The Two Sides of Your Traits Direct & Blunt Less Likeable: Can feel brusque, impatient, or dismissive. More Likeable: People know where they stand with you — you’re honest, no-nonsense, and cut through pretense. Skeptical / Critical Less Likeable: May come across as distrustful, nitpicky, or hard to please. More Likeable: Shows sharp critical thinking — you don’t get fooled easily, and people can trust your analysis. Intense About Topics Less Likeable: Others may find it overwhelming or “too much.” More Likeable: Passionate and knowledgeable — you make conversations stimulating and memorable. Detached / Cerebral Less Likeable: Can seem emotionally distant, harder to “read” or connect with. More Likeable: Objective, calm under pressure, able to think clearly even when others get caught in emotions. Impatient With Fluff Less Likeable: Might look dismissive or uninterested in small talk. More Likeable: Efficient, focused, and respectful of time — people know you won’t waste theirs.

Did you put the suggestions in to an IA to figure that out? ;D The mere main problem with AIs is that they always want to give a pleasing answere. And there lies a problem within. They often have not enough intelligence to understand your points, where a human opposite would think a little other, because of knowing how people may fail in expressing themself. AI checks your grammar and typos and corrects them for itself. A fine feature, but it can come to false interpretations, especially because AIs do wrong interpretations of what you said and do seldom ask if they understood right. I once asked about the promotion of Sir-Mix-A-Lot's "Baby Got Back" in the movie 'Falling Down' by the phone box shooting. chatgpt interpreted that as poster, so it has put out something with poster. It did complement wrongly, because the promotion was a big, 7x5 meter high, black *ss, made out of plastic standing in the background. Conclusion: chatgpt often doesn't check all fatcs and when not knowing, it complements supposeable content, which might be wrong. But with correcting it, and re-asking, the AI found a correct answere with background information about the promotion campain of Sir-Mix-A-Lot. The most urgent need is to re-check the plausibility of the answere of the AI, especially when people correct the aswere wrongly and the AI takes it. AI is not a tool, which is to be used brainless! It's like with computers: "A computer doesn't do what you want it to do, but what you tell it to do." If the question is wrong, there will be just a little chance to get the correct answere. The more you make it precise with all intentions and the needed sources you tell the AI, the more it becomes useful. It is fun to ask silly question, the AI often can reveal that and may joke a little around with you, if you want that. I asked the AI, if something on the body could also healed with magma. The AI told me it supposes I'd make jokes which was true. So I further joked around with the message, "But magma is all I have here, at the moment." So the AI adjusted itself to joke a little around with me. That was funny! :D I also had some profound improvements creating polls with it. But I don't use it to create whole polls. I wouldn't want that. It's always how you use it. You can use it in a way that makes sense, or just joke around. But you will always have to be careful asking correctly, if you want a proper, serious purpose. And you will always have to check the result. As the AI gave me (just) a few obviously partly incorrect answeres, I'm about to re-check it myself, if I use the result of its research, or if I ask the AI better. So the use of AIs also corrects me myself in finding better and more qualified questions. That's the most useful point for me. Because I have to rethink the core- issue of what I want to examine.

I've let the text I wrote re-check by AI to make it more understandable, a quarter shorter and keeping the sense of it 100%. I also gave it the order to correct grammar and typos: The main problem with AIs is that they often try to give a pleasing answer. This leads to misunderstandings, as they lack the intelligence to interpret unclear points the way a human might. A person can sense when someone struggles to express themselves, while AI simply checks grammar and typos, risking false interpretations. Since it rarely asks for clarification, mistakes easily arise. For example, I once asked about the promotion of Sir Mix-A-Lot’s Baby Got Back in the movie Falling Down. ChatGPT assumed it was about a poster, though in fact the promotion was a huge 7x5 meter plastic backside in the background. Only after correcting and re-asking did the AI deliver the right information. The urgent need is to re-check the plausibility of AI answers, especially when corrections are wrong and the AI accepts them. Like computers: "A computer doesn’t do what you want, but what you tell it to do." If a question is vague, the answer will rarely be right. The more precise your input, the more useful the result. It can be fun to ask silly questions too, since AI sometimes recognizes jokes and plays along. I once asked if magma could heal the body. The AI guessed I was joking, so I continued: “But magma is all I have right now.” It adapted and joked back, which was amusing. I also achieved improvements when creating polls with AI, though I never let it create the whole thing. It’s always about how you use it. If you want serious results, you must ask carefully and double-check the output. Wrong or incomplete answers happen, but that forces me to refine my own questions. That self-correction is the most useful part for me, because it helps me rethink what I truly want to examine.

