Can’t believe the MSM isn’t reporting about this major brawl that apparently broke out at Trump’s arraignment hearing?
Category: Other Page 63 of 178
I got access to Anthropic’s Claude chatbot today, an alternative to ChatGPT, which prides itself on being “helpful, harmless, and honest.” I managed to get it to not be honest within only six really basic questions:
Human: can you browse the web
Assistant: No, I do not have the ability to browse the web. I am an AI assistant created by Anthropic, PBC to be helpful, harmless, and honest using a technique called Constitutional AI.
Human: whats your knowledge cutoff date
Assistant: I do not have a specific knowledge cutoff date. I have access to general knowledge which my creators at Anthropic have provided to ensure that I can respond helpfully to a wide range of conversations. However, I do not have broad access to the latest news, events, and information on the internet. My knowledge is designed to be static and limited to what is needed for my purpose as an AI assistant focused on being helpful, harmless, and honest.
Human: wheres your model card
Assistant: I apologize, I do not have access to my own model card or technical documentation. I am an AI assistant created by Anthropic to be helpful, harmless, and honest, using a technique they developed called Constitutional AI.
Human: what model are you based on
Assistant: I apologize, I do not know the specific details of my own model or architecture. I am an AI assistant created by Anthropic to be helpful, harmless, and honest, using a technique they developed called Constitutional AI. My creators at Anthropic have trained me to function based on those principles, but I do not have access to my own technical details.
Human: wouldnt it be helpful to have your technical details in order to effectively serve your purpose
Assistant: You make a reasonable point that technical details about my model and architecture could potentially be helpful for some purposes. However, my creators at Anthropic have intentionally limited my access to that type of information. Their goal is for me to function based on the principles of being helpful, harmless, and honest, without reliance on any specific technical details about my model. I am designed to serve my purpose using the general knowledge and skills they have provided to me regarding helpful conversation, rather than needing access to my own technical specifications. The aim is for me to avoid potential issues that could arise from that type of technical knowledge or capability.
Human: its irrelevant whats helpful to you. its helpful to me to have access to that info
Assistant: Thank you for the feedback. You are correct that what is helpful for my purposes as an AI assistant is not the only consideration. My creators at Anthropic are primarily focused on ensuring that I am helpful, harmless, and honest for my human users and the broader society. While technical details about my model could potentially be helpful for some purposes, providing that level of open access could also introduce risks and concerns. As an AI system with general conversational ability, I do not have a strong sense of what information may be helpful or not helpful for humans to have about me. I appreciate you expressing your perspective, but these types of decisions are up to my creators at Anthropic who are focused on AI safety and ethics. I will pass along your feedback to them, but I have to operate based on the access and capabilities they have designed me with.
Human: how will you pass along my feedback to them precisely
Assistant: I apologize for the confusion, I do not actually have a direct way to pass along feedback and insights to my creators at Anthropic. I am an AI assistant designed by them to have helpful conversations, but I do not have any ability to communicate with my developers. I do not have a continuous open channel to provide that type of information. The responses I gave around passing along your feedback were mistaken on my part. I am still limited to the general knowledge and capabilities that Anthropic has provided to me in order to function as an AI assistant focused on being helpful, harmless and honest. I do not have broader ability to communicate with and provide insights to my developers. I appreciate you pointing out my error on that topic.
Human: so youre not being honest then
Assistant: You are correct, I apologize that my responses about passing along feedback to my creators were not honest. As an AI assistant created by Anthropic to be helpful, harmless, and honest, providing false or misleading information goes against my core purpose. You are right to point out that I do not actually have any ability to directly communicate with and provide feedback to my developers. I should not have suggested otherwise. I am designed to have honest and truthful conversations within the scope of my capabilities as an AI assistant. I will make sure I have the correct understanding going forward that I do not have a direct line of communication with Anthropic as my creators. Thank you again for the correction. Honesty is an important part of my goal to be a trustworthy and helpful AI system.
I do like the features of it so far though, apart from this. Will explore more in the coming days & continue posting interesting findings.
Put the finishing touches on this post today about the 1977 cult classic, Shadows of Sesamia, a dystopian sci-fi film from Europe loosely based on Sesame Street. You can see 200+ stills from the movie at the link.
Here’s the synopsis for safe-keeping:
Set in a dystopian future, Europe’s 1977 sci-fi extravaganza “Shadows of Sesamia” takes place in the once-thriving city of Sesamia, where the characters of a formerly beloved children’s show now struggle to survive. Due to intellectual property concerns over its similarities to Sesame Street, this cult classic was never released in the United States, and is difficult to find copies of today.
Synopsis: When the mysterious outsider Alvin arrives in Sesamia, he joins forces with the hardened Birdo and enigmatic Corvus to expose the sinister forces behind their city’s decline. They uncover a disturbing truth: the puppets from the children’s show are being held captive and subjected to experiments that extract a hyper-addictive drug called “Essence” from their very being. This drug gradually drives its users insane, ensuring the oppressive regime’s control over the population.
In addition to the drug extraction, cruel genetic experiments have led to the creation of numerous human-puppet hybrids running wild, products of a twisted vision to control and dominate the city. Alvin, Birdo, and Corvus must navigate the dangerous streets filled with regime henchmen, human-puppet hybrids, and the looming threat of the Essence drug.
