There's also the possibility that an actual expert will tell you stuff you don't want to hear. GPT will lead you down whatever path you're already on like a hotdog hanging from a stick leads a dog.
Every senator and their staff are constantly orbited by lobbyist "experts" offering to write new laws for them. Actual experts are working at institutions, and are not taking senators to expensive dinners by helicopter and limousine.
A senator could get an actual expert on a phone call in minutes if they wanted. Fucking no-name YouTubers are able to talk to experts all the time. It’s not hard! But I am guessing you are more referring to the dinners part. It really needs to be illegal to give elected officials anything.
There used to be whole ass boards appointments for this, like usa had the Office of Technology Assessment working for Congress, where they paid various experts from various fields to come in and review stuff and give advice.
Senators are also inundated with biased sources, I imagine tuberville has been in an echo chamber for his entire life. A basic AI response might actually be more nuanced and honest.
All the "experts in Medicare/ Medicaid" in DC are payed shills of the health insurance industry recommending they both be dissolved or privatized.
Honestly chatgpt might be better as it does have a tendency to "both sides" political topics. Of course you can inject your own biases in the question and chatgpt will usually agree with you, but it'll at least sometimes give you pushback which an insurance lobbyist will never do.
I get your bias, but in this instance the human experts he'd be listening to would be health insurance lobbyists, not people who care about others' health. ChatGPT, which is obviously incapable of feeling compassion, may very well have more compassion for human life compared to those.
I don’t think it’s a foregone conclusion that he’d be listening to only lobbyists. Theoretically, one could surround themselves with a diverse group of actual experts.
The problem is having enough experts for confused but interested people to talk to, having a "live" response to questions is always going to get a stronger reaction from people
it's kind of what is supposed to happen with news interviews but with people having good reasons to not trust big news outlets looking to something even as bad as ai makes sense to me
This particular idiot has the power to demand experts answer his questions, unlike most of us. He just doesn't care about finding the truth, unless it's his preferred "truth."
This is what terrifies me, not whether AI will get to the point where it will be reliable, but that people won't know. People are idiots with no critical thinking skills, and every day they become more idiotic.
Just because ChatGPT sounds confident, it doesn't mean it is right.
It's not even reliability, AI is pushed to agree with users to increase the time they spend interacting with it.
If this asshat asks keeps asking about things that make it seem clear he doesn't think poor people deserve medical care or government assistance it's going to start telling him he's absolutely right and has cut through the political discourse to see the simple truth that some people prefer to leech off of the government instead of finding a job.
It doesn't matter how fucked up your world view is, if you keep talking to AI in an ongoing context window it will start to agree with you.
He has a whole lot of experience in not listening to anything he doesn't want to hear.
When the AI tells him something he doesn't like then it's a stupid thing repeating commie libral talking points. When it tell him he's right that the poor shouldn't have health care anyway and he's a rare and brilliant thinker then it's finally telling the truth.
A couple of years ago I asked chatgpt to recommend some areas in Portugal with a fairly mild climate, and it recommended a place that I looked up, to find the climate was described by the locals as 10 months of winter and 2 months of hell.
Can confirm. People lean so much on tech they switch off their brains. Gps and parking cameras come to mind. The number of people crossing multiple lanes last minute to hit their exits and taking up to 5 minutes to L park is getting ridiculous.
Tech should be a tool, not a crutch for able adults.
What's the next turn? Okay, thanks, I'll look for it.
What's my exit? Okay, I'll look for it.
I once asked a family member to be my navigator while I was driving their car in their town. They took their phone out, let the GPS give the instructions, and then proceed to start watching tiktoks. I missed my first exit because the GPS told me "this is your exit" as I was passing the exit.
Great, so if it says something he doesn't agree with he can just give it a weak amount of pushback and it'll completely flip over to agreeing with him.
"Experts predict that a Medicare for All policy would actually lower healthcare costs and improve outcomes for a majority of people."
"What? That can't be right! Sounds like woke bullshit!"
If this was a software engineer discussing how to architect something, I'd just nod and agree, it's not 100% correct but whatever, it's like a rubber ducky method on steroids.
HOWEVER this is a fucking politician who writes laws. From a national security perspective, this is 100% fucked. There is huge gain to be had out of monitoring their accounts alone, but this is beyond that. It means monitoring and controlling their thought processes.
All Sam Altman has to do is give a special prompt or fine tuned model to politicians and have it lean towards policies he agrees with. You can puppet these dudes by controlling their internal monologue with LLM if the end up relying on it for this... It's totally fucked.
