I routinely check out the pelicans you post and I do agree, this is the best yet. It seemed to me that the wings/arms were such a big hangup for these generators.

How likely this problem is already on the training set by now?

If anyone trains a model on https://simonwillison.net/tags/pelican-riding-a-bicycle/ they're going to get some VERY weird looking pelicans.

Why would they train on that? Why not just hire someone to make a few examples.

I look forward to them trying. I'll know when the pelican riding a bicycle is good but the ocelot riding a skateboard sucks.

But they could just train on an assortment of animals and vehicles. It's the kind of relatively narrow domain where NNs could reasonably interpolate.

The idea that an AI lab would pay a small army of human artists to create training data for $animal on $transport just to cheat on my stupid benchmark delights me.

When you're spending trillions on capex, paying a couple of people to make some doodles in SVGs would not be a big expense.

The embarrassment of getting caught doing that would be expensive.

They were caught using all the data on the internet without asking for permission or compensating anyone. And it has cost them nothing and earned them billions so far.

Vetting them for the potential for whistleblowing might be a bit more involved. But conspiracy theories have an advantage because the lack of evidence is evidence for the theory.

Huh? AI labs are routinely spending millions to billions to various 3rd party contractors specializing in creating/labeling/verifying specialized content for pre/post-training.

This would just be one more checkbox buried in hundreds of pages of requests, and compared to plenty of other ethical grey areas like copyright laundering with actual legal implications, leaking that someone was asked to create a few dozen pelican images seems like it would be at the very bottom of the list of reputational risks.

How do you think who's in on that? Not only pelicans, I mean, the whole thing. CEOs, top researchers, select mathematicians, congressmen? Does China participate in maintaining the bubble?

I, myself, prefer the universal approximation theorem and empirical finding that stochastic gradient descent is good enough (and "no 'magic' in the brain", of course).

Well, since we're all talking about sourcing training material to "benchmaxx" for social proof, and not litigating the whole "AI bubble" debate, just the entire cottage industry of data curation firms:

https://scale.com/data-engine

https://www.appen.com/llm-training-data

https://www.cogitotech.com/generative-ai/

https://www.telusdigital.com/solutions/data-for-ai-training/...

https://www.nexdata.ai/industries/generative-ai

---

P.S. Google Comms would have been consulted re putting a pelican in the I/O keynote :-)

https://x.com/simonw/status/1924909405906338033

Cool. At least they are working across the board and benchmaxing random things like the theory of mind.

I think no matter what happens with AI in the future, there will always be a subset of people with elaborate conspiracies about how it's all fake/a hoax.

I'm not saying it's a hoax. If it gets better because of that data, tant mieux, but we have to be clear eyed about what these models are actually doing. Especially when companies don't explain what they've done.

Would it not be better to have 100 such tests "Pelican on bicycle", "Tiger on stilts"..., and generate them all for every new model but only release a new one each time. That way you could show progression across all models, attempts at benchmaxxing would be more obvious.

Given the crazy money and vying for supremacy among AI companies right now it does seem naive to belive that no attempt at better pelicans on bicycles is being made. You can argue "but I will know because of the quality of ocelots on skateboards" but without a back catalog of ocelots on skateboards to publish its one datapoint and leaves the AI companies with too much plausible deniability.

The pelicans-on-bicycles is a bit of fun for you (and us!) but it has become a measure of the quality of models so its serious business for them.

There is an assymetry of incentives and high risk you are being their useful idiot. Sorry to be blunt.

Or indeed do the Markov chain conceptual slip. Pelican on bicycle, badger on stool, tiger on acid. Pelican on bicycle is definitely cooked, though: people know it and it's talked about in language.

For every combination of animal and vehicle? Very unlikely.

The beauty of this benchmark is that it takes all of two seconds to come up with your own unique one. A seahorse on a unicycle. A platypus flying a glider. A man’o’war piloting a Portuguese man of war. Whatever you want.

No, not every combination. The question is about the specific combination of a pelican on a bicycle. It might be easy to come up with another test, but we're looking at the results from a particular one here.

More likely you would just train for emitting svg for some description of a scene and create training data from raster images.

None of this works if the testers are collaborating with the trainers. The tests ostensibly need to be arms-length from the training. If the trainers ever start over-fitting to the test, the tester would come up with some new test secretly.

You can easily make a RLAIF loop.

- Take a list of n animals * m vehicule

- Ask a LLM to generate SVG for this n*m options

- Generate png from the svg

- Ask a Model with vision to grade the result

- Change your weight accordingly

No need to human to draw the dataset, no need of human to evaluate.

You can always ask for a tyrannosaurus driving a tank.

I've heard it posited that the reason the frontier companies are frontier is because they have custom data and evals. This is what I would do too

Is there a list of these for each model, that you've catalogued somewhere?

