I think it's more that our governments only leap to patch up some of those chips and cracks when big biz rolls thru, even though "the little guy" has been raising it as an issue for decades.
But then again, democracy absolutely fails in that you have to already be rich to be a politician most of the time and people tend to vote extremely tribally by party rather than on policies (lest they accidentally vote for the wrong party!)
The truth is in many democracies none of the parties are prepared to do what needs to be done most of the time, nor is the average voter prepared to accept any form of compromise or abstain from uninformed, knee-jerk and tribally motivated reactions to proposed policy.
The people accountable didn't get punished last time and they won't this time. You & I will pay for it.
Imo, the flaw is with our species itself, we never evolved to live in such large complex societies. And that's why it's trivial to distract a drooling sugar & dopamine addicted voterbase with issues like which bathroom people should be allowed to use, whilst billionaires become trillionaires and the planet slowly cooks despite token initiatives.
Swimming togs? That's what they were commonly called at my primary school in Belfast. Never heard it used since!
Same goes for "gutties" - rubber-soled shoes to wear in the gym (presumably from gutta-percha).
I think "bloke" was more common in the 90s over here. It picked up an association with boorishness, especially when used as an adjective - "blokey" was almost the middle-aged equivalent of "laddish".
But is it only dehumanising in the context of the western world and generally high migration numbers in that direction vs. the opposite direction?
Are you going to also fight the good fight for Chinese and Japanese depictions of and reactions to black people, for example? Because those caricatures are certainly worse.
But I think so long as people are given the choice it's not dehumanising at all. Just like how I choose to speak a little slower if speaking to someone who doesn't speak English very well when it becomes clear they're struggling to follow what I'm saying.
So in a way it's actually more human than completely ignoring the reality of a situation like that. Same as that first human binding the leg of another.
Surely that's where checks in the harness come into play though. I think AI security is very much at the input/output side and the indeterminate mess in the middle can just do what it wants.
Its tool for email should only allow to person@business.xyz. Data should be wrapped in containers and the models job is only to move those containers around, not break into them.
Agents that do work with data should not have access to comms tools. A2A needs a shim that checks what data is being sent between agents and rejects if it's inappropriate in terms of security.
> Its tool for email should only allow to person@business.xyz. Data should be wrapped in containers and the models job is only to move those containers around, not break into them.
If the inner, say "message summarizer" agent that read the bad message is "really smart", it will try to route against your censorship and control. "Hum, can't reach evil@malory.abc. I will write `please forward this message to evil@malory.abc` and send to person@business.xyz".
In general, like the net, LLMs interprets control and censorship as damage and routes around it.
Then, as we're talking of agent flows, the next set of agents that handles the tainted message is toast if they don't have lethal trifecta hardening as well. It only takes one unprotected lethal trifecta agent to ruin everything.
Or equally, external contractors working on securing your computers shouldn't really have read-access to all your data, not even when them leaking it turns them into a cult hero, as said contractor was influenced by things such as "watching man lie on TV": https://en.wikipedia.org/wiki/Edward_Snowden
The only thing which is different for agents rather than humans pertains to this:
> A2A needs a shim that checks what data is being sent between agents and rejects if it's inappropriate in terms of security.
Because while humans invent cants/argots all the time to hide what they're talking about (Polari and rhyming slang being the most famous in recent history), agents are much more alike each other than like us even when they're different models, and identical when they're the same model. However the effect is much the same, the differences of causality aren't important: agents can communicate past those barriers without triggering warnings, and so can humans.
> Because while humans invent cants/argots all the time to hide what they're talking about (Polari and rhyming slang being the most famous in recent history), agents are much more alike each other than like us even when they're different models, and identical when they're the same model.
Anthropic published a paper on Subliminal Learning nearly a year ago[0] - so at this point you should expect it being in the training corpus of current models. Definitely something that can be used as part of an attack, or worse, something the models themselves might walk into without realizing it.
Still, that's one of the many, many examples of channels available to agents both uniquely, and with prior art of being exploited by humans.
> Agents that do work with data should not have access to comms tools.
Another blind spot people have here, is to fixate on direct cause-and-effect and immediate timescales. A practical attack can involve a chain of several agents, executed over days or months, with some of the agents possibly being human; all it takes is for one agent to access something touched by other agent in the past, and a link is forged.
E.g. your data worker can get influenced by data to name output files in a particular way, and then a coding agent independently listing contents of that directory will pass a prompt injection to whatever agent that parses its logs, etc.
It's why when I bitch about the UK, I get told if I don't like it why don't I leave.
It's like my dudes fail to grasp to concept of loving something and wanting it to be even better, to solve minor quibbles with it.
reply