Hacker News Reader

Comment by bhouston | Hacker News Reader

100%. This person is trying to find patterns in random noise and believes they are meaningful. The original post hurts my head with its bad logic.

View on HN

I'm sorry it hurts your head. I'm happy to sponsor a packet of paracetamol or some water if that helps. Ultimately, this is fun, not science. I'm just happy that after all these attempts, it finally got to a unicorn.

It got to a unicorn quite easily when the output was set to Tikz as I’ve done here:

https://bobjansen.net/drawing-with-chatgpt/

That looks nothing like a unicorn, but the three triangles at the top right in the third picture look like the mask of some evil manga villain alien robot.

Well done for being a good sport, but I'm willing to bet that tomorrow's shape will not resemble a unicorn and you'll have to figure out how that works with your assumption that the model is improving.

I think we're going to need to wait a few more years, at least, to see any improvement. I expect to see 4 new models a year, before GPT-5 arrives. I'll just keep using the latest model and we can all reconvene in 1, 2, 5, 10 years.

The 'random noise' from a prompt "Draw a unicorn in svg" should still return Unicorns.

This is absolutly fine and it should start showing unicorn like drawings over a longer period and potentially finetuned ones over a longer period of time when the model changes.

>The original post hurts my head with its bad logic.

Huh? What "original post"? This is an experiment, today the model drew something resembling a unicorn. Tomorrow we will see how the experiment goes again. I see no associated analysis, so what makes your "head hurt".

https://adamkdean.co.uk/posts/gpt-unicorn-a-daily-exploratio...

> The idea behind GPT Unicorn is quite simple: every day, GPT-4 will be asked to draw a unicorn in SVG format. This daily interaction with the model will allow us to observe changes in the model over time, as reflected in the output.

Sure, each day it will draw a unicorn. Each time the model changes, we'll have a new group of drawings. No they're not drawn at T=0 but even at T=0 GPT-4 is not deterministic. This isn't science -- this is just a bit of fun.

I believe the logic is fine. You seem to think you need multiple data points from the same version of the model (i.e. multiple samples per day at least) would be necessary to judge the actual performance on each particular day.

That's worse logic. How would you visualize the very large sample you would get? Even with the current 118 samples (one per day) it's already difficult to find a pattern.

Would you "average" the samples?? That would not help IMO, you would need to average the score of each image, which requires either manually doing it or finding a reliable algorithm to do it automatically, but good luck with that.

So, a sample per day which allows clearly visualizing any change in the results over months and years is a valuable thing to do and I find it hard to improve on the methodology. You just need to keep in mind that one single picture from the sample is not enough, no one is going to disagree with that... but that doesn't make it "bad logic" and it's pretty thoughtless to say so.

I agree that trying to determine the distribution of these drawings is hard because it isn't a simple floating point number in its current form.

But maybe you could covert it to a linear monotonic measure?

You could pass it to an image recognition model and see record the degree to which it thinks it is an:

animal horse unicorn

Basically if it fails to be a unicorn, see if it is a horse and if it fails to be a horse check if it is an animal. This gives you some type of linear measure if you place these three measures along the same axis adjacently. Then you can transform each image to a floating point number and then characterize the distribution.

>You could pass it to an image recognition model and see record the degree to which it thinks it is...

We don't care what another algorithm "thinks". We want to see if what it draws is humanly interpretable as a unicorn.

Then you could use mechanical turk to have them rate each image to figure out how close it is to a Unicorn...

Mechanical turk 'workers' use ChatGPT

We could but this is also a fun project, which is why when I checked it today I was surprised that what I saw was not a turd with eyes (2023-05-18) nor a strange sea creature (2023-07-08) but something which, for the first time I think, actually resembled a unicorn.

I appreciate all the comments around determinism, sampling, scientific method, but as I said when I posted this just after building, it really is just for fun and to see, over time, if the general mish mash of outputs become more refined without any changes to the prompt (which doesn't aid it through CoT/ToT or improving on previous attempts etc.)

You don’t have to justify yourself to the HN peanut gallery :-)

Then simply train the model to predict whether a human can interpret it as a unicorn.

What if, instead of your ridiculous strawman, they believe in waiting to get consistency?

I could also believe in not having to have beliefs and just letting it run until I either die or run out of money, and the former may not result in an immediate shuttering of the project either

Expecting a convergent series is bad logic.

I don't see the bad logic.

He said he is "Asking GPT-4 to draw a unicorn every day to track changes in the model."

The variance he is seeing in the output is primarily the product of random chance, rather than changes in the model. Specifically this "unicorn" that he found today is likely just random chance and there was no changes in the model between yesterday and today that lead to it arising.

