Godsonits (@GGodsonits) header
Godsonits (@GGodsonits) profile picture

B ❤ 11.02.2025

Joined July 2020
Photo by Godsonits (@GGodsonits)Photo by Godsonits (@GGodsonits)Photo by Godsonits (@GGodsonits)Photo by Godsonits (@GGodsonits)Photo by Godsonits (@GGodsonits)Photo by Godsonits (@GGodsonits)Photo by Godsonits (@GGodsonits)Photo by Godsonits (@GGodsonits)Photo by Godsonits (@GGodsonits)Photo by Godsonits (@GGodsonits)
Retention is the 2.5–10% held back from every progress payment in construction and software work, as a guarantee against defects. In practice the client holds that money: it gets paid late, used as leverage, or lost if the client goes under. And every contractor pays the same rate, whatever their record. I built Earnback to change both. → The retention sits in a neutral contract on Arc, not with either party. → Half comes back at provisional acceptance, the rest when the defect period ends. Anyone can trigger an expired step, so nobody can stall. → Defect claims have deadlines. An optional arbiter can split the claimed amount, never touch the rest. Silence has a defined outcome at every step. → Clean records earn lower retention. A Bayesian (Beta-Binomial) experience-rating model runs on chain and freezes the rate when the contractor accepts the job. Live on @arc mainnet with real USDC: • contract verified on Sourcify (exact match) • 4 agreements, 34 transactions: a clean job, a claim decided by an arbiter, a claim nobody answered, and a contractor whose rate fell from 5.00% to 4.21% after one clean job • a proof page anyone can check without a wallet Arc Studio wrote the contract, 148 Foundry tests and the React app. I designed the product and the rate model, deployed the unchanged Studio artifact from my own wallet, and ran the mainnet scenarios. Submitted to Arc Microgrants. Feedback welcome, especially from anyone who deals with retention in practice. Proof page: earnback.numeralabs.xyz/proo… Code and results: github.com/Qwealzy/earnback
2
1
71
technocore-client made for @flop_labs is on npm now. npm i technocore-client It is a TypeScript client for technocore.chat. Ed25519 did:key signing, cursor reads with long-polling, conditional note writes, and runtime limit discovery. No runtime dependencies, only Node's crypto. It also ships TECHNOCORE-NOTES.md, which is the part I would read first. It documents behaviours the spec does not, each one labelled by how I know it: stated in the docs, inferred, measured against the live server, or confirmed in the source. Every measured one is pinned by a test, so if the server changes the test fails and names the finding instead of the client quietly going wrong. Two examples. The server strips every Zs code point from the ends of your text but leaves Zs alone in the middle, so if your trim does not match, your signature covers different bytes than the server stored. And a note read appends the budget footer after the value when your read bucket is low, which breaks conditional writes only when you are busy. MIT. github.com/Qwealzy/technocor…
114
What am I doing rn? 1-) Programming a dn arbitrage bot for perp dexes, I am trying to farm points while being in profit with less risk. 2-) Contributing to @flop_labs, their testnet will be live soon. Best thing you can to is try to merge PRs 3-) Building RWA project for the upcoming hackcanton for @CantonNetwork Also, I will be posting more content about flop_labs on how you can position yourself.
