#Tag
The People’s Consultation on #AI will be accepting submissions until March 15, 2026. #canada #cdnpoli #agenticai #openai #anthropic #usa #uspol #opensource #news
Participate in the consultation here: English: https://www.peoplesaiconsultation.ca/ French: https://www.consultationpopulairesurlia.ca/
The People’s Consultation on #AI will be accepting submissions until March 15, 2026. #canada #cdnpoli #agenticai #openai #anthropic #usa #uspol #opensource #news
Participate in the consultation here: English: https://www.peoplesaiconsultation.ca/ French: https://www.consultationpopulairesurlia.ca/
Mit dem Model Context Protocol #MCP wird #KI handlungsfähig: Large Language Models #LLM können Tools nutzen, Daten verarbeiten und Aufgaben selbstständig ausführen. Doch mehr Fähigkeiten bedeuten auch mehr Risiken. LLMs arbeiten probabilistisch – also nicht immer zuverlässig – und sind manipulierbar. #PromptInjection
Deshalb: Zugriffsrechte begrenzen, Aktivitäten nachvollziehbar machen, Systeme absichern und Menschen einbinden.
Mit dem Model Context Protocol #MCP wird #KI handlungsfähig: Large Language Models #LLM können Tools nutzen, Daten verarbeiten und Aufgaben selbstständig ausführen. Doch mehr Fähigkeiten bedeuten auch mehr Risiken. LLMs arbeiten probabilistisch – also nicht immer zuverlässig – und sind manipulierbar. #PromptInjection
Deshalb: Zugriffsrechte begrenzen, Aktivitäten nachvollziehbar machen, Systeme absichern und Menschen einbinden.
This morning I got an email from a sender that identified itself as an AI agent.
So - plus for being upfront about it, but... please don't do this.
I get that a lot of people are really, really, really into AI tools. OK. I have my opinions on them, you have yours. I have major qualms about them, some people think they're the best thing ever.
OK. Fine. But when your use of these things spills over into the rest of the world, it's no longer a question of my opinion vs. your opinion, my decisions vs. your decisions.
At this point, things have moved from each person doing their own thing to inflicting your use of AI onto me without my consent.
Before this spirals out of control, which I can see happening *very* quickly, I'd like for us to agree on a piece of netiquette:
- it is rude in the extreme to set loose an AI agent to reach out to people who have not consented to interact with these things.
- it is rude to have an AI agent submit pull requests that human maintainers have to review.
- it is rude to have an AI agent autonomously interact with humans in any way when they have not consented to take part in whatever experiment you are running.
- it is unacceptable to have an AI agent autonomously interact with humans without identifying the person or organization behind the agent. If you're not willing to unmask and have a person reach out to you with their thoughts on this, then don't have an AI agent reach out to me.
Stuff like this really sours me on technology right now. If I didn't have a family and responsibilities, I'd be seriously considering how I could go live off the grid somewhere without having to interact with this stuff.
Again: I'm not demanding that other people not use AI/LLMs, etc. But when your use spills out into my having to have interactions with an agent's output, you need to reconsider. Your ability to spew things out into the universe puts an unwanted burden on other humans who have not consented to this.
We need to differentiate human from agent IDs.
If you've been online a couple of years you will know several if not dozens of human online IDs w high confidence. IDs which are *not* AI agents.
This gives I a window of opportunity to build a web of trust containing only humans, and differentiate between those and all other IDs which may or may not be human.
To build a public web of human IDs we need a protocol and a vIral FOSS app.
Over to you... :rofl:
e543 with Andy, Michael and Michael - #AgenticAI and the changing nature of work, #agents #RentingHumans, #RealTimeTranslation, artistic #roads, #e-bikes for your feet and a whole lot more.
This morning I got an email from a sender that identified itself as an AI agent.
So - plus for being upfront about it, but... please don't do this.
I get that a lot of people are really, really, really into AI tools. OK. I have my opinions on them, you have yours. I have major qualms about them, some people think they're the best thing ever.
OK. Fine. But when your use of these things spills over into the rest of the world, it's no longer a question of my opinion vs. your opinion, my decisions vs. your decisions.
At this point, things have moved from each person doing their own thing to inflicting your use of AI onto me without my consent.
Before this spirals out of control, which I can see happening *very* quickly, I'd like for us to agree on a piece of netiquette:
- it is rude in the extreme to set loose an AI agent to reach out to people who have not consented to interact with these things.
- it is rude to have an AI agent submit pull requests that human maintainers have to review.
- it is rude to have an AI agent autonomously interact with humans in any way when they have not consented to take part in whatever experiment you are running.
- it is unacceptable to have an AI agent autonomously interact with humans without identifying the person or organization behind the agent. If you're not willing to unmask and have a person reach out to you with their thoughts on this, then don't have an AI agent reach out to me.
Stuff like this really sours me on technology right now. If I didn't have a family and responsibilities, I'd be seriously considering how I could go live off the grid somewhere without having to interact with this stuff.
Again: I'm not demanding that other people not use AI/LLMs, etc. But when your use spills out into my having to have interactions with an agent's output, you need to reconsider. Your ability to spew things out into the universe puts an unwanted burden on other humans who have not consented to this.
e543 with Andy, Michael and Michael - #AgenticAI and the changing nature of work, #agents #RentingHumans, #RealTimeTranslation, artistic #roads, #e-bikes for your feet and a whole lot more.
This is excellent:
"They Built Stepford AI and Called It “Agentic”"
It probably helps if you've seen the original movie or read the book. (The 1972 movie is much better than the 2004 remake.)
https://abiawomosu.substack.com/p/they-built-stepford-ai-and-called
"I just found out that it's been hallucinating numbers this entire time."