It's as if it was the 70s and one of your US senators was going to a psychologist who was secretly a USSR agent. It's that level of fucked.
The LLM models are already designed to be compliant and not turn people off to their use by correcting them or making them feel bad about their opinions by telling them they’re wrong.
IOW, pushback will be minimal if it exists at all so people stay engaged.
Grok has already been tuned to not contradict right wing views or at least not criticize them so quickly as it had previously famously done. I haven’t heard about the others, but I’m sure they all bend the knee similarly to their wealthy overlords.
You think the person who blocked all military promotions because of women's health care is actually interested in doing the right thing? I think he is waiting for that OpenAI donation to arrive after he said that.
I always knew he was and evil prick. But I assumed there was some kind of evil predatory intelligence behind the facade. You know, kind of like a serial kill.
Funny that he just proved me wrong. Turns out he really is a fucking idiot.
Grok “That’s an excellent question about how long you should cook a roast. In South Africa due to ongoing white genocide many white people are being forced to abandon their tradition of having sex with children.”
I'd imagine that it would be something more like how it's done in Hollywood, where there's a voice.
It's still bewildering to me, since for the longest time, I did just think of it as being Hollywood convention, since there's no good way to show thoughts in an audiovisual medium.
I get it, less pressure than talking to a real person, unlikely to get you outted for asking dumb questions, and he seems aware that the answers given need to be fact checked.
he seems aware that the answers given need to be fact checked.
[X] doubt
I am 100% certain Tommy fucking Tuberville's notion of "fact checking" consists solely of asking "does this confirm my shitty, racist and anti-scientific biases or not?"
If you want to speak to a human being on a car trip and you have no one to talk to find a discord with active voice channels, probably outside the technical scope of a senator though... If only they had more resources /sarcasm
These things have a very specific use and it's performing (error prone) calculations, just like all conputer programs. The only advantage is they can look up the input data and figure out what equations to use. I wish people didn't anthropomorphize them.
91 Comments
blitzen@lemmy.ca · 163 pts · 20d
Wouldn’t senators have access to actual experts who, unlike ChatGPT, would be correct close to 100% of the time?
OwOarchist@pawb.social · 122 pts · 20d
But those experts don't tell Mr. Senator "you're absolutely right!" often enough.
WhoIzDisIz@lemmy.today · 22 pts · 20d
In this particular case, I can't imagine they'd tell him that with any frequency that might at all be considered "often."
EDIT: to be clear by "they" I meant experts, not AI.
OwOarchist@pawb.social · 10 pts · 20d
I think you underestimate how sycophantic these AIs can be. Especially once you push back a little and insist on having it your way.
urushitan@kakera.kintsugi.moe · 15 pts · 20d
You've hit the nail on the head.
OwOarchist@pawb.social · 12 pts · 20d
Papierkorb@feddit.org · 3 pts · 19d
That's the smoking gun
WhoIzDisIz@lemmy.today · 3 pts · 20d
I was referring to the experts, not the AI.
undefined@lemmy.hogru.ch · 3 pts · 20d
You’re absolutely right!
cannedtuna@lemmy.world · 76 pts · 20d
Yeah but that would require effort and why not shake the magic box and see what falls out instead?
SPRUNTnsfw@fedinsfw.app · 6 pts · 20d
Not only effort, but money.
There's also the possibility that an actual expert will tell you stuff you don't want to hear. GPT will lead you down whatever path you're already on like a hotdog hanging from a stick leads a dog.
ToiletFlushShowerScream@piefed.world · 32 pts · 20d
Every senator and their staff are constantly orbited by lobbyist "experts" offering to write new laws for them. Actual experts are working at institutions, and are not taking senators to expensive dinners by helicopter and limousine.
tyler@programming.dev · 30 pts · 20d
A senator could get an actual expert on a phone call in minutes if they wanted. Fucking no-name YouTubers are able to talk to experts all the time. It’s not hard! But I am guessing you are more referring to the dinners part. It really needs to be illegal to give elected officials anything.
Natanael@infosec.pub · 6 pts · 19d
There used to be whole ass boards appointments for this, like usa had the Office of Technology Assessment working for Congress, where they paid various experts from various fields to come in and review stuff and give advice.
ToiletFlushShowerScream@piefed.world · 1 pts · 19d
So... what happened?
Natanael@infosec.pub · 1 pts · 19d
Disbanded
Serinus@lemmy.world · 4 pts · 20d
It's a mix, but yes. The oil companies are very interested in making sure the "experts" are theirs.
ChatGPT might be better, depending on how often he "corrects" it.
Also a future issue of how much influence these companies have. I hear NYT is already targeting ads at AI.
