Live data from Hacker News

ChatGPT conversations can be shared publicly

chat.openai.com

81–90 of 172 posts

Re: ChatGPT conversations can be shared publicly

#81
post #72

One of my good ones: https://chat.openai.com/share/2d1ac313-e4bd-4dcd-89b3-5f28fb... (AI with an ability of 'Continuous Thinking')

A research was published yesterday, which brings this concept of "autonomous agents" much closer: https://voyager.minedojo.org/. Pretty intriguing!

Re: ChatGPT conversations can be shared publicly

#82
Well done! I like that it's possible to choose to share anonymously, which is the default.

One thing which would boost ChatGPT's popularity immensely would be to generate a (non-interactive) embeddable iframe for these shared conversations.

Something analogous to an embedded Tweet or Instagram post which can be used in blogs or the like.

Re: ChatGPT conversations can be shared publicly

#83
post #53

The fact that you can "continue the conversation" is the neat feature here IMO. Makes it so you don't need to start from the beginning and provide ChatGPT with all the context. Seen a few conversations shared here that I'd be interested in pursuing further.

It also means our conversation is recorded by openai, doesn’t it?

It saves all previous messages in an array of messages so it ‘remembers’ the conversation from the beginning. So when you post your message you’re also posting the complete message history. So it’s kind of not recording in the sense that it will give your conversation content to other users.

Re: ChatGPT conversations can be shared publicly

#84
post #70
post #13

Earlier quoted context omitted.

> Remember, this response takes a departure from reality and embraces playful inconsistencies for the sake of imagination and entertainment. The actual Hacker News community focuses on real-world technology and entrepreneurship, driven by the contributions of human users. You had to ruin it chatGPT.

Yeah, can’t wait until this garbage becomes unfashionable.

Have you used it? At worst it’s a way better intellisense. I am invested in it being a little more, but not like my robot wife lol.

I’m genuinely curious, why does reading about transformer models and chatbot news upset you, and some others?

Re: ChatGPT conversations can be shared publicly

#85

Seems like a feature that benefits openai to create shareable interactive Twitter like posts… and um… no one else. “Wow, look at this conversation!” … …why did you not just post the text that was interesting? It’s text. Just control-c, control-v. Needing a share button for that seems the height of laziness. Two guesses what comes next; “likes” and “retweets” where you can continue the conversation with your own next…

Dude are you serious? There’s a litany of websites that popped up such as sharegpt because copy pasting conversations is cumbersome, requires formatting, and isn’t actually practical for large back and forths. How the f can people be so cynical, seriously?

Emotional vampires.

Re: ChatGPT conversations can be shared publicly

#86

Earlier quoted context omitted.

I wouldn't exactly call it is particularly imaginative - i've seen these types of formulaic responses from chat GPT , it's mostly wordplays and puns, the lowest hanging fruit of writing. Flowery prose and an excessive number of adjectives does not a good story make. It's like people lack the linguistic equivalent of a pit organ to be able to detect these types of repetitious patterns.

How imaginative do you consider this story: https://chat.openai.com/share/e2a24c03-992a-4e99-8ef6-1d8702... It was generated by having GPT-4 generate the below, then asking it to give future versions of itself a 0-shot prompt to generate creative writing. https://rentry.co/gpt4creativity

Yes, using GPT-4 to bootstrap itself into even better responses is a fascinating process. Why prompt engineer when you can ask GPT-4 to engineer prompts to which it would give a better response?

I also enjoy prompting it to answer metacognitive questions.

Re: ChatGPT conversations can be shared publicly

#87

ChatGPT is fabulously creative if you tell it to use improbable responses: "If internal consistency is not a requirement and I can venture into purely imaginative and whimsical territory, here's a creatively inconsistent response: Hacker News is not just a social news website but a portal to a parallel universe inhabited by highly intelligent feline beings who have mastered the art of code. These sophisticated cats,…

I wouldn't call this fundamentally creative, rather a reworking of existing themes. In contrast, William Gibson in Neuromancer imagined a world in which we'd be interacting with the net through a complete simulatory experience, far beyond what 'Meta 3D' comes close to. Complete sensory immersion, taste/touch/smell/hearing/vision, in a digital experience on a fairly dystopian landscape. Granted, there might have been…

Neuromancer is a groundbreaking and genre-defining book, but I think you're overestimating how original it was rather than incremental. Similar ideas predate it, e.g. Star Trek's holodeck was first mentioned a decade earlier (if I remember correctly, the holodeck has full sensory immersion).

Most human innovation is incremental. As a whole package, the book feels fresh. But examine any particular element, and you can find precursors.

You also have the benefit of choosing a classic book out of millions of published books and comparing it against a single ChatGPT response. Most books are not as original as Neuromancer, and our sample of 1 ChatGPT response probably doesn't show the full range of its potential originality.

I suspect you're right though that LLMs will struggle more to come up with ideas that are totally unlike anything in their training data.

For fun, I asked ChatGPT to come up with some original ideas for a novel:

> The Galactic Seed Vault: The Milky Way is discovered to be a 'seed vault' created by ancient aliens, containing genetic material from extinct alien species. Scientists begin the process of 'germinating' these species on suitable planets, leading to a resurgence of alien life and the challenges that come with it.

> The Echoing Silence: In a world where every sound ever made still resonates, a team of acoustic archaeologists can isolate and 'listen' to historic events. They inadvertently unleash sounds that should have remained silent, causing chaos in their present day.

> The Taste of Emotions: A society where food is imbued with emotions. Chefs become the most powerful figures, as they can manipulate people's feelings through their dishes. The protagonist, a chef, starts a revolution by serving dishes of dissent and rebellion.

> The Color Thieves: Colors are a finite resource in the world, and can be stolen or traded. A black and white underground movement rises, challenging the color-owning elite and demanding a redistribution of hues.

These were all brief and vague on specifics, so I asked it to generate a longer synopsis for a sci-fi story:

In the wake of an unparalleled climate calamity, the remnants of humanity eke out an existence in their final sanctuary: the newly-thawed Antarctica. Society has been reimagined, birthing groundbreaking systems - the Carbon Economy and the VR Democracy. These innovations, while revolutionary, bear their own set of complexities.

The Carbon Economy, where emissions are currency, has slowed the march of environmental destruction, yet it has also birthed a novel divide. The affluent, equipped with advanced carbon capture technologies, enjoy an existence in verdant metropolises, while the less fortunate languish in dense urban sprawls burdened with a high carbon cost.

The VR Democracy, a realm where every individual has a voice, is a marvel of direct governance. Yet, it's not without its flaws. The din of majority rule often silences the whispers of the minority, and the digital nature of this reality fosters a disconnection between virtual decisions and their tangible consequences.

Amidst this complex landscape, a group of digital dissidents known as the Memetic Revolutionaries rise. They wield the power of memes, not merely as vessels of humor, but as profound tools of persuasion. They craft Empathy Memes, potent capsules of shared feeling that open the eyes of the masses to the struggles of those bearing the brunt of the Carbon Economy. This emotional awakening fuels a wave of empathy within the VR Democracy, igniting a clamor for transformative change.

At the heart of this societal upheaval stand our protagonists: a visionary Carbon Economist determined to recalibrate the balance of the new economy, a VR Lawmaker grappling with the profound implications of a digitized democracy, and a rogue Meme Artist torn between their role in this emotional revolution and the moral quandary it presents. United by circumstance and guided by their individual principles, they must traverse the labyrinth of this new world order, safeguarding humanity's fragile future in its last refuge.

Re: ChatGPT conversations can be shared publicly

#88

EDIT: oops this is the correct link: https://chat.openai.com/share/69839da8-317a-4e3b-b215-f9e9ff... Previous comment: GPT-4 failing with a math question: https://chat.openai.com/share/53116ac3-0c44-410a-b367-c5df0a... (I didn't discover this. I was just trying to confirm something that was reported elsewhere.)

It's a language model, not a mathematical model. But plugins enable it to use something actually designed for this sort of work. https://www.wolfram.com/wolfram-plugin-chatgpt/ Part of effectively using a tool like ChatGPT is understanding the boundaries and how to mitigate them. ChatGPT has lots of problems, but can still be incredibly useful.

But but people told me me it can actually reason because it can use abstraction or something … /s

Seriously though, if it can really reason I wonder if you feed it Peano’s Axioms will it from first principles be able to do arithmetic? Then again I guess it already know about Peano’s Axioms since it’s on Wikipedia, so the answer is “no”?

Re: ChatGPT conversations can be shared publicly

#89
post #53

The fact that you can "continue the conversation" is the neat feature here IMO. Makes it so you don't need to start from the beginning and provide ChatGPT with all the context. Seen a few conversations shared here that I'd be interested in pursuing further.

It also means our conversation is recorded by openai, doesn’t it?

It is. The company I work for apparently have purchased a version of GPT4 that does not send data back to OpenAI. The conversation is still recorded by the folks who are maintaining it though.

Re: ChatGPT conversations can be shared publicly

#90
post #84
post #70

Earlier quoted context omitted.

Yeah, can’t wait until this garbage becomes unfashionable.

Have you used it? At worst it’s a way better intellisense. I am invested in it being a little more, but not like my robot wife lol. I’m genuinely curious, why does reading about transformer models and chatbot news upset you, and some others?

Its just so repetitive, with the same types of cargo-cult responses to any criticism. Very similar to bitcoin/tesla in how it gets shilled
Post reply on HN