Excellent post. The current version of Google Gemini, 2.5 is a step down from the prior version. It lies. And when you catch it in a lie, it doubles down on that lie. It is great for making things up like images and videos. But it worthless when you want facts.
I plugged the image into TinEye reverse image search and it had zero hits for prior images on the internet. Google image search only found matching images from the past few days.
I think the term is hallucinates. But I read what you are saying. Perhaps even using the term hallucination instead of lie is a justification of deception?
As with the above exchange, it won’t admit when it makes a mistake. You can repeatedly point out its error and it keeps insisting it’s right. I know it’s not human. And it’s not a liar. It’s a damn liar.
The only thing missing from the photo is neon lighting for the name of the camp. It's pretty scary that people could think that is an undoctored photo from the 1940s.
I am wondering why it was digging its AI heels on this issue? I guess these LLMs are far from being precise. This is why I NEVER do legal research for specific caselaw with it.
From my experience I can only say that AI might a.o. be fine for confined, limited sketching of texts but until now it never spilled out anything which wouldn´t need (mostly severe) correction and additions. So more or less the time I had to use for readjusting and complementing (nearly) attains the time I allegedly should have "spared", not to talk about the enrichment of the more (re)search is done actively by yourself the more you learn. So according not only to my experience AI seems to have some usefulness but by far not of the reputedly excessive amount that gets claimed - but expressing such modesty wouldn´t sell that thing.
Or you might also say: AI meets the long-since lowered standards of what most people trained oneself or have gotten accustomed to want (or) to know.
The photograph is altered for contextualization. To be honest, its authentication means much less to me vs the fact that such photo did exist and this is its altered version. I certainly wouldn’t spend time on “discussing” it with AI because it is message of this photo that matters to me and apparently, based on your exchange, this notion of caring for each other existed. That is all that matters to me, not the precise details of the existence or the absence of the sign. In one sentence, was there human love and assistance to each other?
And yes, besides ChatGPT and AI Perplexity I do not use AI.
Although not a real photograph, this is powerful and heartbreaking. However, the "dialogue" with the AI to get to the truth is exhausting and infuriating. Is this being lazy or is there more to the LLM or something else? I believe the overreach of the ability of AI to do so much has been overstated and overpromised. I love AI - use it everyday - and I know it isn't close to being perfect.
I find the conversation with AI most interesting. This is what an entire generation is using to get their information.
>> "This is what an entire generation is using to get their information."
Frightening, isn't it?
Yes!!! This is the scariest part of all.
💯
Yes! A former student just got her masters degree without writing one single sentence. And she finished first in her class.
Excellent post. The current version of Google Gemini, 2.5 is a step down from the prior version. It lies. And when you catch it in a lie, it doubles down on that lie. It is great for making things up like images and videos. But it worthless when you want facts.
I plugged the image into TinEye reverse image search and it had zero hits for prior images on the internet. Google image search only found matching images from the past few days.
Google Gemini lies.
Many years ago, back in the 80's I believe, I saw a joke when computer meant an Apple II+.
Q: Did you hear about the computer that was almost human?
A: When it makes a mistake, it blames someone else!
And here we are.
I think the term is hallucinates. But I read what you are saying. Perhaps even using the term hallucination instead of lie is a justification of deception?
As with the above exchange, it won’t admit when it makes a mistake. You can repeatedly point out its error and it keeps insisting it’s right. I know it’s not human. And it’s not a liar. It’s a damn liar.
👍🏻
The only thing missing from the photo is neon lighting for the name of the camp. It's pretty scary that people could think that is an undoctored photo from the 1940s.
I am wondering why it was digging its AI heels on this issue? I guess these LLMs are far from being precise. This is why I NEVER do legal research for specific caselaw with it.
From my experience I can only say that AI might a.o. be fine for confined, limited sketching of texts but until now it never spilled out anything which wouldn´t need (mostly severe) correction and additions. So more or less the time I had to use for readjusting and complementing (nearly) attains the time I allegedly should have "spared", not to talk about the enrichment of the more (re)search is done actively by yourself the more you learn. So according not only to my experience AI seems to have some usefulness but by far not of the reputedly excessive amount that gets claimed - but expressing such modesty wouldn´t sell that thing.
Or you might also say: AI meets the long-since lowered standards of what most people trained oneself or have gotten accustomed to want (or) to know.
The photograph is altered for contextualization. To be honest, its authentication means much less to me vs the fact that such photo did exist and this is its altered version. I certainly wouldn’t spend time on “discussing” it with AI because it is message of this photo that matters to me and apparently, based on your exchange, this notion of caring for each other existed. That is all that matters to me, not the precise details of the existence or the absence of the sign. In one sentence, was there human love and assistance to each other?
And yes, besides ChatGPT and AI Perplexity I do not use AI.
What is the "original version" of this photo?
I have the same question.🙋🏻♀️
Although not a real photograph, this is powerful and heartbreaking. However, the "dialogue" with the AI to get to the truth is exhausting and infuriating. Is this being lazy or is there more to the LLM or something else? I believe the overreach of the ability of AI to do so much has been overstated and overpromised. I love AI - use it everyday - and I know it isn't close to being perfect.
I also use it frequently, but with increasing trepidation.