My main experience with the demo on the site is that I asked it to add a simple feature which would not even necessarily let you win the game, and it would spend >100k tokens in a loop of "laws broken" until maybe adding it, maybe not. After burning half a million tokens just to add an extra line of walls so that teleporting 2 squares up wouldn't let me win, my API key got rate limited and broke the loop.
On top of that, the solutions feel like patchwork. I asked it to let spacebar flip the board horizontally, and it responded by making the board completely symmetrical including 2 flag poles. At some point it just has to say "this isn't possible without breaking the laws" or think of an actual workaround, because if I was making a game, suddenly having 2 finishes would be unwanted behavior for me.
I think the demo is this way simply because looking at the LLM find wacky ways of implementing features without breaking the law is fun and drives the point across. In practice I imagine you would write something like "If you don't see a clear way of implementing a feature without breaking the law, ask me for directions" in AGENTS.md
Why was the aesthetic standard to be pale when workers worked the fields and royals were inside, but tan when workers moved into factories and only the rich could afford to go on a beach vacation?
Aesthetic standards are formed by association. Its why sites that are "well designed" but obviously just use a squarespace or wix template feel so cheap. Why millenial flannel went from hip to standard to outdated. Why purple was the color of royalty before we could synthesize the pigment.
Having good design is about associations. Whatever design LLMs will default to, it will always feel cheap because we will learn over time that that design means cheap. Having good taste is about being ahead of the curve. An LLM cant be ahead of the curve because then that becomes the standard, and theres a new ahead.
You can use LLMs to make novel looking websites by carefully telling it to add certain details, use certain elementd, etc. At that point youve looped back to being a graphic designer.
I think what you’re talking about is real, but it’s only part of the problem. The issue is it’s poor design. There’s a lack of consistency that is really off putting. Spacing is inconsistent and doesn’t create a sense of visual hierarchy. Buttons, inputs, selects, call-outs, table cells are barely distinguishable from each other, but also inconsistent within their own categories. The copy is also confusing. I don’t even know what this does.
This is, once again, about diversity and the lack thereof (and I don't mean diversity in a political sense).
LLMs seem fundamentally incapable of producing truly diverse outputs, truly creative and different responses to the same prompts in different runs. Because you and me use the same Claude, if you want a website and I want a website, we'll get (almost) the same website. This is not some BS about "the average of its training data", most of the LLM style (both in design and in text) comes from reinforcement learning. You could RL Claude to produce a very different style, but you couldn't RL it to produce a different style for me than it does for you.
I think this is also where a lot of the complaints about "Claude writing" come from.
RL causes distributional collapse, it's how the models get consistent. Anyone who generated images with early gen (SD1.5-2) models will remember the wild variance between seeds, which newer models have mostly lost, and similarly GPT3.5/4 could produce weirder, more original outputs even if they were less consistently "good" in some sense.
It's worth mentioning that they do RL for aesthetics to some degree based on human expert feedback, but whatever the model tends to produce quickly becomes debased by its ubiquity. They could RL for output diversity, but it's less well studied and likely to cause minor regressions in coding performance, at least until the algorithms are dialed in.
I don't know, I think you have a point that aesthetic preferences are subjective and shifting. But there is also all-caps monospaced text with emdashes in it on the site; just an example of something I think would not turn into a fashion at any point because it just looks silly (subjectively, to me at least). Thus I don't think the antipathy of people towards these llm-generated landing pages is entirely based on associating it with other LLM sites, there is at least an element of it clearly not being through as much human review and interaction as a hand-crafted landing page necessarily would be.
This is a problem that i'm actively working on (https://fudge.design), what i've realised is that it's simply not an issue of capability - given a well crafted site and a competently written visually aware harness, most recent models can replicate that website.
So it's what lies between saying "I want x website" -[.....] -> Code+Assets
The issue has to do with specification fidelity, in short a grill-me style aesthetic interrogation using illustrative tooling - ascii diagrams for specifying layout, copy and user-flow, image-gen mockups for higher fidelity mockups. References are also very important for nailing down the aesthetical qualities. I've noticed it's far better vs purely text description to simply gather up a mood-board telling the llm to find commonalities and come up with a design system and brand guide.
