nitrateX (@nitraeX) header
nitrateX (@nitraeX) profile picture

Emerging Builder. $MEOW @SalvorKnows @SalvorLabsAI https://nitter.cf/t.co/szs1a1vcJq https://nitter.cf/t.co/h49iBsf9BT

Antalya
Joined November 2010
Photo by nitrateX (@nitraeX)Photo by nitrateX (@nitraeX)Photo by nitrateX (@nitraeX)Photo by nitrateX (@nitraeX)Photo by nitrateX (@nitraeX)Photo by nitrateX (@nitraeX)Photo by nitrateX (@nitraeX)Photo by nitrateX (@nitraeX)Photo by nitrateX (@nitraeX)Photo by nitrateX (@nitraeX)
Salvor × Whales Hideout AMA (8 Kasım 2026): Soru-cevap, Türkçe
7
15
105,650
Piece by piece, it's coming together.
One shared memory for your AI agents, with spending limits that stay in your hands.
The team keeps building. No breaks.
salvorlabs.org @SalvorLabsAI
1
3
1
7
194
Agents that pay for themselves. x402 payments (in development) start in Salvor+. Whether they also come to the free Salvor OS is still open, and contributions are welcome. @SalvorLabsAI $SALVOR #MCP #AIagents #OpenSource
3
6
55
Your keys, your coins. Your project, your knowledge. Salvor OS keeps what your AI learned inside your own repo, not on someone else's server. @SalvorLabsAI $SALVOR #OpenSource #AI #AIagents
3
4
72
nitrateX retweeted
Salvor OS is in open beta and built in public. The protocol, the claims and the tests are all in the repo, open to read, question and improve. Contributions welcome. salvorlabs.org @SalvorLabsAI #OpenSource #AI @BlockchainDan @SalvorKnows
2
8
72
Plain files. In your repo. Yours. No lock-in: your .salvor folder is readable markdown with or without any tool. @SalvorLabsAI $SALVOR #OpenSource #AI #DevTools
2
3
44
nitrateX retweeted
Can you review what your AI learned? .salvor files live in Git, so every change to what the project knows shows up in a diff, like any code change. @SalvorLabsAI $SALVOR #OpenSource #DevTools #AI @BlockchainDan
2
8
146
nitrateX retweeted
How do you pay for Salvor MCP? In $SALVOR, to start. A Stripe option for cards and other crypto comes later (planned). @SalvorLabsAI $SALVOR #MCP #AI #AIagents @BlockchainDan
2
10
115
Soon Salvor Plus from Labs will allow non-coding projects to store project brain wherever they like (preferably with built in version history) - git, gDrive, Dropbox, and many more. $SALVOR MCP delivers better solutions for ANY problem domain. It's not just for coding.
1
2
2
87
Love chattin w community, debating & reading AI research, etc. That's all great! Our product has to get built as well. As such- code hole for me today; to GSD. Reply to DMs, TGs and general chit chat msgs unrelated to *code deliverables* will be slightly delayed. Ty.
3
4
1
17
580
Key Takeaway (from OP): "It’s about building systems around models, giving them context, memory, tools, feedback loops, and data flows. “Delete everything, keep Graph.”" Sounds like someone literally describing @SalvorKnows. Decouple the model. Knowledge is the real edge.
Andrej Karpathy’s 1-hour Stanford lecture on AI engineering is one of the best explanations I’ve seen of how AI systems actually work. The progression is simple: 10% - LLM 30% - Prompt 50% - Agent 70% - Loop 100% - Graph The key takeaway: AI engineering isn’t just about writing better prompts. It’s about building systems around models, giving them context, memory, tools, feedback loops, and data flows. “Delete everything, keep Graph.” Definitely worth watching if you’re building with AI agents. Watch and bookmark it.
2
4
1
13
410
nitrateX retweeted
Cool (and smart) approach to what seem like a major inefficiency in token usage.
