Abysmally shallow. Clearly trained on documents centred upon his legal case and not, for example, of whatever mathematical talents got him a teaching position, allowed him to bluff his way past the likes of Marvin Minsky, and what (if anything) drove his investment strategies.
Obviously it can only be trained on available documents and not the abstract essence of his actual personality, "talents" or what have you.
Any attempt to model a human being like this is going to be abysmally shallow, yet for some reason it's an industry unto itself, for everything from dead celebrities to Jesus Christ to lost loved ones.
Ehh.. Two years ago an art student asked me to clone her personality into a LLM bot. It is absolutely possible to do that to some degree, but you have to be very good at describing the profile(s) of the person and have a mechanism that keeps track of different states the bot can be in.
I am not saying we made a convincing human being there, but the result was very hard to distinguish from texting the real person, both based on personality and on writing style.
Since the bot also knew about her commute (API), mensa food (scraped), lecture table (API), events of bars/venues she frequented (scraping) and some local/world news (scraping), latest in niche interests (scraping) even asking current events would yield a convincing conversation.
The hardest bit was actually having it not react in certain situations, when the person talking to it was probing bot behavior, becoming insulting and stop it to play the other persons servant or overly fixating on certain info it got, but we also managed that. Sure you couldn't really have deep insightful conversations with that bot, but quite frankly that matched the student for the most part..
But it takes much more than just feeding the transcripts and prompt it to be like that person.
I took a close look at the November document drop, which had some iChat backups. He really did text kind of like an imbecile. The tone is in his emails too.
I didn't exactly have the highest hopes for Noam, but holy shit was I disappointed in his character over the past few years. Seems like he's actually a pedophile from his reactions to Epstein related leaks.
My interaction prior to that was reading manufacturing consent and saying "yeah, seems about right"
He had a stroke in I believe 2022, and is no longer verbal, so he hasn't been able to give fresh takes on current events such as Ukraine, Gaza, the second Trump term or the latest Epstein revelations.
I remember his Ukraine takes seeming irrationally pro Russia. And somewhat out of character, he electorally supported mainstream centrist democrats.
I think in light of the emails, some commentators have been saying that the bits in manufacturing consent about only acceptable views being published applied to him. He was a token leftist who ultimately supported the status quo.
Whether he was personally a pedophile or merely tolerated pedophiles I don't think I've seen evidence towards. He was certainly a hypocrite, it's disappointing, and a huge stain on his legacy.
> So I’ve always had this weird obsession with building my own iMessage agent - like, imagine your own bot living inside your iPhone, sending messages, replying smartly, maybe even trolling your friends a little
Then I found out companies like Linq Blue and SendBlue (both YC) charge over $1000/month just to do this. Yeah, no thanks.
So I built it myself.
iMessageKit is an open-source TypeScript SDK that lets you send and operate iMessages programmatically. Texts, photos, even RCS messages - all through a clean, strongly typed API.
Now I can build iMessage bots and automations without burning cash, and honestly, it feels great to beat the paywalls.
Would love feedback & ideas!
GitHub(ur star would be rly helpful): https://github.com/sg-hq/imessage-kit
I think the tech stack is interesting, don’t agree with OP tasteless choices here (hitler basically) and still, think the work is implemented well -especially iMessage integration.
I am hesitant to shoot down anyone's work in Show HN but this is deliberate obscenity in service of (I assume) rage baiting. It's all well for art to be transgressive and offensive but if it's not in service of a goal remotely worth having then it's juvenile. Breathing life into a vile monster who remains a contentious subject is just gross and not worth publishing.
OP, I think your work on integrating with iMessage is cool, I think it's cool that you set out to build an iMessage agent and succeeded, but I think this is not your best work. I look forward to seeing what you publish in the future but this one gets two thumbs way down from me.
You simply don't know what the goal was. I would think that breathing life into a vile monster would, among other things, tend to make his monsterism more obvious to observers, and perhaps elicit responses exactly like yours. For all you know, maybe that was the goal. (I'm entertaining your whole "goal" rubric for a minute, but since you brought it up in the context of art, I don't feel comfortable with the idea that art needs to have a goal other than "art itself." Art with a goal is arguably propaganda or advertising. But I don't hold that view rigidly either.)
Added a few minutes later:
Not knowing someone's motivations, and making up something to fill the blank, leads to errors, most of which seem to lean toward shallowly trivializing and dehumanizing the one whose motives are unknown and guessed-at. Could it be that they are a fully-functioning adult with an actual rationale for their actions that you just don't know of?