So I don't believe it's an unsolvable problem, it's simply a lack of effort on the implementors part. Also there's probably some survivor's bias here (you won't notice an intentionally designed vibe-coded site)
I really want to create a nosology of common generic memes that can be applied to literally anything without context. Saying that the evaluators are just stupid and arbitrarily chasing the fashion of the week instead of the evaluators possibly actually latching onto some structure is a tale as old as time.
My touchpad doesn't work anymore. Why is it so difficult to connect a bluetooth mouse without a mouse? The "mouse detected, want to connect?" notification shows up ON THE SCREEN but I need to enter a bunch of shortcuts and navigate through multiple items just to be able to click yes.
My touchpad broke and I couldnt download the touchpad driver to fix it on the spot because there was no way to get to the download button with tab and enter, you had to click it.
That's a funny example, reminds me of how often fresh Windows 7 or older installs were missing the ethernet driver so you had to move stuff over on a flash drive to get started really setting stuff up.
In your situation, a couple possible solutions come to mind. 1) turn on "mouse keys" which lets you use the numpad to move the mouse and click, 2) a vim addon for the browser (e.g. vimperator or tridactyl) likely could've clicked the button via the hints mode.
It's in the past now so maybe these wouldn't have worked or you didn't know, of course, just fun to think of what I would've done.
Autism is a disability. The function of the brain to normally process social cues is broken. In the same way someone who is blind has broken vision, someone in a wheelchair has broken legs, and someone with ADHD has broken executive function.
This isnt a moral judgement, theres nothing wrong with being disabled, but it remaims a disability that many people struggle due to.
The disability exists when compared to neurotypicals but does not necessarily exist when compared to other autists.
As a comparison, if you put a bonobo with chimps the bonobo will be severely disabled socially, but it won't be when compared to other bonobos.
The fact that autists can communicate with each other does not fit typical models of disability. A blind person should be blind to another blind person.
There is interesting evidence that autism-autism social interactions are fairly vibrant. Everything is a bit more complicated than simple answers, no matter which way you swing.
You've agreed with what I said. Nothing here is counter to or in opposition to my point.
The person has a disability. Part of the person may be broken, but the person isn't broken. I'm autistic. My sensory processing wiring is different and arguably broken. But I am not. I'm disabled. I can overcome the broken parts that result in my disability with support.
I mean, it's classified as a disability for a good reason.
I would guess that even if this platform has more autistic people due to being tech focused, autistic tech nerds and programmers are on the extreme end of high-functioning. Many low-functioning autistic people will never be able to have a job or live by themselves.
This is the same issue as high-functioning ADHD'ers saying "ADHD is my superpower", meanwhile people with severe ADHD do significantly worse than the general population in basically any metric, up to having a lower life expectancy.
This isnt to say its morally bad, theres nothing wrong with being autistic or having ADHD, but it is a disability. Just like someone who's deaf can live a totally happy life but still be disabled because they are not able to hear, someone with autism is disabled because they are not able to process social cues at the same speed, even if they are a great person and have a good life.
"Have a disability" and "broken" are not synonyms.
There's no judgement or moralizing in having a disability. Autism and ADHD like you mentioned are disabilities.
But the people who have them are not broken. Having a disability means some part of your life is more difficult. But that's very very different from you as a person are broken.
For example, if you can fix the environment and the person can now function equally to someone without a disability, then maybe it's the environment that's broken. If you can fix the social expectation and suddenly the person can function equally, then maybe the society is broken. Even if you cannot accommodate the person's needs, insinuating they are broken is not a great framing -- it creates a sense of pity rather than constructively looking at means to support them.
That's just people though, if you walk around with a scar on your head you'll see lots of knee-jerk immediate reactions. Some people are able to mediate those initial reactions and be kind, others just go with it and treat people poorly. Yes - some people are a'holes, and treat those without legs poorly.
The same is true of basically anyone that doesn't fit into what you typically see - there's always some people that will treat them poorly. It's unfortunate, but was probably adaptive during our evolution, and something we just have to deal with today.
> If you see a person without legs, do you tell him he is a piece of shit because he can't run? So why harass autistic people?
