4 ms·

>Here’s one of the many astounding finds in Microsoft Research’s Sparks of AGI paper. They found that GPT-4 could write the LaTex code to draw a unicorn. a lot

by sgt101 3y ago

>Here’s one of the many astounding finds in Microsoft Research’s Sparks of AGI paper. They found that GPT-4 could write the LaTex code to draw a unicorn. a lot of people have tried to replicate this, I have tried. It's very hard to get GPT-4 to draw a unicorn, also asking it to draw an upside down unicorn is even harder.

midlightdenight 3y ago
The model of GPT-4 those researchers had was not the same that’s available to the public. It’s assumed it was far more capable before alignment training (or whatever it’s called).
- Xelynega 3y ago
  That's convenient. Typically one of the markers of good science is reproducability. How can we trust any of the information coming out of these studies if it can't be reproduced?
  - sgt101 3y ago
    Also - why not make this clear in all the model access documents? Perhaps call it GPT-4P (for Public?) Perhaps also provide other researchers with vetted access. There are a lot of groups trying to evaluate these things systematically - for example "Faith and Fate", "Jumbled thoughts","Emergent abilities are a Mirage" were all very good papers published this year which really highlighted hype in LLM evaluation. Everyone can see that modern LLM's have some great capabilities, the flexibility you can get in an interface by doing intent detection and categorizations using an LLM is great and it is so much easier and quicker than using previous techniques. It's more expensive, but that's improving rapidly. I firmly believe that a new era of great new systems with better interfaces and more functionality will be built on LLM's and other models from this wave of Big Data / Big Model AI, but these are not the precursors of AGI. The problem with the looky looky AGI bunkum show is that it's pulling money into crappy projects that are going to fail hard and this will then stop a lot of money going into projects that could be successful fast. I am seeing the shape of the dotcom boom/bust in what's happening. Microsoft and Intel used dotcom to build and maintain their monopoly position, I think AWS, MS and Google will do the same this time. I think we will see a wave of new companies like Amazon that will "fail" some of them will really fail and disappear, some will half fail like Sun did, but some will go on to build monopolies anew. In the meantime the technology will evolve not for the greater good but instead to serve purposes like advertising distribution that are trivial compared to the benefit we could have seen. Over all we will not capitalise on the potential of what we have for several decades, ironically because of the failures of capitalism. Children will die, wars will be fought but some of us will have nice sweat pants and fun playing paddleball in the sunshine while it all happens. When historians write this up in 100 years they won't really see any of this - they will just see a huge surge of innovation. The dead have no voices...
nyrikki 3y ago
Stocastic parrots are going to stochasticate. The author also cited a few human assisted efforts as ML only. The fact that the author also is surprised that GPT is better at falsifying user input while it struggles at new ideas demonstrates the fact that those who are hyping LLMs as getting us closer to strong AI don't know or ate ignoring the know limitations of problems like automated theorem solving. I think generative AI is powerful and useful. But the AGI is near camp is starting to make it a hard sell because the general public is discovering the limits and people are trying to force it into inappropriate domains. Over parameterization and double decent is great at expanding what it can do, but I haven't seen anything that justifies the AGI hype yet.
RaftPeople 3y ago
A person commenting on this topic at a different site mentioned that there is a lot of content on the internet around how to draw (animals?) with LaTex and a different tool (can't remember the name), so it's unclear if GPT is just regurgitating or if it's generalizing.
- sgt101 3y ago
  I think it regurgitates badly, an interesting pre-print is the "reasoning or reciting" one from https://arxiv.org/abs/2307.02477 https://arxiv.org/abs/2307.02477 But - it's only a pre-print and I don't think that they have taken it forward to a publication so handle with care. Anyway, the relevant evidence and thinking that I would highlight and somewhat agrees with my findings is seen in figure 7 and section 5.6. I struggle to get the quality of results that they saw, but its believed by some people that ongoing development of GPT-4 and throttling of the reasoning for cost reasons may have limited some of its capabilities by the end of this year so that may be some of my problem.
mk_stjames 3y ago
Ask it to draw an image of a unicorn in SVG format instead. I've did this and got a result. Asked for a cat wearing a hat. It drew a circle with dots as eyes and sorta-whiskers with lines and a triangle for the hat. All using just SVG vector code. The claim is result shows that there is understanding of not just words about cats and hats and connections to shapes but also a bit of spatial awareness in the x,y coords needed on the SVG canvas to place the shapes. I was able to do with with several different little characters / cartoons in SVG and while it completely fails every now and then it was better than I would have thought. I think it has been exposed to SVG (via the web) way more than LaTeX vector drawings.

by sgt101 3y ago

midlightdenight 3y ago

The model of GPT-4 those researchers had was not the same that’s available to the public. It’s assumed it was far more capable before alignment training (or whatever it’s called).

