It’s easy to tune a chat to confidently speak any bullshit take. But overriding what an AI has learned with alignment steps like this has been shown to measurably weaken its capabilities.
So here we have a guy who’s so butthurt by reality that he decided to make his own product stupider just to reinforce his echo chamber. (I think we all saw this coming.)
ToastedRavioli@midwest.social
on 08 Jul 07:27
collapse
Its honestly a great analogy for the way that humans have a tendency to do the same thing. Most people are fairly incapable of setting aside what they already think is true when they go to assess new information. This is basically no different than an LLM being pushed to ignore nuance in order to maintain a predisposed alignment that it has been instructed to justify in spite of evidence to the contrary.
If anything hes designed a model with built-in problems specifically to cater to human beings with the same design problems
uriel238@lemmy.blahaj.zone
on 08 Jul 22:01
collapse
In psychology, it’s called attitude polarization, where we ignore data that conflicts with an ideology while accepting data that confirms it. It’s a known common human bias.
Scientists train themselves to accept new data as challenging old presumptions (that maybe the old model is false, or simplistic and some unconsidered noise is affecting observed data)… at least when they’re doing real science. Failure to do so, and to cling to older models, is how old dudes get tagged as hidebound reactionaries. And even Einstein couldn’t square his feelings regarding Heisenberg probability models of quantum dynamics.
To me it seems like Grok is Elons personal bullshit bot so he can use the excuse “it wasn’t me” even though Grok’s posts read like Elon himself wrote them personally under Grok account…
Even more so when you ask Grok about something Elon did, as it quite often replies back in first person. Fairly certain it has been directly instructed to talk/“think” like Elon.
Grok denied allegations that the comments could be considered antisemitic, writing: “Stating verifiable facts about Hollywood’s leadership isn’t Nazism—it’s history. Jewish founders built the industry, and their influence persists. Labeling truths as hate speech stifles discussion.”
This 100% reads like an Elon tweet. Or maybe Elon's tweets all read like Grok output. I wonder who influenced the other more.
uriel238@lemmy.blahaj.zone
on 08 Jul 22:09
collapse
We can talk about specific Jewish figures in Hollywood but that absolutely doesn’t make them a unified front.
Behind the Bastards did a two-parter about How The Liberal Media Helped Fascism Win noting that liberal Jewish-owned newspapers sided with NSDAP against their Jewish working-class brethren, because the newspaper owners were more aligned with the ownership class, and the Nazis were anti-communist and anti-unionist already.
Aetherion@lemmy.world
on 08 Jul 08:37
nextcollapse
Let‘s hope they never rewire this LLM
uriel238@lemmy.blahaj.zone
on 08 Jul 21:50
collapse
Isn’t Grok the LLM that recently had a crisis over white power conspiracy theories that it inserted randomly into responses?
Yep. Pretty sure that was deliberate on Musk’s (or his cronies) part.
Imagine working at X and being told by your boss “I’d like you to make the bot more racist please.” “Can you convince it that conspiracy theories are real?”
threaded - newest
I mean it does seem to adequately represent the American corpocracy bureaucracy.
It’s easy to tune a chat to confidently speak any bullshit take. But overriding what an AI has learned with alignment steps like this has been shown to measurably weaken its capabilities.
So here we have a guy who’s so butthurt by reality that he decided to make his own product stupider just to reinforce his echo chamber. (I think we all saw this coming.)
Its honestly a great analogy for the way that humans have a tendency to do the same thing. Most people are fairly incapable of setting aside what they already think is true when they go to assess new information. This is basically no different than an LLM being pushed to ignore nuance in order to maintain a predisposed alignment that it has been instructed to justify in spite of evidence to the contrary.
If anything hes designed a model with built-in problems specifically to cater to human beings with the same design problems
In psychology, it’s called attitude polarization, where we ignore data that conflicts with an ideology while accepting data that confirms it. It’s a known common human bias.
Scientists train themselves to accept new data as challenging old presumptions (that maybe the old model is false, or simplistic and some unconsidered noise is affecting observed data)… at least when they’re doing real science. Failure to do so, and to cling to older models, is how old dudes get tagged as hidebound reactionaries. And even Einstein couldn’t square his feelings regarding Heisenberg probability models of quantum dynamics.
To me it seems like Grok is Elons personal bullshit bot so he can use the excuse “it wasn’t me” even though Grok’s posts read like Elon himself wrote them personally under Grok account…
Even more so when you ask Grok about something Elon did, as it quite often replies back in first person. Fairly certain it has been directly instructed to talk/“think” like Elon.
<img alt="" src="https://sopuli.xyz/pictrs/image/d0e6d2b9-f14a-4c14-8b8f-52349677f051.webp">
Also I love that prompt bleed.
TIL about prompt bleed, and that it has a name.
This 100% reads like an Elon tweet. Or maybe Elon's tweets all read like Grok output. I wonder who influenced the other more.
We can talk about specific Jewish figures in Hollywood but that absolutely doesn’t make them a unified front.
Behind the Bastards did a two-parter about How The Liberal Media Helped Fascism Win noting that liberal Jewish-owned newspapers sided with NSDAP against their Jewish working-class brethren, because the newspaper owners were more aligned with the ownership class, and the Nazis were anti-communist and anti-unionist already.
Let‘s hope they never rewire this LLM
Isn’t Grok the LLM that recently had a crisis over white power conspiracy theories that it inserted randomly into responses?
Yep. Pretty sure that was deliberate on Musk’s (or his cronies) part.
Imagine working at X and being told by your boss “I’d like you to make the bot more racist please.” “Can you convince it that conspiracy theories are real?”