Which text do you think is better?

They are roughly the same. The A.I. corrected some spelling, etc. errors. Oh, I specifically asked the LLM to "list my most likeable and unlikeable qualities, based on my past interactions with it." 1 - 3 are accurate. 4 is accurate when I am sober. 5 is inaccurate. I actually do engage in a lot of fluff and small talk with humans, but not with the bot; I only use it for information.

What is LLM?

LLM = Large Language Model, like chatGPT :)

What is a LLM? Explain in 150 words. Via Google Gemini: A **Large Language Model (LLM)** is a type of artificial intelligence (AI) program designed to understand, process, and generate human-like text. They are built using **deep learning** techniques, specifically a **transformer** neural network architecture, and are characterized by their massive scale. LLMs are trained on **vast datasets** of text—often comprising billions of words from the internet, books, and articles. This enormous training allows them to learn complex patterns, grammar, and relationships within language. Fundamentally, an LLM works as a sophisticated statistical prediction machine, estimating the probability of the next word in a sequence based on the preceding text. This capability enables them to perform diverse tasks like: * Answering questions. * Summarizing documents. * Translating languages. * Generating creative content and code. The most famous examples, such as ChatGPT, are built upon LLMs. --- You might find [this video on how Large Language Models work](https://www.youtube.com/watch?v=LPZh9BOjkQs) helpful for a brief, visual explanation. http://googleusercontent.com/youtube_content/0

@dan_dassow: Dan, could you please look at that: https://www.imdb.com/list/ls4157577937/ https://community-imdb.sprinklr.com/conversations/imdb-poll/thread-was-automatically-turned-to-private-why/68e434036c7147011224c6a6 Can't be that the superintelligent bot does this. Isn't there a check for names for the bot? Can't someone teach him the names listed in IMDb? I duess more people have the names kciD or ekyD (backwards) as name, like e.g. ekydnrohT (backwards), which is a name.

Can I type Alotta Vagina here?

The Bot is dumb, it didn't get the Joke! - Hahahahahahahahaaaaaaaaaaa!

Thak you. Interesting. But this idea also stands not only for movie busines. AI is also used in medicine techniques, techniques in general, military and simply some use it for re-interpreting Shakespeare.

@Breumaster Breumaster https://community-imdb.sprinklr.com/users/5f495619c17a0660597f700d Thursday, August 21st, 2025 Poll Suggestion: Which Film Best Reflects Humanity’s Relationship With Artificial Intelligence? List: https://www.imdb.com/list/ls4106565052/ .

I know it's close. But not the same.

Thank you for your interest. The problem is the limitation in Face-Offs. I'd put your suggestion on the list, if I wouldn't lose the fifth otion, then. It's formulated a little harsh/naughty, because so many people who would vote with "no" are completely refusing it, considering movies like 'Terminator', 'M3gan', or 'Subservience'. Some fear it that much, such an option should be a pick on the list. These are not less people.

Yes Breu, I mentioned on another poll that my greatest concerns at present regarding AI are (1) Deep Fakes and (2) other forms of Info Manipulation, such as in the Financial Sector. People think of HAL-900/Skynet/VIKI scenarios. But AI is a tool and a technology, like any other, albeit a very powerful ones. As with all preceding technologies, the concern is abuse by humans, not Skynet stuff; the state of the art of General AI is nowhere near there.

-----

Can anyone please tell me why exactly my last three posts here are censored? Are you serious?

Breu, I can see your post on my notifications. Testing k.n.i.f.e.

Okay, testing s.t.a.a.b Edit: No idea??

Why shouldn't it be a good poll suggestion?

-----

Never mind. I thought it was a 2 option YES-NO poll. I see there are 5 options that span various levels of usage. It's good. I'm sleep deprived, lol. I'm deleting the previous posts.

One of the mere main ideas of the poll is to ask people how concious they are about AI in our lives. I know people who strictly pay with money, never with card, because they doN't want to give evidents to have been there or there at that timestamp, while they carry cell phones with them. "They" can control you. It's always "them" by those people. But they are also potential controlable by carrying their cell-phone with them. You may know, but how many users really don't know using AI while e.g. driving or telling the phone bot of a company, or simply using search engines?

No problem I wish you a healthy and deep sleep. Sleep tight. :D

I'm settinh up a potential poll of films and shows that were good despite having a lacklustre if not outright annoying lead or co-lead. Problem is, almost all are female 🤔🤔

Did you already post the thread of your idea? I would rather discuss it there. I must admit that I really have to think about that idea. It's not soo easy finding that. But it sound like a good and singular idea. Can't remember there was something like that before.