Xelynega 3y ago
That's convenient. Typically one of the markers of good science is reproducability. How can we trust any of the information coming out of these studies if it can't be reproduced?
- sgt101 3y ago
  Also - why not make this clear in all the model access documents? Perhaps call it GPT-4P (for Public?) Perhaps also provide other researchers with vetted access. There are a lot of groups trying to evaluate these things systematically - for example "Faith and Fate", "Jumbled thoughts","Emergent abilities are a Mirage" were all very good papers published this year which really highlighted hype in LLM evaluation. Everyone can see that modern LLM's have some great capabilities, the flexibility you can get in an interface by doing intent detection and categorizations using an LLM is great and it is so much easier and quicker than using previous techniques. It's more expensive, but that's improving rapidly. I firmly believe that a new era of great new systems with better interfaces and more functionality will be built on LLM's and other models from this wave of Big Data / Big Model AI, but these are not the precursors of AGI. The problem with the looky looky AGI bunkum show is that it's pulling money into crappy projects that are going to fail hard and this will then stop a lot of money going into projects that could be successful fast. I am seeing the shape of the dotcom boom/bust in what's happening. Microsoft and Intel used dotcom to build and maintain their monopoly position, I think AWS, MS and Google will do the same this time. I think we will see a wave of new companies like Amazon that will "fail" some of them will really fail and disappear, some will half fail like Sun did, but some will go on to build monopolies anew. In the meantime the technology will evolve not for the greater good but instead to serve purposes like advertising distribution that are trivial compared to the benefit we could have seen. Over all we will not capitalise on the potential of what we have for several decades, ironically because of the failures of capitalism. Children will die, wars will be fought but some of us will have nice sweat pants and fun playing paddleball in the sunshine while it all happens. When historians write this up in 100 years they won't really see any of this - they will just see a huge surge of innovation. The dead have no voices...

nyrikki 3y ago

Stocastic parrots are going to stochasticate. The author also cited a few human assisted efforts as ML only. The fact that the author also is surprised that GPT is better at falsifying user input while it struggles at new ideas demonstrates the fact that those who are hyping LLMs as getting us closer to strong AI don't know or ate ignoring the know limitations of problems like automated theorem solving. I think generative AI is powerful and useful. But the AGI is near camp is starting to make it a hard sell because the general public is discovering the limits and people are trying to force it into inappropriate domains. Over parameterization and double decent is great at expanding what it can do, but I haven't seen anything that justifies the AGI hype yet.

RaftPeople 3y ago

A person commenting on this topic at a different site mentioned that there is a lot of content on the internet around how to draw (animals?) with LaTex and a different tool (can't remember the name), so it's unclear if GPT is just regurgitating or if it's generalizing.

sgt101 3y ago
I think it regurgitates badly, an interesting pre-print is the "reasoning or reciting" one from https://arxiv.org/abs/2307.02477 https://arxiv.org/abs/2307.02477 But - it's only a pre-print and I don't think that they have taken it forward to a publication so handle with care. Anyway, the relevant evidence and thinking that I would highlight and somewhat agrees with my findings is seen in figure 7 and section 5.6. I struggle to get the quality of results that they saw, but its believed by some people that ongoing development of GPT-4 and throttling of the reasoning for cost reasons may have limited some of its capabilities by the end of this year so that may be some of my problem.

mk_stjames 3y ago

Ask it to draw an image of a unicorn in SVG format instead. I've did this and got a result. Asked for a cat wearing a hat. It drew a circle with dots as eyes and sorta-whiskers with lines and a triangle for the hat. All using just SVG vector code. The claim is result shows that there is understanding of not just words about cats and hats and connections to shapes but also a bit of spatial awareness in the x,y coords needed on the SVG canvas to place the shapes. I was able to do with with several different little characters / cartoons in SVG and while it completely fails every now and then it was better than I would have thought. I think it has been exposed to SVG (via the web) way more than LaTeX vector drawings.