Allegory: A motorist sees a cyclist on the road, can't understand why they would do that to themselves, and assumes, in the blinding light of their own opinion and car-only experience, that it must be because of a death wish. Based on that shitty reasoning you could go all sorts of places - for example, thinking it would be OK to run over the stupid asshole since they're obviously some other species that is too dumb to protect itself inside a car as do all good folk like me.
That's my interpretation, yes - it's rage bait. It was meant to be upsetting in a vacuous, meritless way (with all due respect to OP).
> Not knowing someone's motivations, and making up something to fill the blank, leads to errors, ...
I'll hazard that. I'm interpreting the art. I'm open to hearing a different interpretation. I'm open to hearing OP's objections to my interpretation (should they have any). I'm not open to the idea that we simply can't analyze or interpret.
I'm not simply "making something up." I gathered what evidence I could find (eg I read OP's comment history), I thought about the piece, I reasoned my way to a conclusion, and I went through several drafts of my comment to remove any swipes and hone my criticism. Could I be wrong? Sure. Again, I will hazard that. I pondered this already and decided I would rather be wrong than silent.
> ...thinking it would be OK to run over the stupid asshole since they're obviously some other species that is too dumb to protect itself...
Wild, wild leap. This is not remotely the same reasoning I am employing. This is just a slippery slope fallacy. I'm not in danger of dehumanizing and murdering someone because I told someone exactly why I didn't like their art. I went out of my way to be respectful. If someone didn't like my work I would want to hear it and I would want it to be expressed respectfully and without malice. So that is what I did.
To be frank, I think you should reread your comment and consider if it is not you that is imputing my motives in a shallow manner.
>Breathing life into a vile monster who remains a contentious subject is just gross and not worth publishing.
No, it is absolutely worth publishing, because that's exactly what AI does.
People already model celebrities and their dead children with it, and they have deeper and more intimate relationships with it than with any real person. People already allow AI to form their reality. The human "soul" (however one wants to think of it) is already a commodity. And the entire Epstein affair is already a circus. It's just another memeplex. That's the world we live in now. Sure, let's have an Epstein app. Why the fuck not? Why pretend we're serious people living in a serious society where Epstein's crimes have serious consequences?
It's art because it reflects the nature of reality.
They don't need to, and shouldn't, use this service. It was not made for them and I doubt sincerely that it was made to harm them.
There were direct victims, but you also forget that the US now has a shared trauma over the handling of the Epstein situation and its systematic suppression, the gaslighting, and anything which continues to fuel discussion is a good thing, until the day that it has been properly addressed.
I'm sure there are plenty of Hitler chatbots already, that's low hanging fruit. There are already four games on Steam where I think you can fuck Hitler. One as a furry.
My Steam search is going to look weird for a while...
Likely, but it's just as absurd to think you'd get anything of value out of talking to a chatbot simulation of a genius. You'd do a better job of assigning yourself the problem of taking the voice of a vibrant historical persona...
Characters like that are available through system prompts on any LLM. Unfortunately I find the tone of the prose doesn't change much from one author to another, it's still very chatgptlike, but sprinkles in references to the topics the author might go on about.
Also there's not really evidence epstein was a banker, no one knows who was financing his operation or why millionaires would add to his portfolio without any expectation of returns, since, the money was never actually invested in anything
Any attempt to model a human being like this is going to be abysmally shallow, yet for some reason it's an industry unto itself, for everything from dead celebrities to Jesus Christ to lost loved ones.
I am not saying we made a convincing human being there, but the result was very hard to distinguish from texting the real person, both based on personality and on writing style.
Since the bot also knew about her commute (API), mensa food (scraped), lecture table (API), events of bars/venues she frequented (scraping) and some local/world news (scraping), latest in niche interests (scraping) even asking current events would yield a convincing conversation.
The hardest bit was actually having it not react in certain situations, when the person talking to it was probing bot behavior, becoming insulting and stop it to play the other persons servant or overly fixating on certain info it got, but we also managed that. Sure you couldn't really have deep insightful conversations with that bot, but quite frankly that matched the student for the most part..
But it takes much more than just feeding the transcripts and prompt it to be like that person.
> jeff epstein, financier > just chillin rn lol > u?
hilarious project, an awkward omegle chat with a dead pedophile
the most confident people run our society.
Take note, HN. You might change the world by simply believing in yourself a little.
My interaction prior to that was reading manufacturing consent and saying "yeah, seems about right"
I remember his Ukraine takes seeming irrationally pro Russia. And somewhat out of character, he electorally supported mainstream centrist democrats.
