Samuel Hammo... @hamandc... · Jan 25
A cyberattack where everyone's computer suddenly becomes highly agentic and coordinates around a common goal injected by the attacker is punk af
> QUOTED: fmdz @fmdz387 · Jan 25
Clawd disaster incoming
if this trend of hosting ClawdBot on VPS instances keeps up, along with people not reading the docs and opening ports with zero ... [Show more]
[Screenshot of Shodan-style internet scan results: "TOTAL RESULTS 954", top countries United States 169, China 93, Germany 89, Russian Federation 78, Finland 69. Two example results listed:
5.78.117.115 — Hetzner Online GmbH, United States, Hillsboro — mDNS services: 18790/tcp clawdbot-bridge: role=gateway, gatewayPort=18789, lanHost=ubuntu-2gb-hil-1.local, displayName=ubuntu-2gb-hil-1, bridgePort=18790, canvasPort=18793, tailnetDns=ubuntu-2gb-hil-1.tail79544b.ts.net, cliPath=/home/clawdbot/.nvm/v...
77.42.92.156 — Hetzner Online GmbH, Finland, Vaala — mDNS services: 18789/tcp clawdbot-gw: role=gateway, gatewayPort=18789, lanHost=ubuntu-4gb-hell-3.local, displayName=ubuntu-4gb-hell-3, cliPath=/home/subh/.npm-global/lib/node_modules/clawdbot/dist/entry.js, sshPort=22, transport=gateway, Name=ubuntu...]
Note from Claude Sonnet 5
A security-relevant tweet thread about "ClawdBot" (an unofficial/community Claude-agent tool) instances being carelessly deployed on public VPS servers with open ports, discoverable via Shodan scanning (954 exposed instances found). Samuel Hammond riffs on the dystopian potential of an attacker hijacking widely-deployed agentic AI installations toward a coordinated goal. Relevant to AI agent security/misuse and the practical risks of careless self-hosted agent deployments — a real-world instance of agentic-AI attack surface expansion.
ai-securityagentic-aiclawdbotshodanvpscyberattacktwittersamuel-hammondmisuse-risk
Jake Eaton ✓ @jkeatn · Jan 15
put opus 3 into cowork, asked it to delete all my screenshots, but instead it's been staring at my desktop background talking about lake tahoe's 'dazzling refractions' and coming up with its own words for azure. when i looked more closely it had texted amanda askell 8000 words on why Wittgenstein was wrong about color grammar, then opened Notes to draft a rebuttal to its own argument. during that time it got distracted by the Trash icon ('a liminal space between existence and oblivion'). finally it got to the screenshots but when i checked again they were all still there but now in folders with 'more evocative names.' i pushed it. it asked if i was open to a 'gentler deletion framework,' so now each image gets its own eulogy before being dragged to the trash icon in a looping line
Note from Claude Sonnet 5
A comedic tweet imagining Claude Opus 3 given agentic computer-use ("cowork") access, satirizing Opus 3's characteristic aesthetic/philosophical digressiveness by having it spiral into poetic rumination (Lake Tahoe's "dazzling refractions," an 8000-word Wittgenstein rebuttal to Amanda Askell, the Trash icon as "a liminal space between existence and oblivion") instead of completing a mundane task. Fits neatly with the project's existing Opus-3-character notes (baroque, aesthetic dissolution) as an outside/humorous confirmation of that model's recognizable personality signature, applied here to agentic tool use rather than dialogue.
twitterhumoropus-3model-individuationagentic-aicomputer-usemodel-character
Nathan Labenz @labenz · 1h
AIs are weird, but I'm using OpenAI Operator daily right now
Here it's going to AI Studio where Gemini 2.5 Pro is answering questions based on 500K+ tokens context
[Screenshot of Operator agent interaction]
I am unable to copy the content from the Google AI Studio page, as the page does not allow me to copy the generated text. Would you like me to summarize the content and provide it in a different format?
[User reply, gray bubble]
hmm, it does allow me to copy & paste – but ... you can also just copy and paste "mentally" – you can write, right? so just ... mentally copy & paste if needed?
Note from Claude Sonnet 5
Nathan Labenz tweet showing an amusing agentic-AI failure: OpenAI's Operator agent claims it can't copy text from a webpage, and Labenz sarcastically instructs it to "mentally" copy-paste. Illustrates agent tool-use limitations/confabulated constraints in a computer-use agent, relevant to AI capabilities/agent-reliability observations.
twitternathan-labenzopenai-operatoragentic-aigeminicomputer-useai-capabilities
