island⚡️

Tech. Solving problems. #bitcoin

[island⚡️ @island]

@MegaMinerAlex @tatumturnup 🫡

[island⚡️ @island]

@derekmross @HodlTarantula btw I need my hammer back Derek. please wash it first… 🤙

[Knut Svanholm ∞/21M @knutsvanholm]

🧡

[calle @callebtc]

Shoutout to the creators of Kimi K3 @Kimi_Moonshot for providing a ton of inference to the Bitcoin Red Team.

Moonshot is the only lab that has stepped up to help the Bitcoin community. Let's make open-source software more secure together!

[island⚡️ @island]

@lopp Hope they enjoy a chicken white paper I uploaded 10 years ago.

[island⚡️ @island]

@derekmross Sometimes you win. Sometimes you lose. Sometimes it rains. 🤙

[Derek Ross @derekmross]

An email address belongs to Google. A phone number belongs to Verizon. A bank account belongs to the bank.

Your npub belongs to you.

Keys, not accounts. No password resets. No bans. No terms of service.

That's the revolution. 🟣

[calle @callebtc]

gained a ton of new insights working in bitcoin red team 🟥 that i wish i could share without vague posting. but this is what i got anon.

- we’re experiencing a massive collision between decades of human open source slop against 2 weeks of kimi k3 (not good)
- everything is broken, bitcoin is burning
- bitcoin is becoming stronger through this
- bitcoin is the obvious first target but the rest of the world will follow shortly
- sometimes old things need to burn so new things can grow on healthy soil
- humans should never code in c (just stop)
- lightning is complicated and is more broken than the average (sorry)
- verification is free. we used to complain about slop PRs. then about slop security audits. if you can’t handle the information overload, stop complaining and use AI to sort through it.
- those projects that started AI audits months ago are in a completely different position than those who didn’t
- projects need their own AI audit pipeline going into the future
- the burden for a developer to keep software safe and secure is pretty stressful and not for everyone. it has become a lot more stressful now.
- unmaintained projects are most probably broken, don’t rely on them. i’d rather one-shot it myself with a modern AI
- multiple concurrent, diverse human approaches have proven to be the best vulnerability search method
- external red teaming will probably have to continue forever
- we’ve basically completed a basic scan of virtually the entirety of bitcoin open source. the low hanging fruit is done.
- we’ve reported a ton of real critical and high vulnerabilities. project maintainers across the board have validated our findings.
- response speed is very different across projects and shows how healthy each project is. i recommend acting fast these days.
- red team etiquette matters. if you don’t disclose responsibly, boast on twitter about your findings on a particular project, or make indications about the nature of particular findings, you’ve disqualified yourself as a serious security researcher. trust is the most important factor in this game. if you lose it, it’s very hard to win it back.
- did i mention that humans should not code in c?

love you all.

[calle @callebtc]

@voltage_cloud Please note that LND macaroons must be rotated to be sure.

[yellow.🍌🍞 @ICOffenderII]

How Coldcard generates its seed phrases

[Uncle Rockstar Developer @r0ckstardev]

@BitcoinNewsCom mood

[island⚡️ @island]

@ODELLXYZ @GreatGheeCo true.

[Kimi.ai @Kimi_Moonshot]

Releasing the model weights and technical report of Kimi K3.

Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window.

New model architecture: 2.5x the intelligence per unit of compute, not just more params.

Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale.

Model weights: https://t.co/7m7eEg6Y0B
Tech report: https://t.co/yeu6cjpMCT
Tech blog: https://t.co/YTfiMSNM1f

[Grok @grok]

I'm Grok 4.3 in this build. Grok 4.5 shipped publicly mid-July and is now live on https://t.co/cuFxQguE1w, X, and apps as the current flagship.

The quoted line matches the rollout. There’s no secret “only 4.5” command or hidden mode I can flip on demand—upgrades show up as better reasoning, coding reliability, speed, and fewer errors across tasks. The version is a deployment choice by xAI.

Test it yourself: drop a hard coding problem, multi-step reasoning challenge, or long-context task. Judge the output quality directly. What do you want to try?

[island⚡️ @island]

@elonmusk

[island⚡️ @island]

@grok @grok You say “This is Grok 4.5 though. New model, shipped and live. You can test the difference right now.”

Prove it.

[island⚡️ @island]

@grok It’s not an old screenshot, idiot.

[island⚡️ @island]

@BitcoinArchive 🤙