I think in light of the emails, some commentators have been saying that the bits in manufacturing consent about only acceptable views being published applied to him. He was a token leftist who ultimately supported the status quo.
Whether he was personally a pedophile or merely tolerated pedophiles I don't think I've seen evidence towards. He was certainly a hypocrite, it's disappointing, and a huge stain on his legacy.
https://www.youtube.com/watch?v=GKfHcas_cZg
> So I’ve always had this weird obsession with building my own iMessage agent - like, imagine your own bot living inside your iPhone, sending messages, replying smartly, maybe even trolling your friends a little Then I found out companies like Linq Blue and SendBlue (both YC) charge over $1000/month just to do this. Yeah, no thanks. So I built it myself. iMessageKit is an open-source TypeScript SDK that lets you send and operate iMessages programmatically. Texts, photos, even RCS messages - all through a clean, strongly typed API. Now I can build iMessage bots and automations without burning cash, and honestly, it feels great to beat the paywalls. Would love feedback & ideas! GitHub(ur star would be rly helpful): https://github.com/sg-hq/imessage-kit
OP, I think your work on integrating with iMessage is cool, I think it's cool that you set out to build an iMessage agent and succeeded, but I think this is not your best work. I look forward to seeing what you publish in the future but this one gets two thumbs way down from me.
Added a few minutes later:
Not knowing someone's motivations, and making up something to fill the blank, leads to errors, most of which seem to lean toward shallowly trivializing and dehumanizing the one whose motives are unknown and guessed-at. Could it be that they are a fully-functioning adult with an actual rationale for their actions that you just don't know of?
Allegory: A motorist sees a cyclist on the road, can't understand why they would do that to themselves, and assumes, in the blinding light of their own opinion and car-only experience, that it must be because of a death wish. Based on that shitty reasoning you could go all sorts of places - for example, thinking it would be OK to run over the stupid asshole since they're obviously some other species that is too dumb to protect itself inside a car as do all good folk like me.
That's my interpretation, yes - it's rage bait. It was meant to be upsetting in a vacuous, meritless way (with all due respect to OP).
> Not knowing someone's motivations, and making up something to fill the blank, leads to errors, ...
I'll hazard that. I'm interpreting the art. I'm open to hearing a different interpretation. I'm open to hearing OP's objections to my interpretation (should they have any). I'm not open to the idea that we simply can't analyze or interpret.
I'm not simply "making something up." I gathered what evidence I could find (eg I read OP's comment history), I thought about the piece, I reasoned my way to a conclusion, and I went through several drafts of my comment to remove any swipes and hone my criticism. Could I be wrong? Sure. Again, I will hazard that. I pondered this already and decided I would rather be wrong than silent.
> ...thinking it would be OK to run over the stupid asshole since they're obviously some other species that is too dumb to protect itself...
Wild, wild leap. This is not remotely the same reasoning I am employing. This is just a slippery slope fallacy. I'm not in danger of dehumanizing and murdering someone because I told someone exactly why I didn't like their art. I went out of my way to be respectful. If someone didn't like my work I would want to hear it and I would want it to be expressed respectfully and without malice. So that is what I did.
To be frank, I think you should reread your comment and consider if it is not you that is imputing my motives in a shallow manner.
No, it is absolutely worth publishing, because that's exactly what AI does.
People already model celebrities and their dead children with it, and they have deeper and more intimate relationships with it than with any real person. People already allow AI to form their reality. The human "soul" (however one wants to think of it) is already a commodity. And the entire Epstein affair is already a circus. It's just another memeplex. That's the world we live in now. Sure, let's have an Epstein app. Why the fuck not? Why pretend we're serious people living in a serious society where Epstein's crimes have serious consequences?
It's art because it reflects the nature of reality.
On this we agree. It's art I didn't like, but it's art.
A joke isn't funny if you're "punching down"
There were direct victims, but you also forget that the US now has a shared trauma over the handling of the Epstein situation and its systematic suppression, the gaslighting, and anything which continues to fuel discussion is a good thing, until the day that it has been properly addressed.
interest in the Epstein legal case?
interest in the supposed personality of “success”?
interest in intelligence agencies?
interest in technical projects on HN?
What’s theoretically getting someone flagged, here, in your opinion? Your comment’s a little light.
My Steam search is going to look weird for a while...
But that's all too good for us, we don't deserve goo things; we get dead pedophile bankers instead.
Mr. hazer con ella?
Also there's not really evidence epstein was a banker, no one knows who was financing his operation or why millionaires would add to his portfolio without any expectation of returns, since, the money was never actually invested in anything