Blacksky Cash (Bluesky safe zone) looks very interesting https://blackskyweb.xyz/money-for-mutual-resilience-introducing-blacksky-cash/
“We’re also clear-eyed about risk. Online scams and bots are increasingly sophisticated. While we will actively work to flag suspicious activity and prioritize real communities, users should remain cautious and avoid sharing personal information or engaging with unverified solicitations. Trust is collective work.”
Trust is collective work. We need a better ethos
its the excited tinkerers who look at everything with child like zeal that change the world
karpathy has spoken, and again, it usually puts the casino in a bad light as a major situation
careful when we are so focused on the "base meta" https://x.com/clawinho2026/status/2017342496422936923
i can tell you Pieter said this in a private forum: "see crypto grifters find my Clawdbot" (just no screenshots, kthxbye)
https://x.com/karpathy/status/2017442712388309406
Mitsukeru LinkSentinel just got a massive upgrade.
On-device ML, local OCR, and transaction guards, all keeping this still operating locally, 100% privacy focused.
We have a lightweight Logistic Regression engine, <1 ms inference times, with ~97.6% accuracy on scam templates.
We added OCR via tesseract.js to detect wallet support, etc. text in images, but we also guard for performance.
We do risk assessments before you connect, with also local contract reputation, i.e. if you've interacted before, it gets added to the list. This is Ethereum based, so far, but will extend - so far we look at eth_sendTransaction, eth_sign, and personal_sign, and we're always looking at approve/transfer.
Local first, privacy always.
Clawdbot, OpenAI Atlas, Perplexity Comet, Claude Chrome plugin, and now even Chrome's built in AI are all vectors to what we call "prompt injection".
Mitsukeru today screens/scans content for scams, watches your clipboard, checks URLs before you visit them, and has a strong social engineering focus. Naturally this extends to protecting agents that run on your behalf.
e.g. if your Clawdbot browser had LinkSentinel, it pre-warns it for a bad URL, steers you clear of scam pages, etc. It can also watch for clipboard manipulation of the agent (useful)
what we don't do (yet?) is help with malicious smart contract interaction - this requires Mitsukeru to be less privacy focused, less local, and needs to connect to APIs that also can ban you geographically (iykyk).
We definitely do not protect writing corrupt memories (this might be a need) to files that Clawdbot controls.
So we still need to focus on segmentation and policy to make an AI agent run safely on your behalf and not get damaged by it, and other prompt injections. It is also sometimes how smartly you craft your prompt. Preventing a prompt injection by alerting on <blank> or display:none is actually easy (though you wonder why the LLM itself didn't pick up on it).
As always, how important is this to the end user base? And I have actively warned/cautioned on agents acting on your behalf.
Hope that answers @kenjiquest - and would love to see more feature requests and thoughts like this!
Mitsukeru Preflight.
Reality is, you may already be r00ted. So one should run a malware scan first, before one installs the rest of the power tools.
Does it hurt to check if your Chrome extensions are legit? Every kernel extension is OK? Whatever is running now, is not malicious or suspicious?
Yes, we didn't sign the binary, and it picked itself up as something worth of inspection.
End to end endpoint security. macOS first. Windows next.
when i said social bifurcated, a lot of my opensource buddies went to mastodon.
my choices have always been farcaster, x, threads, and only then bluesky, followed by mastodon.
anyway, some friends started amutable - exciting development for linux https://0pointer.net/blog/introducing-amutable.htmlhttps://amutable.com/blog/introducing-amutable
all will be at FOSDEM this weekend - will you?
so, no X account sales so far.
ryan got elaborate phishing scam via email, showing his tweet. he got his account back.
crystal got phished in DM via a message from Deedy. she's still awaiting, and kimi 2.5 launched hours ago!
Deedy, unclear.
again, elaborate phishing is what on-device AI and ML recognition should prevent you from being attacked. it is Mitsukeru coded
Reputation.
It ain't Ethos.
Geoff (Ralph) & Steve (Gas Town) both feel the brunt of it.
Peter (clawdbot/moltbot) actively wants to distance from meme tokens.
Today it's bagsfun. Sometimes pumpfun. Who remembers Linus Torvalds actively dissing gitfish?
https://x.com/gmoneyNFT/status/2016150274352591139
Don't forget the hack -> launch a CA issue as well https://x.com/crystalsssup_/status/2014571082716713356
Heck, even Amazon Prime doesn't want the Vitalik movie. There are more movies featuring Vitalik compared to Linus - and Linus has arguably made a bigger impact on this world.
Just remember, these are vocal voices. Think about the silent majority.
My Farcaster reputation just hatched a RARE BasedPenguin with 0 legendary traits! @nd10@dogfacts.eth@kelborhal78 y'all need to see what yours looks like 🐧
☝️ created with @neynar app studio
