The problem is obvious: People spend much more attention on cat videos from strangers than on their own friends' posts. Ads turn this attention into money.
A 12.48 Waveshare eink display costs $175.
Sadly haven't gotten it to work with the Raspi Zero and therefore can't use it battery-powered. Got an ugly cord right now. Running power to the right place through the walls is definitely dedication!
That's the point though. The testers wouldn't actually abuse their victims without the conviction of doing something righteous. Or they would, accidentally or intentionally, spill the secrets.
But if you make even the instruction material lie, then there is nothing that could be leaked and "expose" the system.
I always thought the workings of polygraphs were common knowledge.
It's fiction. Analysts get scared and don't do anything wrong preemptively. Analysts admit stuff they'd never do otherwise. The agency gets to show who's in charge. It creates a legal fiction that allows you to abuse your employees. It creates a fiction that the abusers themselves can believe in.
Why should the believe in the non-working polygraph be any weaker than in a nonexistent god?
That's a fair point. I think the distinction is between software that follows deterministic rules (your 2-week-delay scenario) vs agents that make autonomous decisions based on learned patterns. With traditional software, intent is clear and traceable. With AI agents, the operator may genuinely not know what the agent will do in novel situations. Doesn't absolve responsibility — but it does make the liability chain more complex. We probably need new frameworks that account for this, similar to how product liability evolved for physical goods.
If the code you wrote appears to be for something completely different, say software to write patches for open source github projects - yes. Why would you bear responsibility for something that couldn't have been reasonably foreseen?
The interesting thing about LLMs is the unpredictable emergent behaviours. That's fundamentally different from ordinary, deterministic programs.
>fighter jets ARE a threat of violence, and it is widely understood and acknowledged.
That's not a credible threat because there's approximately 0% chance France would actually follow through with it. Not even Trump would resort to murder to get rid of his domestic adversaries. As we seen the fed, the best he could muster are some spurious prosecutions. France murdering someone would put them on par with Russia or India.
Don’t forget that captain of the plane makes decisions not Elon.
If captain of the plane disobeyed direct threat like that from a nation, his career is going to be limited. Yeah Elon might throw money at him but that guy is most likely never allowed again to fly near any French territory. I guess whole cabin crew as well .
Being clear for flying anywhere in the world is their job.
Would be quite stupid to loose it like truck driver DUI getting his license revoked.
>Don’t forget that captain of the plane makes decisions not Elon.
>If captain of the plane disobeyed direct threat like that from a nation, his career is going to be limited. Yeah Elon might throw money at him but that guy is most likely never allowed again to fly near any French territory. I guess whole cabin crew as well .
Again, what's France trying to do? Refuse entry to France? Why do they need to threaten shooting down his jet for that? Just harassing/pranking him (eg. "haha got you good with that jet lmao")?
I think the implication of the fighter jets is that they force the plane to land within a particular jurisdiction (where he is then arrested) rather than allowing it to just fly off to somewhere else. Similar to the way that a mall security guard might arrest a shoplifter; the existence of security guards doesn't mean the mall operators are planning to murder you.
Guards can plausibly arrest you without seriously injuring you. But
according to https://aviation.stackexchange.com/a/68361 there are no safe options if the pilot really doesn’t want to comply, so there is no “forcing” a plane to land somewhere, just making it very clear that powerful people really want you to stop and might be able to give more consequences on the ground if you don’t.
Planes are required to comply with instructions; if they don't they're committing a serious crime and the fighters are well within their international legal framework to shoot the plane down. They would likely escalate to a warning shot with the gun past the cockpit, and if the aircraft is large enough they might try to shoot out one engine instead of the wing or fuselage.
I suspect fighter pilots are better than commercial pilots at putting their much-higher-spec aircraft so uncomfortably close that your choices narrow down to complying with their landing instructions or suicidally colliding with one - in which case the fighter has an ejector seat and you don't.
I felt like you ruled out collision when you said they're not going to murder, though, granted, an accidental but predictable collision after repeatedly refusing orders is not exactly murder. I think the point stands, they have to be willing to kill or to back down, and as others said I'm skeptical France or similar countries would give the order for anything short of an imminent threat regarding the plane's target. If Musk doesn't want to land where they want him to, he's going to pay the pilot whatever it takes, and the fighter jets are going to back off because whatever they want to arrest him for isn't worth an international incident.
In the USA they would be allowed to down any aircraft not complying with national air interception rules, that would not be murder. It would be equivalent to not dropping a gun once prompted by an officer and being shot as a result.
reply