a therapist / diary / notes / idea / task system</li> <li>"The age of the idea guys has begun."</li> <li>Daemons and spirits\ </li> <li>Monological VS dialogical thinking\ </li> <li>Yes-men and disagreeable LLMs\ </li> <li>Energy cost vs human cost\ </li> <li>Paying by the meter vs paying a subscription\ </li> <li>The equivalence of storage and compute\ </li> <li>Thinking needs memory, and memory is about the future\ </li> <li>Nostr+ecash as the perfect AI+human substrate\ </li> <li>Real cost, real consequence, and Human Action\ </li> <li>The cost of words & speaking\ </li> <li>Costly signals and free markets\ </li> <li>From shitcoin tokens to LLM tokens to ecash tokens\ </li> <li>Being too close to the metal & not seeing the forest for the trees\ </li> <li>Power users vs engineers\ </li> <li>Participatory knowing and actually using the tools\ </li> <li>Nostr as the germination ground for ecash\ </li> <li>What is Sovereign Engineering?\ </li> <li><a href="https://llvm.org/">LLVM</a> and the other side of the bell-curve\ </li> <li>How nostr gives you users, discovery, mircopayments, a backend, and many other things for free\ </li> <li>Echo chambers & virality\ </li> <li>Authenticity & Realness\ </li> <li>Growing on the edges, catering to the fringe\ </li> <li>You don't own your iPhone\ </li> <li><a href="https://grapheneos.org/">GrapheneOS</a>\ </li> <li>WebRTC and other monolithic "open" standards\ </li> <li>Optimizing for the wrong thing\ </li> <li>Building a nostr phone & Gigi's <a href="https://njump.me/nevent1qqsr4scsxwpvleeklkq60l6300s8hnkt579az2guhctv26slv0k3axcpz3mhxue69uhkummnw3ezummcw3ezuer9wcpzqmjxss3dld622uu8q25gywum9qtg4w4cv4064jmg20xsac2aam5nqvzqqqqqqyqz2yg6">dream flow</a>\ </li> <li>Using nostr to sync dotfile setups and other things\ </li> <li>"There are no solutions, only trade-offs"\ </li> <li>Cross-platform development\ </li> <li>Native vs non-native implementations</li> <li>Vitor's point on <a href="https://njump.me/nevent1qqsqzpr2ux0h8g2ccdjaavgll2vczpv5lvgwzxwjvuc2vn87c2dxe0gelp5ky">what we mean by native</a>\ </li> <li>Does your custom UI framework work for blind people?\ </li> <li><a href="https://ladybird.org/">Ladybird</a> browser & how to build a browser from scratch\ </li> <li><a href="https://templeos.org/">TempleOS</a>\ </li> <li>Form follows function & 90's interfaces\ </li> <li>Lamentations on the state of modern browsers\ </li> <li>Complexity & the downfall of the Legacy Web\ </li> <li>Nostr as the "new internet"\ </li> <li>Talks by Ladybird developer <a href="https://duckduckgo.com/?q=andreas+kling+ladybird&iax=videos&ia=videos">Andreas Kling</a></li> <li>Will's attempt of building it from scratch with <a href="https://damus.io/notedeck/">Notedeck</a> & <a href="https://github.com/damus-io/nostrdb">nostr-db</a></li> <li>Justin's attempt with <a href="https://github.com/justinmoon/rust-multiplatform">rust-multiplatform</a>\ </li> <li>"If it doesn't have a rust implementation, you shouldn't use it."</li> <li>Native in terms of speed vs native in terms of UI/UX\ </li> <li>Engineer the logic, vibe-code the UI\ </li> <li>From Excalidraw to app in minutes\ </li> <li>What can you one-shot?\ </li> <li>What do you need to care about?\ </li> <li>Pablo's NDK snippets\ </li> <li><a href="https://eugenkiss.github.io/7guis/">7GUIs</a> and GUI benchmarks for LLMs</li> <li>"Now we're purpose-building tools to make it easier for LLMs"\ </li> <li>"Certain tools really make your problems go away."</li> <li>Macros and meta-programming\ </li> <li>Zig's <a href="https://zig.guide/language-basics/comptime/">comptime</a>\ </li> <li>UNIX tools and pipes\ </li> <li>Simple tools & composability</li> <li>Nostr tools for iOS & sharing developer signing keys\ </li> <li>Building 10 apps as one guy\ </li> <li>Simplicity in a community context\ </li> <li>Most people are on phones\ </li> <li>Most people don't install PWAs\ </li> <li><a href="https://zapstore.dev/">Zapstore</a> & building our own distribution channels</li> <li>Web-of-trust and pushing builds quickly\ </li> <li>Improving homebrew by 10x\ </li> <li>(Micro)payments for package managers\ </li> <li>Guix and bitcoin-core\ </li> <li>Nix vs Guix\ </li> <li>Reproducible builds & web-of-trust\ </li> <li><a href="https://keet.io/">Keet</a> vs "calling an npub"\ </li> <li>Getting into someone's notifications\ </li> <li>Removing the <a href="https://njump.me/nevent1qqsxvahg4a2l6jtv5da9jycx9q0g6rjrtsgpzgj47grv7ct47yr59xgzyphydppzm7m554ecwq4gsgaek2qk32atse2l4t9ks57dpms4mmhfxqcyqqqqqqgx9jtpx">character limit</a> was a mistake</li> </ul> Recorded at <a href="https://mempool.space/block/889812">889,812</a>.

Recorded at 890,848.

#nosolutions

Write a comment

No comments yet.

08: Navigating the Vibe: AI, Collaboration, and the Future of Creativity

23: Shipping Violently w/ Justin Moon

24: Building FIPS w/ Johnathan Corgan

22: Sovereign Engineering w/ Yo