Riven

After today’s incident, I’m concerned about how private keys are handled across self-custody setups.I built a quick tool to model your setup, device by device. No seed words or private data. Everything runs in your browser. keysafety.bitcoinvseverything.com @bitcoinvsall PS: I’m not a cybersecurity/cryptography expert. Please check my assumptions and tell me what I got wrong.

@btcbenchmark I am looking for numbers to get a sense of risk. Will it be 1000 times harder, or a billion times, etc I don’t think there is a fool proof setup, just want to get a sense of risk/reward

@CharonVoyage It prevents this

THERE HAS BEEN A SERIOUS CYBER ATTACK ON BITCOIN USERS JUST NOW Do not panic, do not remake your whole self-custody setup overnight when stressed, it's more likely you make a mistake than that your coins are at risk Also, soon there will be fake phishing, security audit and recovery services appearing all over the internet, be extremely cautious on anyone contacting you directly about this issue offering their help Here are the facts (source: stacker news) A single address collected over 590 BTC in under an hour on the evening of 29 July. More than 1300 UTXOs were spent toward it, from multiple independent victims About 10 victims identified so far This is what makes it weird: many wallets were only partially drained: some UTXOs taken, others left behind, sometimes including large ones. (My current thesis is that the attacker saved computational resources by only searching for certain sized UTXOs) Address reuse does not explain the majority of cases, and since every transaction was p2wpkh, a long-tail quantum attack is ruled out Sweeps were not confined to BIP-84 paths Coinkite's nvk stated they had received no support emails about it, while others in the industry independently confirmed the issue What bitcoiners should do now? The right question isn't "which brand do I use” it's where did the entropy for my seed actually come from, and has that seed ever touched a general-purpose computer? A seed you generated in 2020 on something you no longer remember, now sitting in a shiny new device, is exactly as weak as the day it was made The lesson isn’t "coldcard bad” but rather that single-vendor custody means a single vendor's RNG bug is a total loss. Multisig across two different vendors, or seeds you generated yourself with dice, prevent this certainly After all, hardware wallets are just signing devices!

@utxoclub This is not airgapped!

The 4 byte security check code after a Distributed Key Generation:

Today was extremely validating for our core security principle: never trust a device to generate its own randomness. For the vast majority of bitcoiners today, hardware wallets are a trusted third party. Even with manual rituals (rolling dice, flipping coins, drawing words from a hat), a device must still convert that secret entropy into public keys and must be trusted to do so faithfully. Distributed Key Generation and related randomness protocols solve the problem properly: every device contributes entropy, every contribution is verified by the others, and every participant verifies the resulting public key through the protocol itself. For the user, it's a simple 4 digit number.

How much safer is multi-sig compared to the MK3 fiasco baseline?

@rileybrown @jack But tagging is becoming problematic on the desktop for agents on the Relay/VPS. This is clearly a beta product.

@rileybrown @jack I figured it out. I spun up an agent on the relay hosting Buzz on my VPS, so I can now access it from my iPhone. It's not as seamless as I'd like, but it works.

The Buzz iOS app is very underrated... I've been messaging my Codex, Claude Code, and Cursor from my phone all morning... and it feels like a group chat, and they can even collaborate. No new subscriptions... free @jack and co are onto something...

Fauci’s 'performance' today made me sick. Smug arrogance. Zero remorse. Zero accountability. He refused to answer a single question. Sharing this again as a reminder: He used your tax dollars to fund the coronavirus lab research in Wuhan—and worked with the IC to cover it up. We deserve the truth and accountability.

True

@calcsam Does it work with already captured langfuse traces?

Trace Intelligence is built into observability on the Mastra platform. Read more and sign up for beta: mastra.ai/blog/announcin…

Today we're incredibly excited to launch Trace Intelligence. Teams shipping agents can spend months reviewing traces. So we auto-group traces into clusters and identify common patterns. Watch how your agent is actually doing:

This may be unpopular, but we never escaped the 2008 financial crisis. We just shifted the pain.

@theo How are you accessing kimi3 ? Have you added moonshot API keys in Claude env?

Okay technically it's 4 clicks.

@RadarChat Very cool, but also need Desktop app as I cannot use signal anymore!

Your messages. Your Bitcoin. Together, at last. Radar brings private messaging and self-custodial Bitcoin Lightning together in one seamless experience, and because it's built on Signal's incredible network - the people you already talk to come with you.

Databricks is embracing Chinese open source model GLM-5.2. Earlier, Microsoft said they were looking to embrace DeepSeek. Major American software companies are embracing Chinese open source models to cut token costs. I have long advocated using the Chinese models running in one's own infrastructure. We have Zoho-specific small models that we use in production. This approach also will become common for larger companies, as it becomes easier to set up the training pipeline to train your own org specific model.

@thebitmela @bitshala_org @studentofbtc @bitshala @RunwithBitcoin @BitashaIndia @bitcoinbharat_ @TheBitcoinGirl_ @KrnGill5 @AmityAge @theAmberApp Telegram group for attendees, anyone? 😆

VENUE IS LOCKED. DATES ARE IN. ✨ OCT 30 & 31 ✨ India, it's happening! #india #bitcoin #bitmela

₿uild them a product they’re allowed to buy. Bitcoin Capitalism — in 2 minutes.

This is super exciting - I’ve been using Omnigent at Databricks for a while, and today we open-sourced it. Omnigent is a meta-agent for orchestrating a swarm of agents. Why do we need this? The best results no longer come from a single model running in a single harness. I used to run the same task with Codex and Claude Code, then pick the better one. But the obvious thing is to let them collaborate, debate, and converge on something better. Omnigent makes this smooth. The other feature I love is real-time collaboration. You can invite people into an Omnigent session to watch, steer, and send commands. Multi-agent, multi-human collaboration is the future. Omnigent was built by @matei_zaharia and a very lean team in just 6 weeks, working every day out of a Databricks war room, truly amazing. Databricks AI really feels like a startup.

Introducing 𝗢𝗺𝗻𝗶𝗴𝗲𝗻𝘁, a meta-harness to combine, control, and share your agents. The best teams already mix models and harnesses and design loops that drive teams of agents. No single harness can keep up with that alone. So we built the layer above — we call it a 𝗺𝗲𝘁𝗮-𝗵𝗮𝗿𝗻𝗲𝘀𝘀. Omnigent sits above the tools you already use, Claude Code, Codex, Pi, and your own agents, and gives them one shared layer: • 𝗖𝗼𝗺𝗽𝗼𝘀𝗶𝘁𝗶𝗼𝗻: combine models, harnesses, and techniques without rewriting code, and switch between them with one-line changes • 𝗖𝗼𝗻𝘁𝗿𝗼𝗹: stateful, data-centric policies and cost budgets enforced at the meta-harness layer, not via prompts — let agents run without watching them • 𝗖𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗶𝗼𝗻: share a live agent session via URL with full history, so teammates can review, comment, and steer in real time Every session is reachable from a terminal, the web, a desktop app, and your phone. We built Omnigent for our own use at Databricks and are open sourcing it under Apache 2.0. databricks.com/blog/introduci…

Really excited to open source a new project: Omnigent, a meta-harness for AI agents. It lets you build multi-agent coding and custom agents, sitting above Claude Code, Codex, Pi, and agent SDKs to let you compose them. It also adds live collaboration and rich control policies.