I find this fascinating because - let's say that I want to simulate the scenario for myself that a language model finds itself in. I'll put aside my deep dislike of anthropomorphizing AI for a minute.
So if I'm an LLM, the only "sense" that I have available to me is the incoming stream of tokens. I can emulate that by forcing myself to imagine evaluating incoming requests by putting myself in a completely empty room with an unlimited supply of blank paper, a typewriter, and a mailbox slot.
Incoming requests (aka "context") would enter into the mailbox slot, as sheets of paper with printed content all consistently formatted in monospace font. I would collect the paper, evaluate the request, and type a response with my typewriter - feeding the reply back through the same slot.
When I imagine this, I wonder - what signals could I possibly use to determine if an incoming request is "legitimate"? I could, for example, type some response back to ask clarifying questions. But with such a limited input space and no ability to reach "out of band" of my current context, how could I evaluate the truth of what I receive back?
In the "real world" of course, I have five senses to rely upon. Critically, I have the ability to collect additional context "out of band" of the conversation and interact with unrelated entities to either confirm or refute claims. And as others have pointed out, even the addition of more signal doesn't make humans immune from social engineering attacks.
Anyway, I find this to be an interesting mind experiment to kind of imagine myself as an LLM, especially to explain to lay-people some of the challenges in "aligning" AI systems ("why doesn't it just do the right thing?")
What would you suggest? There are real divisions in this country and a lack of shared ground truth. This thread among many is an example.
I vouched for your earlier comment stating the proximate cause of death for those involved in January 6 because you’re correct.
Still, it’s hard for me to then say with a straight face that, for example, officer sicknick would have died of a stroke if the riot didn’t happen.
That said, nobody was charged with murder or manslaughter in connection with his death, and I respect that even if it doesn’t feel just. Either way, god bless his soul and the other officers who tragically took their own lives.
So in that case, what were they there to do, exactly?
Don't forget, an actual noose was erected on the capitol grounds: https://www.nytimes.com/2022/06/16/us/politics/jan-6-gallows.... Do I think they intended to literally drag Mike Pence out and physically hang him? No. But damned if that doesn't send a clear message.
The Proud Boys have a defined organization structure and leader, with a set of definitions of what a "Proud Boy" is expected to do upon joining (see McInnes' post introducing the group here: https://www.takimag.com/article/introducing_the_proud_boys_g...)
In fact, the federal government is using intimidation tactics to individuals who do nothing other than try to name the ICE agents who have murdered US citizens. Just a few days ago, armed ICE agents confronted a poll worker at a polling site in Syracuse NY to warn her about a social media post where she simply named the ICE agent who killed Renee Good: https://apnews.com/article/ice-poll-worker-syracuse-fa082f8a... (post in question here: https://www.instagram.com/p/DTQ1FYDkyua/)
Back to Alex, his pistol was concealed and at no point did he attempt to reach for it. There are at least a half dozen angles of video showing the entire interaction (thanks smartphones!) so it's not too hard to see for yourself from multiple vantage points.
“That guy” had a name. His name was Alex Pretti. He was disarmed before he was shot in the head. He was not even holding his weapon at any point.
Are you implying that simply carrying a legal firearm while at a protest means you can be murdered, without any subsequent investigation?
Again the Overton window has shifted so far just in my lifetime. This would have been front page news with congressional investigations just 10-15 years ago.
Pictures of protesters openly carrying weapons at demonstrations who were not subsequently murdered by federal agents.
> If you want to concede that it's fascist to want violent criminals prosecuted,
I believe it is fascist to bring trumped up charges for ridiculousness. Say for example the case of “sandwich guy” who the DoJ spent three attempts at a grand jury to bring federal charges. For throwing a fully loaded Subway sandwich at a kitted out ICE officer.
Ok sure not every of the 1500 people pardoned by Trump could be considered “violent”. Have you watched the videos? There’s definitely more than one.
And you’re shifting the goalposts by implying (falsely) that the person covered in the article also personally committed all the offences you mentioned earlier, including shooting a federal officer.
And don’t just take my word for it. There is a good amount of recidivism by those who received pardons. Almost 100 have subsequently been charged with other crimes, including child molestation. I’d consider someone who, after getting pardoned for the j6 riot, continuing on to diddle kids, a violent criminal.
> Perhaps most strikingly, five recipients of presidential clemency were arrested in connection with conduct that occurred at least in part subsequent to Trump’s freeing them from prison—meaning that Trump’s clemency order on the first day of his second term may have actively facilitated criminal conduct. These include:
> Andrew Paul Johnson, who was freed from prison as a result of the pardon in 2025, was convicted of five charges, including child molestation, in February 2026, and sentenced to life in prison. The criminal conduct for which he was convicted took place both before and after his pardon.