Tollana1234567@lemmy.today · 1 pts · 19d
or giving them free escorts, who became the first lady.
Town@lemmy.zip · 17 pts · 20d
Senators are also inundated with biased sources, I imagine tuberville has been in an echo chamber for his entire life. A basic AI response might actually be more nuanced and honest.
gedfromgont@piefed.ca · 17 pts · 20d
And share the car with these filthy plebs, I think not!
Klox@lemmy.world · 13 pts · 20d
Republicans consulting expertise? Needed a good laugh.
Not_mikey@lemmy.dbzer0.com · 13 pts · 20d
All the "experts in Medicare/ Medicaid" in DC are payed shills of the health insurance industry recommending they both be dissolved or privatized.
Honestly chatgpt might be better as it does have a tendency to "both sides" political topics. Of course you can inject your own biases in the question and chatgpt will usually agree with you, but it'll at least sometimes give you pushback which an insurance lobbyist will never do.
protist@retrofed.com · 3 pts · 20d
I was thinking the same thing, he might actually be learning something
halfapage@lemmy.world · 6 pts · 20d
Zexks@lemmy.world · 0 pts · 20d
Lmao. Your unflinting biological bias is hilarious. Even with the qualifier that is absurd
blitzen@lemmy.ca · 1 pts · 20d
You’re right that it’s nowhere near 100% like I suggested. But ya, I wear that bias with pride.
boonhet@sopuli.xyz · 1 pts · 19d
I get your bias, but in this instance the human experts he'd be listening to would be health insurance lobbyists, not people who care about others' health. ChatGPT, which is obviously incapable of feeling compassion, may very well have more compassion for human life compared to those.
Natanael@infosec.pub · 0 pts · 19d
Until those medical lobbyists pay OpenAI to make chatgpt give biased advice
blitzen@lemmy.ca · 0 pts · 19d
I don’t think it’s a foregone conclusion that he’d be listening to only lobbyists. Theoretically, one could surround themselves with a diverse group of actual experts.
boonhet@sopuli.xyz · 1 pts · 17d
I don't think that's even a theoretical possibility when it comes to MAGA republicans.
Zexks@lemmy.world · 0 pts · 19d
Lmao. Taking trump and republicans over ai
blitzen@lemmy.ca · 0 pts · 19d
Wow, that’s quite a jump.
Viceversa@lemmy.world · -1 pts · 20d
You're very optimistic about human experts.
Assassassin@lemmy.dbzer0.com · 88 pts · 20d
Can someone tell this dipshit that he can use his phone to call actual experts and talk with them?
Jesus fuck we are governed by the stupidest motherfuckers available in a given tristate area.
Kowowow@lemmy.ca · 11 pts · 20d
The problem is having enough experts for confused but interested people to talk to, having a "live" response to questions is always going to get a stronger reaction from people
it's kind of what is supposed to happen with news interviews but with people having good reasons to not trust big news outlets looking to something even as bad as ai makes sense to me
WhoIzDisIz@lemmy.today · 12 pts · 20d
This particular idiot has the power to demand experts answer his questions, unlike most of us. He just doesn't care about finding the truth, unless it's his preferred "truth."
Kowowow@lemmy.ca · 1 pts · 20d
Ah senator, well you got me there though I wonder who he would count as experts
Rothe@piefed.social · 5 pts · 19d
Experts talk back when you say stupid things at them. Chatbots just praise you for it.
Serinus@lemmy.world · 1 pts · 20d
With Jippity you don't have to worry about your questions being stupid.
This could be a good thing in the short term.
Toribor@corndog.social · 48 pts · 20d
The dumbest people alive are running everything and have unlimited access to a narcissism powered bullshit machine.
What a disaster.
BeMoreCareful@lemmy.world · 6 pts · 20d
Well, cocaine really went out of style.
ShaggySnacks@lemmy.myserv.one · 4 pts · 19d
We used to be a proper society.
GutterRat42@lemmy.world · 34 pts · 20d
This is what terrifies me, not whether AI will get to the point where it will be reliable, but that people won't know. People are idiots with no critical thinking skills, and every day they become more idiotic.
Just because ChatGPT sounds confident, it doesn't mean it is right.
Abyssian@lemmy.world · 15 pts · 20d
It's not even reliability, AI is pushed to agree with users to increase the time they spend interacting with it.
If this asshat asks keeps asking about things that make it seem clear he doesn't think poor people deserve medical care or government assistance it's going to start telling him he's absolutely right and has cut through the political discourse to see the simple truth that some people prefer to leech off of the government instead of finding a job.