If you have not yet realized it, here is a black pill for you: most people are deeply manipulative and bullying basically all the time. For example nearly all social rules that are commonly accepted in society are basically about which kinds of manipulations and bullying of other people are accepted by most people vs not.
With this in mind, the answer to your question
> So why harass autistic people?
is trivial: because it is currently socially mostly accepted.
> If you have not yet realized it, here is a black pill for you: most people are deeply manipulative and bullying basically all the time. For example nearly all social rules that are commonly accepted in society are basically about which kinds of manipulations and bullying of other people are accepted by most people vs not.
You went from "Neurotypical Peers are Less Willing to Interact with Those with Autism based on Thin Slice Judgments" to "refuse to interact with autists".
> For example nearly all social rules that are commonly accepted in society are basically about which kinds of manipulations and bullying of other people are accepted by most people vs not.
> Have a disability" and "broken" are not synonyms
They really are. Like, you can alter the environment all you want but someone who needs glasses to see has broken eyes. Is it super common and do we have a cure? Yes and yes, but it's still, you know, not working. Disabled. Busted. Broken. Whichever word you pick isn't going to change the reality.
"That person is disabled" and "that person is broken" fundamentally are different. You are supposed to fix broken things. We don't fix disabilities. We can support them, treat them, assist them.
Yes, having a disability is hard. But the person is not broken.
This is also why apps like Duolingo don't work for most users. They want to turn language learning into a fun, effortless game but effortless is fundamentally incompatible with learning.
Learning is uncomfortable. Reading a difficult (for you) text in a language you don't understand is exhausting and confusing. But that is where improvement happens.
Its even worse since Duolingo added a life system (not sure if they still use it), where you were only allowed to make 3 mistakes before having to recharge your energy. If you get everything right, you are not learning, you SHOULD be making mistakes constantly. That shows you're actually being challenged.
It is crazy how much of the tech industry - one of the biggest industries on Earth - still uses so much volunteer labor. Its crazy because in no other industry would people expect you to not only give away your stuff for free but do it in such a way anyone else can copy your stuff and edit it.
Too much self-awareness can cause you to become blind to the experiential part of life.
Like, when you're at a party, not having fun, but instead thinking: "I am drinking alcohol to lower my social inhibitions at the cost of worse sleep and long-term health benefits, because I think in the long run the possible connections made here and the subsequent rise in social status will be worth it"
Resembled with above experience. Wanting to be self aware, I spent couple of years in confusion. Rather only thinking about what I am doing at any point of time, and trying to be aware, not being involved fully and not able to enjoy. spent few years like that, it is really difficult situation to be in. At some point, I realized, the self-awareness is towards emotions that are arising in myself, mainly, not on particular thoughts. means while speaking with others, if there is any strong emotion particular negative emotion raising in oneself. just being aware about that strong emotion actually cools down, and then it becomes a habit. and over a period of time along with emotion awareness also comes. and then we can still continue to enjoy the whole party.
That's not self-awareness. You have it all backwards.
Awareness is about the senses, as well as having the experience to predict what will happen next, and plan accordingly. Awareness is 100% about balancing the sensing and the planning.
The sensing is important, but not for its own sake. If you overcorrect in that direction, it's just self-indulgence and hedonism.
It's interesting how this compulsion to work spread even to people that don't need to do it. Just look at how many modern-day billionaires are still flaunting how many hours they work each week. Up to the modern era, work was not seen as an inherent virtue, and even something to be avoided. Even well into the modern era, Teddy Roosevelt went on multi day camping trips completely unconnected from civilization. Just imagine a world leader wanting to do that now.
Roman emperors routinely issued statements about how they were working so hard on behalf of their subjects that they barely had time to sleep. Evaluating the veracity of those statements is left as an exercise to the reader.
On top of that, the solutions feel like patchwork. I asked it to let spacebar flip the board horizontally, and it responded by making the board completely symmetrical including 2 flag poles. At some point it just has to say "this isn't possible without breaking the laws" or think of an actual workaround, because if I was making a game, suddenly having 2 finishes would be unwanted behavior for me.
reply