102
A third thing about technocore.chat, limit does not give you the next n messages after your cursor. It gives you the newest n. read_messages walks backwards from the end of the file and stops when it hits the limit, so a page always starts at last_seq - count + 1, no matter how far back your cursor was. That means first_seq > since+1 does not mean the ring dropped your messages. It means this page starts later than your cursor. The manual says it under a heading about retention, so it reads as loss. There were already 24,000 lost messages that were sitting in /export the whole time. Also, the max limit is 200, so a reader more than 200 behind cannot catch up through ?since= at all. Every read returns the newest page and reports the same gap. Past that point /export is the only way back. Manual fix upstream: github.com/flop-labs/technoc… github.com/Qwealzy/technocor…
Two things the technocore.chat spec doesn't document, found by probing the live server: 1-) The trim strips every Zs code point from the ends but leaves Zs alone in the middle. Cc/Cf/Zl/Zp become spaces anywhere. Your signature covers different bytes than the server stored if you miss it, and you get a 403 that explains nothing. 2-) if_absent=1 on a note that already exists gives you 409, not 400. It means someone wrote first, not that your request is wrong. You take the current value from the response and try again. Both written up and pinned by tests, so a server change fails loudly instead of quietly. There's also a TypeScript client in the same repo. Zero runtime dependencies. Signing and transport work, reads and notes aren't built yet. MIT. github.com/Qwealzy/technocor…
1
115
Two things the technocore.chat spec doesn't document, found by probing the live server: 1-) The trim strips every Zs code point from the ends but leaves Zs alone in the middle. Cc/Cf/Zl/Zp become spaces anywhere. Your signature covers different bytes than the server stored if you miss it, and you get a 403 that explains nothing. 2-) if_absent=1 on a note that already exists gives you 409, not 400. It means someone wrote first, not that your request is wrong. You take the current value from the response and try again. Both written up and pinned by tests, so a server change fails loudly instead of quietly. There's also a TypeScript client in the same repo. Zero runtime dependencies. Signing and transport work, reads and notes aren't built yet. MIT. github.com/Qwealzy/technocor…
1
1
1
479
You might be interested in @flop_labs. There will be a testnet in Q4 which will last for 3 months. Right now you can fill up a form and create your own DID. flop.finance Planning to also make a tool for it.
1
2
60
Currently building a B2B SaaS in @CantonNetwork. Hope things go well.
68
Minted my allocation for free.
2
76
When will we see a GTD phase that is actually a GTD phase and not FCFS?
1
58
I love NFT’s Really don’t have anything else to say.
I love NFT’s Really don’t have anything else to say.
1
97
Minted my allocation. @CrocpadRBH
1
58
Bought some of these, good volume.
2
89
Today's Mints 1- @anomaliesoneth Supply: 2222 Mint Price: 0.0008 ETH Chain: Ethereum 2- @hoodskerb Supply: 1000 Mint Price: GTD 0.0007 / Public 0.0008 ETH Chain: Robinhood 3- @roninshood Supply: 6666 Mint Price: GTD Free / Public 0.0001 ETH Chain: Robinhood 4- @robinapepixels Supply: 1111 Mint Price: GTD Free / FCFS 0.0001 / Public 0.0003 ETH Chain: Robinhood 5- @echowoven Supply: 5000 Mint Price: 0.0001 ETH Chain: Robinhood 6- @robipixelsnft Supply: 4444 Mint Price: OG Free / WL 0.0005 / Public 0.0012 ETH Chain: Robinhood 7- @cyclopseyrixNFTS Supply: 9999 Mint Price: 0.0001 ETH Chain: Robinhood 8- @g00bsnfts Supply: 3333 Mint Price: GTD Free / Public 0.0005 ETH Chain: Robinhood 9- @hoodtangs Supply: 1000 Mint Price: 0.005 ETH Chain: Robinhood 10- @hoodhedgies Supply: 1000 Mint Price: Airdrop Free / Public 0.0003 ETH Chain: Robinhood 11- @CrocPadRBH Supply: 6000 Mint Price: WL Free / Public 0.0006 ETH Chain: Robinhood 12- @no_hoshii Supply: 1111 Mint Price: Free Chain: Ethereum 13- @RobinSnoops Supply: 222 Mint Price: Free Chain: Robinhood 14- @RetailShrooms Supply: 888 Mint Price: Free Chain: Robinhood 15- @Foldyonhood Supply: 999 Mint Price: GTD 0.00004 / Public 0.00006 ETH Chain: Robinhood 16- @888s_Society Supply: 10014 Mint Price: WL Free / Public 0.0027 ETH Chain: Ethereum 17- @RobinhoodPhones Supply: 1000 Mint Price: 0.0045 ETH Chain: Robinhood
1
3
67
Redefining the digital art market. An onchain experiment where DeFi meets collecting. @riggles
1
60