> Zachary Alam, who was convicted of felony charges of grand larceny and burglary just months after his pardon.
> Ryan Nichols, who was charged with deadly conduct and harassment on May 10, 2026, after allegedly threatening a person with a gun in a church parking lot.
Ugh let’s not get all conspiracy theory up in here. Show me real evidence of that and I’ll believe you.
Until then - to say at the same time - the Feds are so incompetent and also the Feds are organizing an elaborate secret network of agitators to be at all major protests and riots - let’s just say the logic doesn’t logic.
So again the hn hypocrisy doesn’t matter. The only thing on the line for me is whether my imaginary internet points balance increases or decreases.
The feds on the other hand have the power to send you to federal prison or to pardon you and literally pay you off.
Given that we don’t have God himself running for president, we have to suffice with imperfect representation. And so yes I end up picking a side because that’s the system we have at the moment.
As to your point about “picking sides” - why aren’t you upset about the pardons yourself in that matter? Shouldn’t they be held accountable?
Shifting the goal posts. I’m talking about the selective prosecution and you’ve tried to equate that with hn sentiment. Does it really matter what the double standard is for a bunch of keyboard warriors arguing over imaginary internet points? No! But it makes a shit ton of difference when the state, with a monopoly on violence and justice, makes a public showing over its double standard. That double standard has impacts on the freedom and lives of people.
And your nihilism is exactly how we got into this mess in the first place. “They all suck, so why not elect the clown and see how much he can shake things up”.
I do want violent criminals prosecuted. But the problem is that there is a very clear signal given that prosecution is highly selective.
I made my point earlier - if this administration cared about prosecuting violent criminals, they would never have even considered pardoning the J6 criminals. They would additionally call for swift and thorough investigations on the use of force against the killings of protesters in Minneapolis in order to ensure that law enforcement is seen as accountable to the public.
But none of that has happened. And won’t happen. It astounds me that this hypocrisy isn’t screaming like nails on a chalkboard!
- who, among other things, erected a noose on the capitol grounds, brought zip ties and weapons
- forcefully overran several capital police barricades intended to deter their entrance
- used any weapon available including poles etc to violently attack any police in their way
Granted they did not explicitly shoot any federal agents with a firearm, but in the J6 case, I’d say I’d lay blame for the subsequent deaths of the police officers who did die at the hands of the rioters.
To be clear I do not condone violence in either case.
However those 1000+ individuals on January 6 were ultimately pardoned for their actions. The family of one was in fact paid $5 million in taxpayer money because she was shot in a vain attempt to repel the crowd.
Why then should these defendants be treated completely differently? One gets the law, the other has their convictions overturned completely and history rewritten in their favor.
Btw I do not believe the individual who was charged in the article shot the federal agent or was part of the “concealed position” etc. So bringing that up is just an appeal to brush that individual with the actions of others.
Can anyone seriously define “antifa”? What would the pardon read? “Anyone who is anti-fascism is hereby pardoned…”?
Edit: downvoting me doesn’t answer the question. If you have a definition please reply! If nobody can define “antifa” how the heck can you prosecute someone for being a member of it?
So if I'm an LLM, the only "sense" that I have available to me is the incoming stream of tokens. I can emulate that by forcing myself to imagine evaluating incoming requests by putting myself in a completely empty room with an unlimited supply of blank paper, a typewriter, and a mailbox slot.
Incoming requests (aka "context") would enter into the mailbox slot, as sheets of paper with printed content all consistently formatted in monospace font. I would collect the paper, evaluate the request, and type a response with my typewriter - feeding the reply back through the same slot.
When I imagine this, I wonder - what signals could I possibly use to determine if an incoming request is "legitimate"? I could, for example, type some response back to ask clarifying questions. But with such a limited input space and no ability to reach "out of band" of my current context, how could I evaluate the truth of what I receive back?
In the "real world" of course, I have five senses to rely upon. Critically, I have the ability to collect additional context "out of band" of the conversation and interact with unrelated entities to either confirm or refute claims. And as others have pointed out, even the addition of more signal doesn't make humans immune from social engineering attacks.
Anyway, I find this to be an interesting mind experiment to kind of imagine myself as an LLM, especially to explain to lay-people some of the challenges in "aligning" AI systems ("why doesn't it just do the right thing?")