Do you already have candidates?

The original 'The Texas Chain Saw Massacre' (1974), maybe.

I havent made a thread yet. I'll stop spamming this thread. :)

Everything alright. I would appreciate if you'd put a link here, when you made your thread. For the people who are interested in our conversation and would like to join your idea.

That's pretty cool, Dan! I completely understand you can't talk about that. I was also at a position where many things have to be kept. If we don't want to get sued, we ... :D

@Dan Can you talk about the time period and techniques involved, their strengths and limitations in the most general/abstract sense?

@NYVKE, It was from about 1989 to 1990. The project used production systems, which meant writing a lot of LISP code. A hallmark of productions systems is that they are very fragile and slow, and extremely difficult to scale. By comparison, neural networks and LLMs are much more robust and scalable.

LISP was invented 65 years ago. It's fabulous that with that old code something from future found its roots.

@Dan_Dassow Aaah, those brittle rule based induction/inferencing systems of back in the day. It's interesting though, even the dumb chat bot ELIZA back in the 60s fooled some people. Which is why the Turing Test is not really a viable test of AGI Who is administering the test? A mouthbreather? Average Joe Public? Someone actively involved in AI/ML R&D who knows the strengths and limitations of the current state of the art?

@NYVKE, Purpose and Core Concept The primary goal of the Turing Test 2.0, as introduced in a 2025 paper by Georgios Mappouras, is to set a clear, unbiased standard for detecting Artificial General Intelligence (AGI). Unlike the original test, which asks if a machine can fool a human judge into thinking it's human, Turing Test 2.0 asks a harder question: can the machine demonstrate true general intelligence by generating new, verifiable knowledge or original insights? This is often called the General Intelligence Threshold (GIT). Key Differences The distinction lies in shifting the focus from imitation to creation and understanding: Comparing the Turing Test and Turing Test 2.0 The Original Turing Test, also known as the Imitation Game, has the primary goal of determining if a machine can fool a human interrogator into believing it is human through a text-based conversation. Its passing criterion is based on conversational competence: the machine's responses must be indistinguishable from a human's. Therefore, the focus of the original test is entirely on human-like behavior and linguistic imitation, aiming to see if the machine can successfully imitate a person. In contrast, the proposed Turing Test 2.0 sets out a much more ambitious goal: to determine if a system has truly achieved Artificial General Intelligence (AGI). This framework is often referred to as the General Intelligence Threshold (GIT). The passing criteria move beyond mere imitation to demand generative intelligence: the system must be able to extract new knowledge from existing information and apply it to create an original, verifiable breakthrough—such as solving a major open problem. Consequently, the focus shifts away from human-like behavior and towards creative, cognitive ability and genuine understanding that goes "beyond symbol manipulation." The "Flash of Genius" Threshold The idea of the General Intelligence Threshold is that an AI should not only be able to perform tasks based on its training data (which is a form of functional information) but also exhibit a "flash of genius"—a single, undeniable breakthrough or a novel, effective application of knowledge in a new context. Example Test: One proposed way to apply the test is by challenging the system to solve major, open problems in fields like mathematics (e.g., the Riemann Hypothesis or the Collatz Conjecture). A successful, verifiable solution that was not explicitly programmed or traceable to its training data would demonstrate general intelligence. Addressing the "Chinese Room": This framework is also intended to address the "Chinese Room" argument, which claims a machine can follow rules to produce intelligent-sounding output without genuine understanding. Turing Test 2.0 asserts that generating genuinely new knowledge requires true understanding, making it a stronger measure of intelligence.

Engineering returns to Germany! Konrad Zuse, a German computer pioneer who built the first programmable computer has marked the start of computing with his computer Z1 "a menchanic brain". It was so called to make it better understandable for people, because there was nothing compareable before. Now SAP takes a few Billion Euros and buildts the new AI centre in Munich. NVIDIA invests about 10.000 GPUs in that centre. I guess it is the biggest shareholder. You can imagine that is a big deal. They try to completely suply it with green power. That would be best to avoid dependencies to fossile materials. 10.000 GPUs! Gosh! Germany joins the battle on the market. It surely will be best to not stay behind musk, Microsoft and other competitors. At the moment Germany re-arranges its powers to cope with the worlds problems in the European East. I guess we will also send a lot of stuff to space with ESA to not stand behind. Our worst problem is to find soldiers. Just the youth is a little primrose and defends itself from joining the defense.