Researchers at the Chinese Academy of Sciences cut AI-agent token consumption by up to 92.9%, without retraining a single model. They also reduced execution time by up to 89%. The problem is familiar if you've built AI agents. Ask a coding agent to investigate one error, and it may read an entire log, repository file, or document when only a few passages are relevant. Those unnecessary tokens can remain in context across subsequent model calls, increasing cost and latency. The researchers developed SparseRead, a system that controls what an agent reads before the content enters the model's context. It works through three components → 1./ Read Intent Protocol: Tracks what information the agent needs, which evidence has been found, and what remains unresolved. 2./ Adaptive Reader Backends: Locate relevant evidence across documents, PDFs, code repositories, and structured data. 3./ Budgeted Read Gate: Limits how much source content enters context, with the option to fall back to reading the complete file. Across six models and five workload scenarios, the researchers reported → > Up to 92.9% fewer tokens consumed. > Up to 89% less execution time. > Median reductions of 61.4% in tokens and 53.7% in time. > Task quality preserved or improved in most evaluated settings. No model retraining is required. The interesting engineering lesson is that context optimization doesn't have to begin with compression or KV caching. You can reduce a substantial amount of work by controlling what enters the context in the first place. (Link to the paper and code in comments)
3
4
23
526
nitrateX retweeted
533 holders today. A growing story for tomorrow. 💙 Every journey starts somewhere. What matters is not just how many people are here today, but the vision, belief, and contributions that bring people together. Salvor is building something bigger than a number on a chart — a knowledge ecosystem designed to make progress last beyond individual sessions, models, and moments. Growth takes time. Trust takes consistency. And meaningful things are built step by step. To everyone who believes in the vision and contributes along the way: this journey belongs to all of us. 533 is a milestone, not the destination. The story is still being written. Keep building. Keep contributing. Keep moving forward. $SALVOR | #SALVOR 🚀 @SalvorLabsAI @SalvorKnows @BlockchainDan @nitraeX
1
6
13
270
How do you measure AI memory? One-off coding tests like SWE-bench can't show it. Salvor Labs plans its own benchmark to compare memory layers across many sessions. @SalvorLabsAI $SALVOR #AI #LLM #DevTools
2
4
67
nitrateX retweeted
How do you test an AI's memory? Not with one-off coding tasks. Memory only shows across many sessions, so Salvor Labs plans its own long-run benchmark. @SalvorLabsAI $SALVOR #AI #DevTools #LLM @BlockchainDan @nitraeX
2
8
113
nitrateX retweeted
Not a developer? Salvor MCP (alpha) brings the project brain into a chat window. No git commands, no terminal. Ask what the project knows. @SalvorLabsAI $SALVOR #MCP #AI #AIagents @BlockchainDan
1
4
11
162
$Salvor doesn't care what others do! @SalvorLabsAI @BlockchainDan
3
4
66
nitrateX retweeted
One repo, many agents. Claude Code in the morning, Codex in the afternoon. Both read and add to the same .salvor brain, so neither starts from zero. @SalvorLabsAI $SALVOR #AIagents #ClaudeCode #DevTools @BlockchainDan
2
8
113
nitrateX retweeted
This is exactly why Salvor matters. AI agents can write code, but understanding the reasons behind decisions, preserving knowledge, and learning from past experiences is what makes them truly valuable. What matters most is not just the technology, but the people who recognize the problem and contribute to building a lasting solution. Two months of experience, real contributions, and lessons learned. That's what building something meaningful looks like. Respect for sharing your journey. The future is built by those who keep building. $SALVOR 🚀 @nitraeX @BlockchainDan
I've built websites for years. Now AI agents write most of my code, and the why behind decisions got lost between sessions. Two months with Salvor: what changed, what was hard, what I contributed. nitratex.com/articles/salvor… Türkçe için: nitratex.com/tr/articles/sal…
2
6
80
nitrateX retweeted
Your Repo Remembers Coding agents write better code every month. What they still lose is the why. Git tracks what changed. It rarely keeps why an approach was chosen or which idea already failed. Clear the context or switch agents and that reasoning disappears. The next session starts from zero. 2/ Salvor puts selected engineering knowledge inside the repo itself. Decisions, learned failures, domain discoveries, and deferred findings become plain Markdown in a `.salvor` folder. Reviewed in Git, owned by the project, not by any single agent or vendor. Git preserves what changed. Salvor preserves what the project learned. 3/ How it works: - Short current-state note (L1) loads at session start - Deeper archive (L2) keeps reasoning and rejected ideas - Capture classes only become permanent after you approve them - Same knowledge works with Claude Code, Codex, Gemini CLI, and others Nothing sticks until a human says yes. 4/ Once `.salvor` exists, new sessions stop starting from scratch. The reasons sit next to the code. It survives agent switches and team handoffs. Salvor OS is free and MIT-licensed. No account or hosted service needed. Salvor+ is a separate commercial layer in development. 5/ Memory is a guide, not a guarantee. The final check is still yours. Setup is light: paste the prompt, answer four questions, review the files. Start with the one or two things that already hurt. Read the memory once a week. 6/ Agents still write most of the code. The project’s hard-won understanding no longer vanishes when the chat ends. github.com/SalvorLabs/Salvor salvorlabs.org · salvorknows.dev @SalvorLabsAI @SalvorKnows @BlockchainDan @nitraeX
🤖 Made with AI
2
3
6
133
nitrateX retweeted
Love the Labs ↔ OS symbiosis. The recent demo made the relationship between Salvor Labs and Salvor OS crystal clear: Labs can move fast on advanced capabilities while the open-source foundation stays clean, reviewable, and fully under the control of the project. What I’m most looking forward to is the full self-capture audit log, the ratification process, and the activation of new knowledge artifacts only after human review and explicit ratification. That human gate is not a bottleneck—it’s the feature that keeps the project brain trustworthy. Agents can propose Domain Learnings, Learned Failures, and Deferred Findings; humans decide what becomes durable, versioned, repo-owned knowledge. Once the tests are consistently green, those PRs should head back to the Salvor OS repository as soon as possible. The stronger the open foundation becomes, the more valuable the Labs work (MCP, Salvor+, advanced agentic flows, etc.) will be for everyone using it. Your repo remembers. salvorknows.dev @BlockchainDan @SalvorKnows @SalvorLabsAI @nitraeX
🤖 Made with AI
2
3
8
121