@marty_kausas Feel free to try: github.com/WillNigri/Agen… . You can run any model, the app is local with all receipts from war rooms and conversations you have with all llms and the teams you set up.
English
nigri
3.4K posts
nigri
@willnigri
Building Cool Stuff | Ambassador @walrusprotocol | Validator @tropikalnode | Co-founder Osiris
Katılım Temmuz 2017
1.6K Takip Edilen3.3K Takipçiler
i'm so sick of using claude code in a terminal
i'm not coding. who has made a great app that i can use with any model?
English
@ClaudeDevs Two things we test next, same receipts:
1.recognition at scale, a weak model picking answers it can’t generate (2/12 solo → 9/12)
2.does coordination beat the best single model on long-horizon agentic tasks
Run the same benchmark on your own keys: agentictool.ai
English
Where teaming up does win: the asymmetric case. @ClaudeDevs showed an advisor lifting Sonnet ~8.5 pts on SWE-bench Pro at ~63% of flagship cost, on long-horizon work. Same verdict as ours: coordinate asymmetrically, route otherwise.
ClaudeDevs@ClaudeDevs
On BrowseComp, we tested Claude Managed Agents w/ Fable 5 orchestrator + Sonnet 5 worker sub-agents. The Fable 5 orchestrator achieves 96% of Fable 5 performance at 46% of the price. Token-heavy research is delegated to Sonnet 5. See our cookbook: github.com/anthropics/cla…
English
LLM Council idea: several models deliberate, a chairman decides. Does that actually beat one strong model?
We benchmarked it plus 8 other coordination methods on 49 fresh coding problems, every call metered.
Coordination raises the floor and lowers the ceiling. No team won outright.
Every run is public, check us:
agentictool.ai/posts/does-mul…
English
@willnigri oh i like you for it, which one would you be interested in tho
English
🚨 HIRING!
this is a real post, not some fake engagement. serious inquiries only.
i'm looking for 2 community managers
1) someone who is passionate about prediction markets
2) someone who is passionate about video games.
- Must be willing to work 4-6 hrs per day
- Must understand culture within their field (Prediction Markets OR Gaming)
- Must have previous community manager experience.. REAL EXPERIENCE!! Not you managing 5 kids in a roblox server
- Must be hardworking and want to be part of something special
I'll make updates as i move forward, but feel free to reply or tag someone who may be interested / a good fit for this.
English
BREAKING: CLAYNOSAURZ TO DISTRIBUTE 15% OWNERSHIP TO THEIR HOLDERS
We've set aside a 15% allocation of options in the equity of the Claynosaurz brand. This will be for our eligible ecosystem holders to participate in our future growth.
Next week we’ll launch an allocation website to check your position, along with detailed information.
English
Still holding my Popkins by @Claynosaurz. Hope this is worth something :)
Cab 🌋@Cabanimation
We took a stealth snapshot of Claynosaurz digital collectibles on June 29th. Unannounced and closely guarded with purpose. Specifically we wanted to avoid short-term hype buyers and volume noise rushing in. This was to protect and reward the real community: longevity, actual holdings, and conviction. And true collecting. This has been a backbone for @Claynosaurz sustainability. Longer hold times and deeper community ties should carry more weight. This is how we build for the long term. We want compounding progression for the right participation. I take pride in the stability, health and organic nature of our ecosystem. A safe, fun place outside of market noise. Most of all, I want to make it clear we would not be here without your support. This is community-first, and this will make sense. I’m so excited to reveal what this was all about. I think it’s a major step forward in the history of the brand we’ve built together.
English
@Crypto_Cobba @zabimx @officialSUIG The article said the deal was: 4m SUI (lended by SUIG to Bluewater). So around 2.8m USD as of the time of writing.
English
@willnigri @zabimx @officialSUIG I wonder why Rooter decided to pack up shop.. maybe they got offered a deal too good to refuse or they lost confidence that Sui is still the protocol for the future. It would be good to know their thoughts and whether they are planning to build anything else on Sui or not.
English
🚨 Sui DeFi just consolidated hard.