If he wanted to track changes in the model for real, he would have to ask multiple questions per day and try to infer some type of distribution characterization and then see if that changes over time. That is much more complex and not what he is doing.

This is just a curious experiment that doesn't mean much.

It takes a picture from GPT every day, so we'll be able to see if this was a fluke by looking at future days' outputs. I think it will work to track changes in the model.

Yes, this makes sense. You are agreeing with me. In order to see if it is just a fluke or not, you need a lot of samples so you can characterize the distribution yourself and try to see if it changed.

OK but I'm still not seeing any logic error.

That a single sample per day can track changes meaningfully when the noise floor is above the signal strength. And also the fact that today's unicorn-ish sample is meaningful at all.

It is a fun experiment though.

If you look closely, you can see that the model version is attached to each image, and the original blog post clearly states that asking the same model over multiple days is how they intend to track the variance. No bad logic there imo.

If you look at the examples from April, only 2 or 3 can be counted as unicorns. If in 3 months it's the reverse and only 2 or 3 can't be counted as unicorns, that would show a progressive improvement in the model. I agree we shouldn't take much from day N-1 to day N as there will be a lot of variance, but this can show us progressive improvement over model updates.

Perhaps in a few months half of the generated pictures will look like unicorns, perhaps in more months they will all be unicorns but 2 or 3 will look way more detailed instead of drawn by a 4 year old, etc. We just need to wait longer for the signal to break through the noise.

I do agree with your comment, especially this part:

> We just need to wait longer for the signal to break through the noise.

Currently, what we are observing is primarily noise with very little signal.

you're looking at it from the wrong time scale.

The difference over months or years is what is interesting.

I don't think anyone claims this is an iterative linear measure, rather than a step function.

SVG can present arbitrarily complex graphics. The underlying display tech supports what ever fidelity GPT will eventually mature into.

Has GPT plateaud? Will it be stuck forever at this hilariously naive level of competence at SVG art? Will it mature into Midjourney level competence? I have no frigging clue. Since the token context is so small I imagine it will put limitations to the complexity of SVG art piece.

But I don't know. And it's fun to have a daily measure.

As a software engineer with a penchant for graphics asking GPT to draw complex graphic shapes was one of the first tests I did for it. It's extremely interesting for me to collect progress data, no matter how noisy.

I have no idea if GPT will ever mature beyond these squiggles but if it does, this track record will have at least considerable artistic value, if nothing else.

I mean, how many humans can draw art by writing out svg? If that's not in the training set, I don't even see how GPT-4 gets much better at this over time.

And so, if we see that it /does/ get better, over the next few years, will that not lead us to ask /how/?

Let's think about it:

1. It has to output SVG [1] 2. It is given a text based representation of what it must draw[2] 3. It must then somehow convert words -- the concept of a unicorn: equine with a horn, white, maybe rainbows? -- into SVG code, and attempt to convey both their location, shape, colour, appearance, with code.

And keep in mind, this is just a token predictor. I doubt there is much data in its training that is this specific.

So while it's quite far from science, for me, it's a bit of fun and I get emails every now and then remarking on things like the turd of May (2023-05-18) and it lightens the mood every now and then, which I think ultimately, is worth it.

[1] System: You are a helpful assistant that generates SVG drawings. You respond only with SVG. You do not respond with text.

[2] User: Draw a unicorn in SVG format. Dimensions: 500x500. Respond ONLY with a single SVG string. Do not respond with conversation or codeblocks.

See: https://github.com/adamkdean/gpt-unicorn/blob/master/src/lib...

GPT-4 is really great at transferring concepts between domains.

That's one the reasons why GPT when it works, feels magical. SVG art does not need to be in it's training set, as long as it knows how to present geometric concepts in SVG.

A good unicorn would require capabilities something like "the outline of unicorn is composed of lines {...}." -> "export lines as svg".

Ok that makes perfect sense, thanks.

There's a linked blog post[0] that goes more into the methodology and reasoning.

"As mentioned in the hacker news discussion, the model doesn't change daily. [...] As OpenAI releases incremental updates, we'll see the model change automatically and be able to judge outputs. A single sample per day leads to quite different results, but that's fine I think. What I expect to see a year from now is an evolution of output. In variance: how varied are the outputs over a month?"

So yes, they could just produce 100 images with each new model release, but chose to spread those out over 1 per day instead. Is it the most scientific way to measure progress? No. Is it more fun and interesting to check back daily? Probably.

[0] https://adamkdean.co.uk/posts/gpt-unicorn-a-daily-exploratio...

Thanks, fun and light-hearted is the approach I went for

Reply on news.ycombinator.com