#Moltbook is the most interesting place on the internet right now - https://simonwillison.net/2026/Jan/30/moltbook/ "Here’s #Clawdbot buying AJ Stuyvenberg a car by negotiating with multiple dealers over email." probably the future. alas. #agenticai
The Agentic AI Handbook: Production-Ready Patterns
https://www.nibzard.com/agentic-handbook
#HackerNews #AgenticAI #Handbook #ProductionReady #Patterns #AI #Development #TechTrends #Innovation
Signal leaders warn agentic AI is an insecure, unreliable surveillance risk
#HackerNews #Signal #AI #surveillance #insecure #unreliable #agenticAI #cybersecurity
Coywolf
'Signal' President and VP warn agentic AI is insecure, unreliable, and a surveillance nightmare
With agentic AI embedded at the OS level, databases storing entire digital lives accessible to malware, tasks whose reliability quickly breaks down at each step, and being opted-in without consent, Signal leadership is sounding the alarm for the industry to pull back until threats can be mitigated.
⁂
More from
Jon Henshaw
https://steve-yegge.medium.com/welcome-to-gas-town-4f25ee16dd04
Does anyone who actually plays with this stuff have a view on qhether this is a powerful evolution or simply a fever dream?
It doesn't remotely address the moral, IP theft etc issues of LLMs but does it start to generate some kind of simplistic "memory" and "understanding" analogues?
I get the impression we are still the traditional "few months" before "lift off".
Andreas Wagner
and 1 other
boosted
📣 NEW w/#UdbhavTiwari Mapping the technical reality & privacy/security perils of pushing AI agents into our infra
We offer palliatives, but the core issues are paradigmatic: 'agency' relies on pervasive data access + ability to act w/o explicit consent.
RE: https://mastodon.world/@Mer__edith/115854211176763097
Excellent #39c3 talk on so-called "agentic AI" and how it's infiltrating into operating systems. Key quote from the end: "Without implementation of the proposed [palliatives] we risk locking ourselves into a digital infrastructure where we are no longer the users of our devices but the managed resources of an automated economy" #agenticAI #security #agency #consent
📣 NEW w/#UdbhavTiwari Mapping the technical reality & privacy/security perils of pushing AI agents into our infra
We offer palliatives, but the core issues are paradigmatic: 'agency' relies on pervasive data access + ability to act w/o explicit consent.
RE: https://mastodon.world/@Mer__edith/115854211176763097
Excellent #39c3 talk on so-called "agentic AI" and how it's infiltrating into operating systems. Key quote from the end: "Without implementation of the proposed [palliatives] we risk locking ourselves into a digital infrastructure where we are no longer the users of our devices but the managed resources of an automated economy" #agenticAI #security #agency #consent
I wrote the manual Karpathy said was missing for agentic AI
https://github.com/nicolasahar/morphic-programming
#HackerNews #agenticAI #morphicProgramming #Karpathy #manual #AIdevelopment #techinnovation
I looked at just one of those multi-agent-systems, called MAS-ZERO. In its minimal setting, it dispatches 5 concurrent queries to e.g. gpt-5 and loops that 10 times.
Every one of these queries is turned into tens of thousands of tokens by the system prompt, and routinely dispatches secondary queries to 5 other models. So if your initial query was 100 tokens, this turns it into 10,000 x 5 x 50 tokens, so several million.
And that is without even going to "reasoning"
(2/3)
With reasoning, there is yet another explosion of at least 10x, often 100x.
Which takes us rapidly into tens to hundreds of millions of tokens for a single query that started out as a few hundred tokens.
This is why Google now processes quadrillions of tokens monthly.
(3/3)
#FrugalComputing #AgenticAI #GenAI
Today we launch the Agentic AI Foundation (AAIF) with project contributions of MCP (Anthropic), goose (Block) and AGENTS.md (OpenAI), creating a shared ecosystem for tools, standards, and community-driven innovation.
Learn more about this major step toward open, interoperable agentic AI: https://www.linuxfoundation.org/press/linux-foundation-announces-the-formation-of-the-agentic-ai-foundation
#AgenticAI #AAIF #OpenSourceAI #AIStandards #Interoperability #OpenInnovation #LinuxFoundation #Anthropic #Block #OpenAI #MCP #goose #AGENTSmd #FOSS #TechNews
Debby ⁂📎🐧:disability_flag:
boosted
I see google’s new agentic AI coding platform is going well.
RE: https://aus.social/@bastardsheep/115641546050077309
If you're going to allow an AI agent to work on your computer on your behalf, create a separate user account for it and only give it access to the specific resources it will control. Follow the principle of least privilege.
This is also true, by the way, if you allow other people to do work on your computer on your behalf. Both people and AIs can make terrible blunders.
It's bad enough that you might accidentally delete a needed resource. You don't want to make it more likely by allowing an AI or another person to also have such access.
Carlo Gubitosa :nonviolenza:
boosted
If you don't actually know how #genAI and #agenticAI works, then you shouldn't let it anywhere near your data (or anyone else's).
If you do actually know how they work, then you *know* not to let it anywhere your data (or anyone else's).
@DanielMReck not even in an on-prem, self-hosted machine with LLMs running locally, no exposure to the public intenet and available only inside the corporate VPN?
@gubi In this particular instance, user reportedly gave the #agenticAI direct access to execute commands on their local file system.
Doesn't really matter whether the #AI is on-prem, self-hosted, running locally, with no access to the public internet… It had access to his local data, and it used that access to destroy the data.
Even isolated AI is dangerous AI, especially if the user assumes that it is competent and does not triple verify every AI-proposed line of command/program execution.
If you don't actually know how #genAI and #agenticAI works, then you shouldn't let it anywhere near your data (or anyone else's).
If you do actually know how they work, then you *know* not to let it anywhere your data (or anyone else's).
I see google’s new agentic AI coding platform is going well.
RE: https://aus.social/@bastardsheep/115641546050077309
If you're going to allow an AI agent to work on your computer on your behalf, create a separate user account for it and only give it access to the specific resources it will control. Follow the principle of least privilege.
This is also true, by the way, if you allow other people to do work on your computer on your behalf. Both people and AIs can make terrible blunders.
It's bad enough that you might accidentally delete a needed resource. You don't want to make it more likely by allowing an AI or another person to also have such access.
Fara-7B: An efficient agentic model for computer use
https://github.com/microsoft/fara
#HackerNews #Fara7B #EfficientModel #AgenticAI #ComputerUse #MicrosoftGitHub
The ever-delightful @adamengst joins me to talk about agentic web browsers. I think you’ll get a kick out of his assessment as he puts them through their paces using phrases like “full cuckoo”, “daydreams in class”, “hapless third graders”, requiring “academic integrity referral”, and “like a scared, insecure intern”.
🦎 SUSE Linux Enterprise 16 Announced: "Enterprise Linux That Integrates Agentic AI" - Phoronix
｢ SUSE's announcement today for SUSE Linux Enterprise 16 proclaims SLES 16 to be "the industry's first enterprise Linux that integrates agenetic AI" and "reduces operational costs and complexity through AI readiness." SUSE Linux Enterprise 16 supports the Model Context Protocol (MCP) standard and can bridge to any LLM provider ｣
Lessons from 70 interviews on deploying AI Agents in production
https://mmc.vc/research/state-of-agentic-ai-founders-edition/
#HackerNews #LessonsFromInterviews #AIAgents #Deployment #AIInProduction #TechInsights #AgenticAI
Syllabi – Open-source agentic AI with tools, RAG, and multi-channel deploy
#HackerNews #Syllabi #OpenSource #AI #agenticAI #RAG #multiChannelDeploy