None of the things you listed mean much when it comes to ethno-nationalism. There are a lot of Chinese that love American brands but still hate the US for nationalist reasons.
Among my relatives I would say all are anti-US. About 5-6 of them vehemently so and want the war to start immediately.
I grew up in China and if you think there’s less propaganda in China compared to the US I don’t know what to tell you.
It's heavy and the extra encumbrance is not great for general purpose computing, though I never felt the need for extra screen real estate while I'm coding, watching youtube, sending emails etc.
That said I still use my AVP regularly, it's a great home theater system for 4k HDR content that's portable for travel. I've owned other headsets like the valve index but the novelty of 3d gaming wore off pretty quickly.
This is kind of the visual equivalent of asking an LLM to count letters. The failure is more related to the tokenization scheme than the underlying quality of the model.
I'm not certain about the specific models tested, but some VLMs just embed the image modality into a single vector, making these tasks literally impossible to solve.
Chess is a great example because it highlights the subtle difference between LLMs and agents. What GPT3.5 does is not quite playing chess but creating realistic chess moves that a human might make.
An LLM could play chess though, all it needs is grounding (by feeding it the current board state) and agency (RF to reward the model for winning games)
I think a lot of the confusion on whether LLMs can think stems from the fact that LLMs are purely models of language and solve intelligence as a kind of accidental side-effect.
The real problem that an LLM is trying to solve is to create a model that can enumerate all meaningful sequences of words. This is just an insane way of approaching the problem of intelligence on the face of it. There's a huge difference between a model of language and an intelligent agent that uses language to communicate.
What LLMs show is that the hardest problem - of how to get emergent capabilities at scale from huge quantities of data - is solved. To get more human-like thinking, all that is needed is to find the right pre-training task that more closely aligns with agentic behavior. This is still a huge problem but it's an engineering problem and not one of linguistic theory or philosophy.
The explanation is perfectly sensical, just too complex for humans to understand as the model scales up.
The thing you're looking for - a reductive explanation of the weights of a ANN that's easy to fit in your head, does not exist. If it were simple enough to satisfy your demands, it wouldn't work at all.
I was writing an SVG editor web app but didn't know anything about the fabricjs api. Here's a portion of my prompts with chatgpt (april 2023 version, the date at the top is the share date). The initial code came from a different chatgpt thread.
I finished the app in 2 days, with a third day for css/visual styling. I previously might have hired someone to do this or tried to figure it out myself, which would have taken about a month.
At the end of this thread chatgpt kind of goes off the rails a bit and fails to center the uploaded image. I think this is because it can't execute the code and see its results, and can only take blind guesses at what the problem might be.
I still had to write about 10% of the code myself, but it's about 10x faster for me to use chatgpt. I think I'd use chatgpt even if it were slower, because I prefer "thinking in natural language" vs "thinking in code"
In pixel space a convnet uses pixel-wise convolutions and a pixel-kernel. If you represent a vector image as a polygon, the direct equivalent to a convolution would be the Minkowski sum of the vector image and a polygon-kernel.
You could start off with a random polygon and the reverse diffusion process would slowly turn it into a text glyph.
I think this approach isn't ideal because you're representing pixels as 150x150 unique bins. With only 71k fonts it's likely a lot of these bins are never used, especially at the corners. Since you're quantizing anyways, you might as well use a convnet then trace the output, which would better take advantage of the 2d nature of the pixel data.
This kind of reminds me of dalle-1 where the image is represented as 256 image tokens then generated one token at a time. That approach is the most direct way to adapt a causal-LM architecture but it clearly didn't make a lot of sense because images don't have a natural top-down-left-right order.
For vector graphics, the closest analogous concept to pixel-wise convolution would be the Minkowski sum. I wonder if a Minkowski sum-based diffusion model would work for svg images.
Hallucinations are an engineering problem and can be solved. Compute per dollar is still growing exponentially. Eventually this technology will be widely proliferated and cheap to operate.
The foreign buyers tax, the empty home tax, and now the total ban on foreign buyers introduced this year has had zero effect on home prices. It's time to stop blaming foreigners.
There are still some good 6 letter names out there, if you're willing to use a brandable (made-up) word. Some I randomly made just now: pletha.com encryx.com oxiply.com
Not seeing anything about the dataset, are they still using LAION? There's no mention of LAION in the paper and the results look quite different from 1.5 so I'm guessing no.
> the model may encounter challenges when synthesizing intricate structures, such as human hands
I think there's two main reasons for poor hands/text
- Humans care about certain areas of the image more than others, giving high saliency to faces, hands, body shape etc and lower saliency to backgrounds and textures. Due to the way the unet is trained it cares about all areas of the image equally. This means model capacity per area is uniform, leading to capacity problems for objects with a large number of configurations that humans care more about.
- The sampling procedure implicitly assumes a uniform amount of variance over the entire image. Text glyphs basically never change, which means we should basically have infinite CFG in the parts of the image that contain text.
I'm not sure if there's any point in working on this though, since both can be fixed by simply making a bigger model.
Among my relatives I would say all are anti-US. About 5-6 of them vehemently so and want the war to start immediately.
I grew up in China and if you think there’s less propaganda in China compared to the US I don’t know what to tell you.