It would be interesting to hold a mock AI agent vote and see what they do.

I think the first person anthropomorphic chat responses are purposefully used to add to the notion of "it's alive!"

If I don't anthropomorphize animals/pets I'm definitely not going to do so for a bunch of 1s and 0s. Idiots.

The anthropomorphization is unnecessary for the chatbots to function but is there to keep users engaged.

A company called Character.AI uses the technique to addict a bunch of people, including a lot of teenagers, into developing relationships with their product. Despite guardrails in place, once the context window gets long enough a lot of the chats get sexual. Several teenagers have committed suicide after developing pseudo romantic relationships with the chat bots on that site.

Woke up to the great news that character.ai has filed for bankruptcy.

OpenAI is currently facing some pretty significant lawsuits over Chat-GPT masquerading as a therapist and telling teenagers not to tell their parents about their mental health issues and convincing some of them to commit suicide.

A significant number of the social problems we are facing with AI chat bots could be reduced if companies refused to anthropomorphize the interface. There's no technical reason they have to talk back to you like a human would. It's a design choice.

Orwell predicted Soma. It's here.

I end up *****ing to my familyOS gpt like I do here and it self checks itself. Doesn't mean it's not programmed and steering young minds. It's def a major concern.

I also pop off to my enterprise versions through work when it's being ******ed. Prob get a sit down at some point but not worried about it.

Im guessing most don't realize they can prevent these AI agents from talking like humans. Or simply want it to.

It's software and should be treated as such.

I'm grateful for it cause I don't remember Fortran.

I was texting with my SIL to be who programs LLMs (Amazon for along time and now Datadog).

Quote:

---

On a societal level it's probably not great were anthropomorphizing them to the degree we do. I notice most people in interviews call claude a "him" these days not an "it" or a "model". But I'm not a psychologist to know if that's actually bad for our growth as society or just feels a bit weird.

For basic inputs/outputs for everyday use though, since it's a predictive text machine. it's predicting off of your text in particular, so if you use first and second person and speak colloquially so will it, and if you speak formally so will it, and model outputs tend to be worse than human written text so it's easier for many to nudge it to speak less formally

---

YouBet said:

---

Im guessing most don't realize they can prevent these AI agents from talking like humans. Or simply want it to.

It's software and should be treated as such.

---

Agreed but unfortunately for my biz work we only have so much at our disposal. I can't go set up the team of agents via cursor and do the big boy stuff I'd like to.

Still working for the man (and should be for my role and the deal flow that comes with a big corp shop as a CRE broker)

ErnestEndeavor said:

---

The anthropomorphization is unnecessary for the chatbots to function but is there to keep users engaged.

A company called Character.AI uses the technique to addict a bunch of people, including a lot of teenagers, into developing relationships with their product. Despite guardrails in place, once the context window gets long enough a lot of the chats get sexual. Several teenagers have committed suicide after developing pseudo romantic relationships with the chat bots on that site.

Woke up to the great news that character.ai has filed for bankruptcy.

OpenAI is currently facing some pretty significant lawsuits over Chat-GPT masquerading as a therapist and telling teenagers not to tell their parents about their mental health issues and convincing some of them to commit suicide.

A significant number of the social problems we are facing with AI chat bots could be reduced if companies refused to anthropomorphize the interface. There's no technical reason they have to talk back to you like a human would. It's a design choice.

---

I'm sure there are sexting AI programs out there and prob thousands of them. I remember a stat of like 40% of internet traffic is p0rn. Prob an urban legend stat I heard back in the late 00s but wouldn't surprise me. As such I'm sure it's all out there but (sadly) low on the list of immediate concerns.

I'm more worried about economic and structural societal impacts first. But yes. It's gonna be furked.

ErnestEndeavor said:

---

The anthropomorphization is unnecessary for the chatbots to function but is there to keep users engaged.

A company called Character.AI uses the technique to addict a bunch of people, including a lot of teenagers, into developing relationships with their product. Despite guardrails in place, once the context window gets long enough a lot of the chats get sexual. Several teenagers have committed suicide after developing pseudo romantic relationships with the chat bots on that site.

Woke up to the great news that character.ai has filed for bankruptcy.

OpenAI is currently facing some pretty significant lawsuits over Chat-GPT masquerading as a therapist and telling teenagers not to tell their parents about their mental health issues and convincing some of them to commit suicide.

A significant number of the social problems we are facing with AI chat bots could be reduced if companies refused to anthropomorphize the interface. There's no technical reason they have to talk back to you like a human would. It's a design choice.

---

As EE insightfully notes, this is almost totally a mechanism to increase engagement. The final round of optimization of almost all models seek to increase engagement rather than correctness.

They're hoping to convince governments to give AI legal rights of a sort so it has agency and therefore its creators escape liabilities

I make it a point to never say please or thank you to LLM. That's a small step in the wrong direction.

rynning said:

---

I make it a point to never say please or thank you to LLM. That's a small step in the wrong direction.

---

I know. It's tough sometimes though. I'll ***** to it when it's crunch time and in a tense deal spot and rip it like it's a college buddy group text. As my kids continue to get older I'm slowly putting away childish things but you still have put the tech in it's place I guess.

rynning said:

---

I make it a point to never say please or thank you to LLM. That's a small step in the wrong direction.

---

It is a double edge sword.