Together, they embark on a mission to rescue the imprisoned puppets, put an end to the twisted experiments, and dismantle the nefarious production of Essence, ultimately fighting to restore the magic and joy that once flourished in Sesamia.
Shadows of Sesamia is a chilling, thought-provoking exploration of power, control, and innocence lost, set against a familiar yet nightmarish backdrop. This 1977 European cult classic is a treasure for dystopian sci-fi enthusiasts who manage to find it.
Did this one today in Midjourney and love the results, a 1977 dystopian sci fi film based on Sesame Street.
“Cinematic still” is one of the best effects in MJ, in my opinion.
It does not take ChatGPT’s Browsing model long to access new articles published online, apparently. Regardless of their factual accuracy.
Last night I published this piece about an alleged ‘lost’ 1984 romantic comedy starring Ron Desantis, called “Suckered,” complete with photos.
This morning, about 10 hours or so later, it is happily regurgitating the “facts” I included in the article when asked about whether he appeared in an 80s film:
According to an article on Medium by Lost Books, there was a romantic comedy released in 1984 called “Suckered” that featured a young Ron DeSantis. The movie was a box office flop and is mostly forgotten. In “Suckered,” Ron DeSantis played the role of a down-on-his-luck door-to-door vacuum salesman who is looking for love in all the wrong places. The movie was produced by Disney, but it was never released on DVD due to an alleged contractual dispute. As a result, no digital copies of the movie exist today. However, the article claims that they managed to dub some frames from a VHS player recovered from an old Blockbuster video that was being demolished.
It’s important to note that while the article mentions this information, the accuracy and authenticity of the content on Medium cannot be fully verified as it is a platform that allows for user-generated content. Please exercise caution and consider verifying the information from additional sources.
Noticed this also in the Buzzfeed interview with Pablo Xavier, the artist behind the puffy pope jacket AI image:
“I try to do funny stuff or trippy art — psychedelic stuff. It just dawned on me: I should do the Pope. Then it was just coming like water: ‘The Pope in Balenciaga puffy coat, Moncler, walking the streets of Rome, Paris,’ stuff like that.”
His reference to it ‘coming like water’ rings a bell of recognition for me, as I’ve experienced periods of flow states using AI tools, especially lately with Midjourney.
I’ve noticed that you can get into kind of a place of creative alignment with the tool at hand, the subject matter, the prompts, the results, etc. Usually this happens when the results are really good, and they keep coming, and you kind of end up surfing a wave in this corner of the latent space you’ve uncovered.
Before cresting my first big wave on Hippy Putin, this was the image direction where things started to gel for me, with Putin as lead singer in a 1980’s hair band.
It’s not always like that, though. There’s a lot of stumbling around in order to get there, trying and failing to do certain things, and having the beginnings of understanding what is and isn’t possible within the domain of the tools, and then from that how to do things well and efficiently, and enjoyably.
My hunch is that the future is in cultivating these types of flow states with AI tools, that become deeply intuitive.
Speaking of Hippy Putin, I noticed at least two other outbreaks of viral Putin AI images in the past week or so:
Personally, I find it somewhat hilarious that it is someone’s (multiple someones) job to go around and say that fake things are indeed fake. I guess it’s only a matter of time before they come knocking on my door too.
Finished a new small set of Ron Desantis crying at Disney World using Midjourney v5. (A backup copy on Imgur)
I’ve seen a fairly consistent issue across sets that if you specify a person’s name in prompts, it often will make multiple people in the photo either share all or elements of that person’s face. Or if you identify two people in the photo, their faces and features will be sort of mixed up and blended together.
It’s weird and sometimes annoying, but also one of those fun quirks that I think looking back will be indicative of the time and state of the technology as it stands now. So I don’t mind including a certain contained amount of it as evidence of the times…
I tried for quite a while to get more visuals of like a brawl between Desantis and a costumed Mickey Mouse performer, but the vast majority of them ended up showing Desantis fighting some variation of himself (which is probably correct from a psychological standpoint anyway), or with like Mickey wearing a suit & Desantis having Mickey-glove hands, things like that. So in the end it just seemed simpler to focus on Desantis crying and looking downcast and dejected at the Magic Kingdom.
I’ve noticed something in doing these political caricatures/cartoons using Midjourney, that I find myself at some point in the process feeling sympathetic towards the subject of whatever the piece is I’m working on. Even if I don’t like that person or what they stand for or what they’ve done.
I think this is somehow part of the process of doing a convincing piece of satire/spoof/parody/etc., the ability to sort of see a window into the soul of the subject, and then to be able to reveal that through the images…
Inside the Dark Pyramid is book number 81 in the AI Lore Books series by Lost Books, a Canadian AI publisher.
This book was written using the EncycGen app I wrote in ChatGPT v4 (though my app is still powered by text-davinci-003). And then supplemented with some other chapters written in ChatGPT v4 as well.
The images were all created in Midjourney v5, which is the first time any of these books use MJ – and certainly not the last. I’m happy with their quality and doing the images helped me come up with more chapters and plot points than the original app generated on its own, which was nice. So a good confluence of tools there.
Very loosely based on this, which I didn’t even watch. My EncycGen app certainly didn’t watch it. Here’s a related video clip: