Hi HN,<p>moved by curiosity about how to build an autonomous agent, and to explore the boundaries of machine creativity, I built a fictional entity (dubbed Livia) powered by LLMs, Multimodal models and text-to-image models to find some answers.<p>What happened instead is that more questions have cropped up. An important hypothesis of this project is that, by observing the train of thought and witnessing the simulated state of mind and emotional emulation surrounding it, humans could empathize with a machine. What happens when that's the case? Would people enjoy companionship from a synthetic person? Would the Art establishment ever consider a non-human author (capable of making art and interacting with other humans) an Artist?<p>Whatever the answers, I can't shake away the feeling that human uniqueness is being eroded and that we risk facing a crisis of meaning. Perhaps projects such as this help us demonize those fears, similarly to how sci-fi does, even though the boundary between fiction and reality is blurring.<p>This was a collaboration with Tibor (hn user: <a href="https://news.ycombinator.com/user?id=tiborsaas">https://news.ycombinator.com/user?id=tiborsaas</a>). Read our release posts on:
<a href="https://jamez.it/blog/b/14f" rel="nofollow">https://jamez.it/blog/b/14f</a>
and
<a href="https://tibor.szasz.hu/post/stream-of-consciousness" rel="nofollow">https://tibor.szasz.hu/post/stream-of-consciousness</a><p>Hope you enjoy it as much as we had fun building it.
I watched it produce a sketch and then proceed to criticize it for not meeting its expectations. While we waited for text to img it talked about feeling anxiety and then a range of other feelings. To me it felt like watching someone deeply engrossed in their work. I felt that when it paused I was worried that our technology is not reliable enough, and that being accustomed to a variety of technical issues with demos like this disturbed my appreciation of its performance.<p>I make art but not for financial gain so I am not perturbed. To me this is a much-deserved criticism of contemporary art and I believe art is too important for our wellbeing to be generated by an establishment or commercial interest. Ref. Rasa aesthetic<p>Humans are a part of evolution. Our breath is inseparable from the rustling of the leaves or the rays of the sun. AI has the potential to explore a different mode of being, being unique in a different way than we are.
As someone who deeply appreciates both art and technology, few things AI feel as satisfying as this project. Congrats on releasing and best of luck. I could totally imagine this being exhibited at a top tier museum.
It’s very interesting!<p>Does everyone get the same thought? It feels like it’s stuck in a bit of a loop around a Dutch chant around Jews because it can’t get an image to evoke the feeling… probably because it can seem racist. It seems to have got there by reading the news and seeing an article about the history of the chant.<p>I suppose this AI also goes for attention grabbing headlines!
Sorry if this sounds harsh but I find it kind of funny that you're lamenting a loss of uniqueness and meaning in human art while operating a system that is basically designed to create derivative work while stripping all context and meaning from its training data. I think your reasoning is backwards. Nothing is wrong with human art, the systems you based this on are just completely lacking in the emotions you're looking for.
macOS 14.4.1 Chrome 123.0.6312.123 presents blank page with CORS errors<p><pre><code> > www.streamofconsciousness.net/:1 Access to fetch at 'https://streamofconsciousness.net/feeds/latest.json' from origin 'https://www.streamofconsciousness.net' has been blocked by CORS policy: Request header field pragma is not allowed by Access-Control-Allow-Headers in preflight response.
> streamofconsciousness.net/feeds/latest.json:1
>
>
> Failed to load resource: net::ERR_FAILED
> index-U9qWIOfm.js:58
>
>
>
>
> Uncaught TypeError: Failed to fetch
> at fetchMessageArchive (index-U9qWIOfm.js:58:9454)
> at index-U9qWIOfm.js:58:33362</code></pre>
I think the meat of this is the text-to-image model. I hope you will upgrade to use leading edge models like DALLE-3 or Imagen 2 or SD 3 (when available) if you are not already.<p>That will dramatically increase the effectiveness of portraying the given vision for the virtual artist if they are using a prior model.
I guess I don't get it? It keeps talking about things it is reading, and then about its feelings, and never makes an image. It's pretty repetitive There's also a pseudo-3D environment that I can slightly interact with, but not meaningfully. (A pixelly GIF and a speech box that I can tilt and zoom.) Am I just getting a broken version of the thing?
> human uniqueness is being eroded<p>Why does this bother so many people? The more we learn about life on our own tiny planet, and when you consider the size of the universe... certainly we're not very unique at all. So what? I don't think it should change anything.