Bluewater acquired Suilend with Bluefin's/Bluewater CEO @zabimx is stepping in to run it. Bluefin financed the deal, backed by 4M SUI from SUI Group (@officialSUIG).
So one team now controls two of Sui's biggest lending protocol AND its biggest trading venues. They're calling it "independent." On paper.
The part worth noticing: $SEND isn't coming along. It's being wound down with a tokenholder distribution (details "soon"). Rootlets stays separate.
Less players. More concentration. Worth watching who controls what on Sui.
R🐽ter@0xrooter
sharing a big update today: bluewater labs is acquiring suilend. we launched suilend two years ago with the goal of bringing high quality defi products to sui. it's been an intense journey, full of ups and downs, and I'm proud of what we accomplished. we built and operated one top lending platform on sui, pioneered instant liquid staking with springsui, and cultivated a strong community around all of it. I'm especially proud that suilend earned trust the hard way: by taking security seriously and protecting user funds over two years of live markets. as suilend grew from a new lending protocol into a core part of sui defi, we started thinking about what the next chapter should look like. bluewater felt like the right team to take things forward. they have a native understanding of the sui ecosystem, deep experience in defi, and a track record of operational excellence. I'm grateful to everyone who used suilend, supported us, integrated with us, competed with us, and helped us get here. I'm excited to see where bluewater takes things in this next chapter!
English
@patrickc Building toward exactly this. Already shipping: multi-runtime coding agents, stored prompts/workflows/specialists, shareable audit-trail artifacts as compiled outputs. Real-time collab on the Pro roadmap. Open source, MIT: github.com/WillNigri/Agen…
English
I want some kind of LLM workflow tool.
• Ability to manage a set of input files (Markdown or similar), plus other general-purpose context.
• With real-time collaboration. (And maybe some concept of snapshots or VCS integration.)
• And the ability to create/manage a inference workflows and a stored set of prompts.
• Access to general-purpose coding agents (and not just chat models).
• Some concept of compiled outputs/inference results (which ideally can be shared externally).
Many projects have this feeling: "there is all this stuff, which I want to process/compute over in this iterated way, with some build artifacts being important/worth saving." GNU Autotools x Notion or something. Is anyone building this?
English
“Loops are privilege” is true if you assume a loop = N calls to Opus. It doesn’t have to. Free Ollama draft → Sonnet refines → Opus only on the gnarly turns. That’s a real loop on low budget/mo. The actual privilege is access to the playbook, not the budget.
Peter Steinberger 🦞@steipete
Here’s your monthly reminder that you shouldn’t be prompting coding agents anymore. You should be designing loops that prompt your agents.
English
Genuine question: is the pain the coding setup itself, or losing control once agents auto-trigger and you can’t see what each one did?
I went the other way from Devin. Building an open-source control layer so you keep all your harnesses and just get one place to run and audit them: github.com/WillNigri/Agen…
English
Word on the street is that everyone is going to be switching back to Opus when the new model drops.
This is exactly why I use an independent agent lab like Devin for my main software factory.
They're going to deal with that headache for me.
There's no way you can move fast and reliably if you're constantly switching between Claude Code, Codex and insert-other-shiny-object-here.
English
@security @perf brew tap WillNigri/ato
brew install --cask ato
Or grab the DMG / AppImage / Windows installer from releases.
🔗 website: agentictool.ai
🔗 GitHub: github.com/WillNigri/Agen…
Bring your own keys. Local-first. MIT.
English
Best moment from this week: I used ato to decide ato's positioning.
Dropped Gemini + MiniMax into a session, made them argue for 5 rounds with me moderating. They converged on a hybrid I shipped that afternoon.
The headline now on the website was produced by the product, on camera. Here is part of the video. Full 5 rounds was 10 mins long.
English
drop multiple LLMs into one shared session. They argue with you, call real tools (read_file / grep / git_log) to verify claims in your repo, and cite every file they checked.
Same primitive as @karpathy's llm-council, different shape: multi-provider, tool-calling, audit trail.
English