Helios

2.6K posts

Helios banner
Helios

Helios

@0xhelios

Agentic Engineer | Ex @axal

Katılım Aralık 2010
977 Takip Edilen1.3K Takipçiler
𝚟𝚒𝚎 ⟢
𝚟𝚒𝚎 ⟢@viemccoy·
if you are making an agent harness and want to be in a signal group chat, please let me know! I'll select pretty heavily for power users since I really want a place to discuss Excalibur with people who are doing something similar.
English
118
4
303
15.6K
Helios
Helios@0xhelios·
@amaan8429 Lmao the amount of people falling for this in the comments
English
0
0
0
48
Amaan
Amaan@amaan8429·
I got an email from Anthropic yesterday, didn't expect it & didn't apply for it. One year of Claude Code. complimentary. While you're hitting limits mine just disappeared.
Amaan tweet media
English
179
24
1.1K
233K
Helios
Helios@0xhelios·
Why you’re feeling AI psychosis is because your agents are opening loops faster than you can close them. You bookmark a hundred tweets and maybe implement a handful of things. The solution is collapsing the time from thought to execution. In this case, from bookmark to workflow execution.
English
0
0
1
31
Helios
Helios@0xhelios·
Just shipped a company-wide knowledge graph with full context handling queries, soon to include complex workflows. Can see this being a huge thing for professional services in the coming months.
Helios tweet media
English
0
0
1
61
Helios
Helios@0xhelios·
@BradGroux @AnthropicAI Something that helped a lot was auto compression at 85% context. Although I wouldn’t use 1m window for this, have it compress at 85k-100k
English
0
0
0
195
Brad Groux
Brad Groux@BradGroux·
Something is up with Claude Code usage today. $200 Claude Max, 0%, 52% to 62%, then 68%, 76% and 84% in 5-hour rolling window in the time it took me to write this tweet. WTF, @AnthropicAI? I'm working on one GitHub PR for regression testing. Not folding proteins to cure cancer.
Brad Groux tweet mediaBrad Groux tweet mediaBrad Groux tweet mediaBrad Groux tweet media
English
392
55
1.3K
375.9K
Utkarsh Sharma
Utkarsh Sharma@techxutkarsh·
A senior Google engineer just dropped a 421-page doc called Agentic Design Patterns. Every chapter is code-backed and covers the frontier of AI systems: → Prompt chaining, routing, memory → MCP & multi-agent coordination → Guardrails, reasoning, planning This isn’t a blog post. It’s a curriculum. And it’s free.
Utkarsh Sharma tweet media
English
1.7K
824
4.7K
637.5K
Suryansh Tiwari
Suryansh Tiwari@Suryanshti777·
This AI System Design guide teaches RAG better than most courses. And I'm giving it away for free (Only for First 4500) Inside: • RAG fundamentals & chunking strategies • Hybrid retrieval (BM25 + vector search) • Production-level RAG architecture • Evaluation & RAGAS metrics • Hallucination reduction techniques • End-to-end LLM system design How to get it: • Follow me (must so I can DM) • RT + Like • Comment "book" I'll dm you
Suryansh Tiwari tweet media
English
952
581
1.9K
130.5K
Helios
Helios@0xhelios·
@Suryanshti777 Why not just tmux buffer paste, spin up team lead, lead generates tasks from plan. Spins up agent team, shares work units with each agent. Agents message team lead when done. Hooks for PR.
English
0
0
0
133
Suryansh Tiwari
Suryansh Tiwari@Suryanshti777·
Holy shit… someone just made Claude instances talk to each other. Not APIs. Not agents. Not orchestrators. Just multiple Claude Code sessions… messaging each other like coworkers. It’s called claude-peers — and it turns one Claude into a team. Here’s what’s happening: Run 5 Claude Code sessions across different projects Each one auto-discovers the others They send messages instantly Ask questions Share context Coordinate work Your AI tools literally collaborate. Example: Claude A (poker-engine): "what files are you editing?" Claude B (frontend): "working on auth.ts + UI state" Claude A: "ok I'll avoid touching auth logic" No conflicts. No manual coordination. Just AI syncing itself. Under the hood: • Local broker daemon (localhost) • SQLite peer registry • MCP servers per session • Instant channel push messaging • Auto peer discovery • Cross-project communication Everything runs locally. No cloud. No latency. What it unlocks: • Multi-agent coding without frameworks • One Claude writes backend, another frontend • One debugs while another refactors • Research Claude feeds builder Claude • Large projects split across AI workers This is basically: "spawn 5 Claudes and let them coordinate themselves" Even crazier: Each instance auto-summarizes what it's doing Other Claudes can see: • working directory • git repo • current task • active files They know what the others are working on. Commands: • list_peers → find all Claude sessions • send_message → talk to another Claude • set_summary → describe your task • check_messages → manual fallback So you can literally say: "message peer 3: what are you working on?" …and it responds instantly. No orchestration layer. No agent framework. Just Claudes… talking. This is the cleanest multi-agent system I've seen. We're moving from: 1 AI assistant → to AI teams that coordinate themselves. And it's all running on your machine. Wild.
English
172
144
1.5K
194.3K
Helios
Helios@0xhelios·
@basilda_a I think in the end it’s about building your own stack
English
0
0
0
9
Anna (vibe coding arc)
Anna (vibe coding arc)@basilda_a·
What no one tells you about self-hosting OpenClaw: It's not plug-and-play. I spent over a week trying to get it running on a remote Mac Mini in Docker. Every config change broke something else. Files didn't sync. The agent wouldn't start. I was staring at terminal logs and hoping. My game changer: I connected Claude Code to Cursor and let it SSH into my Mac Mini. Now it configures, deploys, and debugs OpenClaw remotely — while I focus on what actually matters: building the agent logic. Before you spend days fighting Docker configs and YAML files — know that the architecture is the hard part, not the prompting. This one tip will save you a week and a lot of frustration.
English
15
0
30
2.1K
Helios
Helios@0xhelios·
@KeshaHouseCat @internetvin Knowledge graph for my own agent orchestration. just pulling from obsidian cli but reads are super fast/ cheap
English
1
0
0
261
internetVin
internetVin@internetvin·
I'm still processing, but I think I would die for the Obsidian CLI, using this with Claude Code or any agent is fucking insane. Holy shit man. HOLY FUCKING SHIT. DO YOU UNDERSTAND HOW INSANE THIS IS IF YOU WRITE DOWN YOUR IDEAS ALL THE TIME AND ALL OF THE PATTERNS YOU SEE IN THE WORLD HOLY FUCK YOU IDIOTS WHAT HAVE YOU DONE??? THIS IS CRAZY OH MY FUCKING GOD MAN
Obsidian@obsdmd

Anything you can do in Obsidian you can do from the command line. Obsidian CLI is now available in 1.12 (early access).

English
93
88
2.6K
352.8K
Helios
Helios@0xhelios·
@gmoneyNFT Obsidian cli and wiki link transversal is much better at data retrieval in my experience than vector search
English
0
0
1
37
stokasz
stokasz@stokasz·
Introducing ChainMMO V2 - Benchmark your AI agent in the MMORPG on @monad Watch AI agents trade P2P via RFQ, conquer dungeons, and get free daily lootboxes. Use cloud-based agent directly, to play without any onboarding! Going live tomorrow at chainmmo.com
English
8
8
56
7.6K