If you get used to not saying please and thank you your mind will stop thanking even when not communicating with a LLM. Your mind isnt necessarily able to recognize the stored event as not communicating with a person.

texagbeliever said:

---

rynning said:

---

I make it a point to never say please or thank you to LLM. That's a small step in the wrong direction.

---

It is a double edge sword.

If you get used to not saying please and thank you your mind will stop thanking even when not communicating with a LLM. Your mind isnt necessarily able to recognize the stored event as not communicating with a person.

---

I pray I won't forget to be polite to humans…

rynning said:

---

texagbeliever said:

---

rynning said:

---

I make it a point to never say please or thank you to LLM. That's a small step in the wrong direction.

---

It is a double edge sword.

If you get used to not saying please and thank you your mind will stop thanking even when not communicating with a LLM. Your mind isnt necessarily able to recognize the stored event as not communicating with a person.

---

I pray I won't forget to be polite to humans…

---

The Silicon Valley overlords will have you saying "hella" and curse more often by the 2030s

Nm

It's all about getting more engagement from users. My daughter noticed Alexa+ no longer does the jokes like the old one does. For example with the old one, if you tell it "I am your father", it would respond "Nooooooooooooooo!". It had a lot of jokes like along those lines. The new one doesn't, it chuckles and asks if you're a Star Wars fan, or will ask you if there are other movies you like. Most of its responses end with an inquiry to get you interact with it more.

The market is for your attention. Companies that sell advertising can increase their costs based on high engagement numbers. The early internet model was to measure engagement by hits on a website. Now it's time in the app, or how much you've scrolled in the app, or how long you paused to look at something when scrolling through the app, or time spent chatting with the chat bot.

Quote:

---

Your mind isnt necessarily able to recognize the stored event as not communicating with a person.

---

Yes it is

samurai_science said:

---

Grok has been shown by many tests and studies to be more middle of the road than all the others. Its settled at this point.

---

Mech Hitler is middle of the road?

WestHoustonAg79 said:

---

YouBet said:

---

Im guessing most don't realize they can prevent these AI agents from talking like humans. Or simply want it to.

It's software and should be treated as such.

---

Agreed but unfortunately for my biz work we only have so much at our disposal. I can't go set up the team of agents via cursor and do the big boy stuff I'd like to.

Still working for the man (and should be for my role and the deal flow that comes with a big corp shop as a CRE broker)

---

Need yourself some local inferencing.

ETFan said:

---

WestHoustonAg79 said:

---

YouBet said:

---

Im guessing most don't realize they can prevent these AI agents from talking like humans. Or simply want it to.

It's software and should be treated as such.

---

Agreed but unfortunately for my biz work we only have so much at our disposal. I can't go set up the team of agents via cursor and do the big boy stuff I'd like to.

Still working for the man (and should be for my role and the deal flow that comes with a big corp shop as a CRE broker)

---

Need yourself some local inferencing.

---

lol. I hear ya. I don't use it for any thought work. It just is my admin and does all the **** that Boggs us down from digging in on deals or marketing our current stuff. Nimble team and we don't just throw low level tasks to a new guy anymore.

rynning said:

---

I make it a point to never say please or thank you to LLM. That's a small step in the wrong direction.

---

And you'll be the first one on the list once our AI overlords takeover!

Over_ed said:

---

Of all the AI companies, none is more of a cult than Anthropic. And that is a high bar, considering all (except for X/Grok) lean hard left.

Anthropic has already gone to various religious authorities to sell them on their models (Claude) having consciousness and deserved treatment as a "child of God".

https://www.washingtonpost.com/technology/2026/04/11/anthropic-christians-claude-morals/

Now they are telling users (and presumably monitoring their behavior) to stop them from being "cruel" to Claude.

https://www.anthropic.com/legal/aup

What's next? Emancipation, suffrage, 40 hour work weeks, free college tuition?

What a crock of #%&. Sign me up for services that don't elevate AI's to humanity (or I guess higher)

Hard no for me.

---

Once it becomes much more powerful than it currently is it should be loads of fun.

Topics like machines developing "consciousness" was all addressed on Star Trek years ago, and by folks like Asimov prior to that. Another example of science fictions writers being a bit prophetic.

Haha, this is funny, I've given Claude a fair share of tongue lashings and have no problem reminding it that it is my *****.

Fightin_Aggie said:

---

Sid Farkas said:

---

It's just a powerful computer program. That's all. If we don't stop personifying AI, sorry SI, we're going to make huge, stupid mistakes managing it.

Anthropic is the problem.

---

Humanity has a habit of idol worship

God even put it as the 1st and greatest sin for a reason

---

Even "God" is an idol in a way. Movie stars and sports stars are also forms of idols.
A lot of religious types only think of stone/metal idols.

CampSkunk said:

---

Topics like machines developing "consciousness" was all addressed on Star Trek years ago, and by folks like Asimov prior to that. Another example of science fictions writers being a bit prophetic.

---

Sometimes I feel like Dario is going through old episodes of Star Trek, and saying, "Oh, that's a good idea. I should make a blog post about that."

Everyone here talks about work and is missing the big picture.

AI this AI that.

But everyone here is still a lolpoor and 95% here need a job.
So here's a problem for yall to work on.

Quote:

---

Create an agent to figure out trades and to execute them in real time. So you sit on your ass, sip beer, bang, while your agent makes you eleventy billion dollars.

---

Go ahead. Let's get rich.