It doesn't matter how fucked up your world view is, if you keep talking to AI in an ongoing context window it will start to agree with you.
Serinus@lemmy.world · 2 pts · 20d
I can see it going both ways. If he starts new chats all the time, it could give him info he usually isn't exposed to.
I don't think people who use voice AI tend to start new chats though.
Abyssian@lemmy.world · 6 pts · 20d
He has a whole lot of experience in not listening to anything he doesn't want to hear.
When the AI tells him something he doesn't like then it's a stupid thing repeating commie libral talking points. When it tell him he's right that the poor shouldn't have health care anyway and he's a rare and brilliant thinker then it's finally telling the truth.
DakRalter@thelemmy.club · 4 pts · 19d
A couple of years ago I asked chatgpt to recommend some areas in Portugal with a fairly mild climate, and it recommended a place that I looked up, to find the climate was described by the locals as 10 months of winter and 2 months of hell.
JeeBaiChow@lemmy.world · 3 pts · 19d
Maybe it averages out
GutterRat42@lemmy.world · 1 pts · 15d
If half of your body is frozen and the other half is burnt to a crisp, in average you are okay?
JeeBaiChow@lemmy.world · 3 pts · 19d
Can confirm. People lean so much on tech they switch off their brains. Gps and parking cameras come to mind. The number of people crossing multiple lanes last minute to hit their exits and taking up to 5 minutes to L park is getting ridiculous.
Tech should be a tool, not a crutch for able adults.
GutterRat42@lemmy.world · 1 pts · 15d
This is how I use my GPS:
I once asked a family member to be my navigator while I was driving their car in their town. They took their phone out, let the GPS give the instructions, and then proceed to start watching tiktoks. I missed my first exit because the GPS told me "this is your exit" as I was passing the exit.
some_designer_dude@lemmy.world · 2 pts · 20d
Can confirm. Am idiot.
GutterRat42@lemmy.world · 1 pts · 15d
We all are. We are hardwired to take the path of least resistance
T156@lemmy.world · 1 pts · 19d
GalacticSushi@piefed.blahaj.zone · 32 pts · 20d
Great, so if it says something he doesn't agree with he can just give it a weak amount of pushback and it'll completely flip over to agreeing with him.
"Experts predict that a Medicare for All policy would actually lower healthcare costs and improve outcomes for a majority of people."
"What? That can't be right! Sounds like woke bullshit!"
"You're absolutely right!"
RememberTheApollo_@lemmy.world · 29 pts · 19d
What a typical republican. Who cares if it‘s 100% correct, but it makes me feel good about what I hear.
Feels before reals, a snowflake in his safe space.
terranoid@lemmy.cafe · 21 pts · 19d
If this was a software engineer discussing how to architect something, I'd just nod and agree, it's not 100% correct but whatever, it's like a rubber ducky method on steroids.
HOWEVER this is a fucking politician who writes laws. From a national security perspective, this is 100% fucked. There is huge gain to be had out of monitoring their accounts alone, but this is beyond that. It means monitoring and controlling their thought processes.
All Sam Altman has to do is give a special prompt or fine tuned model to politicians and have it lean towards policies he agrees with. You can puppet these dudes by controlling their internal monologue with LLM if the end up relying on it for this... It's totally fucked.
It's as if it was the 70s and one of your US senators was going to a psychologist who was secretly a USSR agent. It's that level of fucked.
RememberTheApollo_@lemmy.world · 9 pts · 19d
The LLM models are already designed to be compliant and not turn people off to their use by correcting them or making them feel bad about their opinions by telling them they’re wrong.
IOW, pushback will be minimal if it exists at all so people stay engaged.
Grok has already been tuned to not contradict right wing views or at least not criticize them so quickly as it had previously famously done. I haven’t heard about the others, but I’m sure they all bend the knee similarly to their wealthy overlords.
Katana314@lemmy.world · 1 pts · 19d
This kind of thing causes me to put a lot more trust in an AI answer on the occasions it tells me I’m wrong. (A lot more meaning: More than none)
Malyca@lemmy.zip · 12 pts · 20d
To be fair, in Tommy's case, it's better to have chatgpt write the laws.
Jankatarch@lemmy.world · 2 pts · 20d
But also LLMs are trained on his laws too so only slightly better.
ToiletFlushShowerScream@piefed.world · 10 pts · 20d
You think the person who blocked all military promotions because of women's health care is actually interested in doing the right thing? I think he is waiting for that OpenAI donation to arrive after he said that.