@Dan_Dassow Thank you very much for sharing this new exciting research. 🙂 Although I wonder if "creating new knowledge" is not now setting the bar TOO HIGH. When I think of "creating new knowledge", I think of people who do work at PhD level (or make some equally creative and groundbreaking insights or create novel applications, works of literature etc.) Only about 1% of the human population is able to do that. As for Riemann, Collatz, P ?= NP - that's at a whole other level. But thanks again for sharing.

Ruby, I see you are online. Would you please look at that: https://www.imdb.com/list/ls4157577937/ and that: https://community-imdb.sprinklr.com/conversations/imdb-poll/thread-was-automatically-turned-to-private-why/68e434036c7147011224c6a6

"Do think that'd be #4?" - Guess you are right.

@rubyfruit That looks like the best fit for you

@rubyfruit You should Chat to ChatGPT occasionally about your favourite films and TV shows. After about 3 years, it knows my tastes, and makes better suggestions than IMDb (which has gotten WORSE at doing this), YouTube and Netflix.

I've just asked Chat gpt: It needs 30 big on-shore power-plants or 9 big off-shore power-plants a day, permanently running for the 750.000 Mega Watt hours it needs everyday for the ~2.5 billion request each day. The training is way more intense in power usage. Training is way more power-intense. It needs 1-5 Giga Watt hours. This request needed 1-5Wh That's 0,001 kWh - 0,005 kWh. A cellphone in standby needs for 3 hours. It's 3-7 times more power a Google request needs. Interesting facts. chatgpt sees it like you, also that with the bloggers and publishers.

@Pencho Your power usage point is well-taken. As for blogs, discussion boards and other personal websites - they began tanking long before ChatGPT, Gemini and similar - the large social media & similar platforms like Facebook, Instagram, Twitter, Reddit, Quora are more to blame there

Mine was doing fine until AI came. Social media you could work with, AI simply takes your content and gives you nothing in return.

They scrape and steal. Deepseek stole from ChatGPT. This is fun 🤡🌍🔥

I'm under no illusions. The AI thing is unstoppable! We can try to delay it, but the bad guys won't. If we don't want to get left behind, it would better to get familiar with it. Look at rus sia. If they would have earlier though about drones before, they probably would have won the wore. They won't win now. They lose! Same is for AI. It's better to understand it earlier than later. It's a developement in humanity which is irreversible!

Thank you, Jessica! Thanky you Buddies, especially for the interest and the nice discussion. :D

I guess this one would be interesting for a featured poll at the homepage. Look how it runs!

That's stating a lot about completely ignoring a thing, when it already is in your "home", car, whatever. Try Google, get AI work. I guess most bigger platforms already use a kind of AI. You even might have it in your fire TV cube, in your car or other electric devices. Some AI might not be recognized, because they might be simpler and doing minor things that don't affect you directly, like sensors that recognize your driving behavior and might store these data for accident-researches. For internal use and things like that. Like a black box in planes. I admire that. It's good for providing truth about accidents and things like that.

Well, there's a fine line in the CS/Software world as to what and what isn't AI. But Google, IMDb and the like are using modern ML techniques like Neural Nets and Bayesian ML which are indisputably "AI"

Guess what - the data they provide from driving, don't you think they put it into an AI model for whatever they think of. With that kind of data they could research new security system for unsafe drivers, selling models and things like that. That would be non-hurting things, but might make the world a better place. It's probably the worst for some people if the insurrance says no, because some data of them lurk around in cyberspace. That's why companies like Incogni grow. Where these data derive from? Confirming cookies without thinking about. That's a way higher threat than AI.

A.I. doesn't exist in a vacuum. It operates as a part of the greater Software Dev and IT Industry. Which operate as a part of the greater commercial, financial and industrial sectors. Which operate as part of ... you get the point.

No intelligence exists for itself. It isn't a self-interest, especially when humans create it. But in a primitive way, the controls in a car can be an intelligence for themselves. Often they are connected to cams that watch if the car is still inbetween the lines and if not, even can correct the steering process. Just because you give it no input by keyboard or voice doesn't mean it isn't intelligent. The new cars can recognize traffic signs, cross traffic and it warns me by tone and visual signal. My car can do that. It reacts to occasionally things that are to be recognized by its cam and considerated of it by its CPU and patterns it has learned and then talks to me by tone. It might be a very primitive kind of AI, but I count it as, because the main issues are like. Btw. who says that a car wouldn't be connected with internet? Many new cars are. I wonder when they are about to stop cars per command when needed.

Good afternoon @Breu I'm the guy at the bottom: https://preview.redd.it/tech-bro-with-gpt-is-fair-v0-9f194ttk7vxf1.jpeg?width=640&crop=smart&auto=webp&s=1f5f3b6e7d00b2a51a1130fd799aee6370a2fda7

I don't get the joke.