At the moment that's mostly my tag page here but I really need to formalize it: https://simonwillison.net/tags/pelican-riding-a-bicycle/

The reflection of the sun in the water is completely wrong. LLMs are still useless. (/s)

It's not actually, look up some photos of the sun setting over the ocean. Here's an example:

https://stockcake.com/i/sunset-over-ocean_1317824_81961

That’s only if the sun is above the horizon entirely.

No, it's not.

https://stockcake.com/i/serene-ocean-sunset_1152191_440307

Yes, it is. In that photo the sun is clearly above the horizon, the bottom half is just obscured by clouds.

It's worth noting that you mean excellent in terms of prior AI output. I'm pretty sure this wouldn't be considered excellent from a "human made art" perspective. In other words, it's still got a ways to go!

Edit: someone needs to explain why this comment is getting downvoted, because I don't understand. Did someone's ego get hurt, or what?

It depends, if you meant from a human coding an SVG "manually" the same way, I'd still say this is excellent (minus the reflection issue). If you meant a human using a proper vector editor, then yeah.

maybe you're a pro vector artist but I couldn't create such a cool one myself in illustrator tbh

Indeed. And when you factor in the amount invested... yeah it looks less impressive. The question is how much more money needs to be invested to get this thing closer to reality? And not just in this instance. But for any instance e.g. a seahorse on a bike.

Do you have to still keep trying to bang on about this relentlessly?

It was sort of humorous for the maybe first 2 iterations, now it's tacky, cheesy, and just relentless self-promotion.

Again, like I said before, it's also a terrible benchmark.

I'll agree to disagree. In any thread about a new model, I personally expect the pelican comment to be out there. It's informative, ritualistic and frankly fun. Your comment however, is a little harsh. Why mad?

It being a terrible benchmark is the bit.

Eh, i find it more of a not very informative but lighthearted commentary

It's HN's Carthago delenda est moment.

Highly disagree.

I was expecting something more realistic... the true test of what you are doing is how representative is the thing in relation to the real world. E.g. does the pelican look like a pelican as it exists in reality? This cartoon stuff is cute but doesnt pass muster in my view.

If it doesn't relate to the real world, then it most likely will have no real effect on the real economy. Pure and simple.

I disagree. The task asks for an SVG; which is a vector format associated with line drawings, clipart and cartoons. I think it's good that models are picking up on that context.

In contrast, the only "realistic" SVGs I've seen are created using tools like potrace, and look terrible.

I also think the prompt itself, of a pelican on bicycle, is unrealistic and cartoonish; so making a cartoon is a good way to solve the task.

The request is for an SVG, generally _not_ the format for photorealistic images. If you want to start your own benchmark, feel free to ask for a photorealistic JPEG or PNG of a pelican riding a bicycle. Could be interesting to compare and contrast, honestly.

Tbh they'd have to be absolutely useless at benchmarkmaxxing if they didn't include your pelican riding a bicycle...

We've reached PGI

Wow. I wonder how it would do with pure CSS a la https://diana-adrianne.com/

So, you've said multiple times in the past that you're not concerned about AI labs training for this specific test because if they did, it would be so obviously incongruous that you'd easily spot the manipulation and call them out.

Which tbh has never really sat right with me, seemingly placing way too much confidence in your ability to differentiate organic vs. manipulated output in a way I don't think any human could be expected to.

To me, this example is an extremely neat and professional SVG and so far ahead it almost seems too good to be true. But like with every previous model, you don't seem to have the slightest amount of skepticism in your review. I don't know if I truly believe Google cheated here, but it's so good it does therefore make me question whether there could ever be an example of a pelican SVG in the future that might actually trigger your BS detector?

I know you say it's just a fun/dumb benchmark that's not super important, but you're easily in the top 3 most well known AI "influencers" whose opinion/reviews about model releases carry a lot of weight, providing a lot of incentive with trillions of dollars flying around. Are you still not at all concerned by the amount of attention this benchmark receives now/your risk of unwittingly being manipulated?

The other SVGs I tried from my private collection of prompts were all similarly impressive.

Is there a way you can showcase a few of these?

Not without people later saying "you shared that on Hacker News last year clearly the AI labs are training for it now!"

Couldn't you just make up new combinations, or new caveats indefinitely to mitigate that? It would be nice to see maybe 3-4 good examples for validation. I'd do it myself, but I don't have $200 to play around with this model.

Here's what it gave me for a kakapo on a skateboard https://gist.github.com/simonw/5e2041c32333effd090e3df42b64d...

Thank you!

This benchmark outcome is actually really impressive given the difficulty of this task. It shows that this particular model manages to "think" coherently and maintain useful information in its context for what has to be an insane overall amount of tokens, likely across parallel "thinking" chains. Likely also has access to SVG-rendering tools and can "see" and iterate on the result via multimodal input.

>"The pelican riding a bicycle is excellent. I think it's the best I've seen. https://simonwillison.net/2026/Feb/12/gemini-3-deep-think/"

Yeah this is nuts. First real step-change we've seen since Claude 3.5 in '24.