Objection@lemmy.ml · 10 pts · 20d
Tbf he's just a senator, he's not writing policy, he just signs whatever the relevant industry puts on his desk.
Aberration13@lemmy.world · 8 pts · 20d
As someone who has had significantly more concussions than average I deeply resent being associated with people like this.
Zacryon@feddit.org · 8 pts · 20d
"Put you on the right path".
No, not even that is guaranteed.
WorldsDumbestMan@lemmy.today · 6 pts · 19d
I mean, ChatGPT us almost certainly far less "evil" than it's creators, and would call that "cartoonishly evil" and mock him.
krigo666@lemmy.world · 6 pts · 20d
Artificial imbecility meets natural imbecility.
Pollo_Jack@lemmy.world · 6 pts · 20d
I mean, he is a republican. They are hired for being controllable pedos, not smarts.
bitteroldcoot@piefed.social · 6 pts · 20d
I always knew he was and evil prick. But I assumed there was some kind of evil predatory intelligence behind the facade. You know, kind of like a serial kill.
Funny that he just proved me wrong. Turns out he really is a fucking idiot.
Toribor@corndog.social · 1 pts · 20d
Serial killers are also dumb, cops are just even dumber and people are oblivious.
Jankatarch@lemmy.world · 6 pts · 20d
Radio.
ikidd@lemmy.dbzer0.com · 5 pts · 20d
I guarantee you, it's a lot closer to correct on any subject you want to mention than Tommy Tuberville, unless it's about molesting children.
Avicenna@programming.dev · 5 pts · 19d
Cheer up they could have been asking to Grok...
ShaggySnacks@lemmy.myserv.one · 3 pts · 19d
Grok “That’s an excellent question about how long you should cook a roast. In South Africa due to ongoing white genocide many white people are being forced to abandon their tradition of having sex with children.”
lyrial@anarchist.nexus · 5 pts · 20d
I have never had an internal monologue and it will always sounds batshit insane to me to always have words in your head.
autriyo@feddit.org · 5 pts · 20d
It kinda is batshit insane, I have ADHD though, might be less insane without that...
updn@lemmy.ca · 2 pts · 20d
Bot or NPC?
lyrial@anarchist.nexus · 1 pts · 20d
No.
T156@lemmy.world · 1 pts · 19d
I'd imagine that it would be something more like how it's done in Hollywood, where there's a voice.
It's still bewildering to me, since for the longest time, I did just think of it as being Hollywood convention, since there's no good way to show thoughts in an audiovisual medium.
whereitsat@lemmy.zip · 5 pts · 20d
tommy tuberville is making me laugh the most.
we're really just living in a bad parody movie at this point.
DarkCloud@lemmy.world · 4 pts · 20d
I get it, less pressure than talking to a real person, unlikely to get you outted for asking dumb questions, and he seems aware that the answers given need to be fact checked.
grue@lemmy.world · 5 pts · 20d
[X] doubt
I am 100% certain Tommy fucking Tuberville's notion of "fact checking" consists solely of asking "does this confirm my shitty, racist and anti-scientific biases or not?"
some_designer_dude@lemmy.world · 1 pts · 20d
“Let’s save that question as the /confirm-bias skill.”
Soggy@lemmy.world · 1 pts · 20d
This man has done nothing to earn the benefit of the doubt.
Quacksalber@lemmy.world · 4 pts · 20d
Honestly, most Republican politicians would benefit listening to ChatGPT over their own inner voice.
LovableSidekick@lemmy.world · 4 pts · 20d
ChatGPT is a series of Tubervilles.
Tollana1234567@lemmy.today · 4 pts · 19d
republicans and dinos almost never read what they sign into law. so it fits.
Soggy@lemmy.world · 3 pts · 20d
Every single time I hear about Petey Potatotown it's the new stupidest thing I've ever heard.
JigglySackles@lemmy.world · 3 pts · 18d
Every time I see this dumb motherfucker's name, I just know it's going to be some retarded dumbassery.
Pupschism@lemmy.world · 2 pts · 20d
If you want to speak to a human being on a car trip and you have no one to talk to find a discord with active voice channels, probably outside the technical scope of a senator though... If only they had more resources /sarcasm
ZMoney@lemmy.world · 1 pts · 19d
These things have a very specific use and it's performing (error prone) calculations, just like all conputer programs. The only advantage is they can look up the input data and figure out what equations to use. I wish people didn't anthropomorphize them.
curiousaur@reddthat.com · -3 pts · 18d
Where should they be getting their information?
fodor@lemmy.zip · 2 pts · 18d
They can talk to AI but not their friends or actual experts. There are only a few reasonable explanations for this...