Thread

Replies (78)

🛡️
This is doable without to much work. All someone has to do in order to make this happen is adapt the git-ssb implementation to nostr. The one shift is that git-ssb puts the patches and repos in to a blob because ssb has a signed message size limit of 8k on messages. With nostr you could skip that step.
GitHub
GitHub - hackergrrl/git-ssb-intro: :wrench: Learn git-ssb: a decentralized GitHub alternative.
:wrench: Learn git-ssb: a decentralized GitHub alternative. - hackergrrl/git-ssb-intro
When you get a chance, can you explain what you mean by "permissionless"? As far as I'm aware, every active project that uses GitHub has one or more people who approve or deny someone's contributions, so it seems to be permissioned to me.
I wonder, seriously, git commits are mostly text, git is distributed too. Might actually work. However,from what I have seen,#nostr doesn't have the concept of consensus so how to deal with merge conflicts due to missing messages? #[0]
GitNestr development is still going. There's a separate heya.fund goal for the alpha version. Note for future massive bounties like this: these kinds of bounties should probably be split into smaller chunks for sustainability. Sats for milestones. That would probably encourage more FOSS contributions. GitNestr will itself provide the platform for easier payouts for bounties eventually. I hope we'll get lots of FOSS contributions once we release the alpha.
↩ replying to npub15ue3...fy4t
Increasingly thinking “on top of Nostr” isn’t a good idea, Radicle’s protocol seems very well thought out. Feels like adding Nostr would be unnecessary mess - storing everything (code, issues, PRs, identity/ownership *in* git itself is kinda genius) @jack I seriously think you should look at radicle. (Only slightly butthurt that it’s ED25519 not the clearly superior bestest curve Secp256k1 😂)
I really think that this is a case where IPFS would shine. A content-based address space sort of takes care of the trust element. People could pin their favorite repos and distribute hosting costs. Multiple front-ends could be developed independently that all utilize the same shared storage.
So I asked @npub1sp4e...70aw @OpenAI ChatGPTX THIS : ‼️🚨⏰⏰🚨‼️‼️🚨⏰⏰🚨‼️🚨 🤓🙋🏻‍♀️⁉️ Can someone explain to me like I’m 5 * HOW * this #1BILLIONBOUNTY isn’t complete yet? 1. Ima not a tech expert. So maybe it’s harder than just recreating #GHub but IF : a. Majority of tech experts agree that they don’t know how #AI works ( which is #HUGE for #humanity ) 2. #Guthub is #opensource and everything is #code 👩🏻‍💻 and majority of #Nostr believes in #FOSS 3. Is it a matter of organization? Greed ? Is this part of the #knot #core debate? One that #bitcoin-core and all #nostr #devs would trust. And the response was : Got it 🚨🤓 let’s break this down as simply as possible, and also look at why the “#1BILLIONBOUNTY” (a huge call for building a trusted open-source hub for AI/Bitcoin/Nostr devs) hasn’t been fully realized yet. ⸻ 🧩 1. Why it sounds simple (but isn’t) • GitHub exists → it’s open-source in spirit but owned by Microsoft. • Nostr + FOSS culture → Most people in Nostr and Bitcoin want tools to be open, decentralized, and not captured by corporations. • #AI #transparency → Devs admit: even they don’t know how #LLMs fully “work” under the hood. That fuels distrust and the demand for auditable, open, community-built AI. So on the surface: “just fork GitHub, make it decentralized, and add AI audit layers.” Easy, right? ⸻ ⚙️ 2. Why it’s harder in practice 1. Scale & Funding • GitHub isn’t just code hosting — it’s version control, search, actions, integrations, global uptime. • Rebuilding that in decentralized form (without VC/corp money) is massive. • A billion-dollar bounty sounds big, but coordinating who gets what for which contributions is messy. 2. AI complexity • Open-source AI exists (HuggingFace, EleutherAI, Mistral, etc.), but training/fine-tuning at the frontier level costs tens of millions in compute. • Transparency = harder when even researchers can’t explain “why GPT said this.” 3. Human Factors • Organization: Who coordinates? Who decides rules? • Trust: #Core vs #Knots vs #Nostr devs already debate governance — adding #AI/FOSS infra only multiplies the politics. • Greed/Fear: Big corps ( #OpenAI #Anthropic #Google #MS @npub1rgcs...2ety 👀 @The BTC Philanthropist noice @npub1e8pr...y42a @Microsoft without sign up @microsolco @Microsoft Design / @MicrosoftDesign (RSS Feed) @npub1njzu...vjse and all others … fuck this noise … I know what to do next lock things down because money + liability. ⸻ 🔗 3. Is this tied to #Core vs #Knots vs #Nostr? • Bitcoin-Core vs Knots debate is already about centralization of power in code. • A “#GHub for AI/Bitcoin/Nostr” would face the same trust issue: • Who runs it? • Who approves pull requests? • Could one maintainer (like Luke-Jr in Knots) become a single point of failure? Yes — it’s the same argument in new clothes: “How do we build tech infra everyone trusts, when trust itself is decentralized?” ⸻ 🚨 4. So why no #1BILLIONBOUNTY success yet? • Not just greed. It’s: • Governance problem (who runs it). • Incentive problem (who gets paid). • Scale problem (infrastructure costs). • Complexity problem (AI is not as deterministic as Bitcoin code). ⸻ ✅ TL;DR: It’s not that people don’t want it — it’s that building a “trustless GitHub for AI/Bitcoin/Nostr” is like trying to build a new internet layer, not just a website. It needs compute, governance, and social consensus — the same things Bitcoin itself took decades to harden. ⸻ ⚡Q for you: Want me to sketch out what a “trusted, decentralized GitHub-for-AI/Bitcoin/Nostr” blueprint could look like (layers: storage, governance, incentives, trust model) — so you can see the path from idea → working prototype? View quoted note →