> 3. Good design is as little design as possible.

> 10. Good design is aesthetic.

Unfortunately, these two principles conspire to produce the bland minimalist aesthetic that while some people like[1], leaves me feeling like I'm living in a sterile, clinical, and ultimately soulless environment.

We lament the lack of beauty in architecture and design and here, in this list, is the authoritative voice convincing people to produce more soulless objects. Design is a reflection of the designers interiorality, but what I see when I look at objects designed this way is a lack of interiorality which misses the point of art in it's entirety. It's the human equivalent of AI slop.

1. I understand it appeals as looking "clean" and makes some people's brains feel less itchy.

I think both are reconciled by:

> 5. Products fulfilling a purpose are like tools. They are neither decorative objects nor works of art. Their design should therefore be both neutral and restrained, to leave room for the users self-expression.

Let the decorations be decorations and art be art, while the tools we use to observe or create those other things should be instantly transparent extensions of the user.

I strongly disagree. Anything with an aesthetic is to some degree art. The territory of the world's objects does not strongly conform to a map drawn in black and white.

Can it pre-emptively write the HN comment where someone says it utterly fails for them but no one else is able to reproduce?

Who in the world has it worst? I want to make sure I listen to the right person.

You act like this is an enigma. Let’s start with people without college degrees and no family history of college degrees.

Surely there are people who have it worse than that, no?

It isn't necessary to identify the world's most unfortunate person to recognize that most redditors have privileged lives, yet choose to wallow in misery.

I just choose to not let what redditors do bother me.

Aren't you fairly likely to get a default judgement because they're unlikely to show up to court?

Perhaps! But then you'd get to file a lien on Microsoft, and what a glorious day that would be. Frame it for the office wall!

The point here is that, if we don't exercise the tools available to us to have a human being judge whether Microsoft's behavior is unjust, then we'll never know if a human being would have judged Microsoft's behavior (and terms of service) as unjust. Is LinkedIn summarily destroying one's 'Rolodex' an unlawful act, regardless of whether their boilerplate permits it? One could easily speculate that LinkedIn's terms are a one-sided contract with intolerable terms that exclusively benefit themselves to the harm of the other party, and a judge might well agree if presented such an argument. We'll never know unless people try, though :)

ps. Not particularly relevant, but it came up in research, so:

> Jaffe was asked by defense counsel, "Did you care how he obtained it?" and he answered, "I don't recall thinking that at the time, no sir." Defense counsel then asked, "... all you wanted was the information, and you did not care how it was obtained?" To which Jaffe responded, "Well, I certainly didn't expect him to murder anybody for it." TRM 161. Casper testified that when he handed Jaffe the rolodex file stolen from Wolstencroft's desk Jaffe "was rather elated."

— United States v. Payner, 434 F. Supp. 113 (N.D. Ohio 1977) footnote 34

They’re actually pretty diligent about finding relatively low-cost local lawyers to show up and provide some boilerplate.

Obligatory, if you don't agree with the mayor, it's because you have Bomberman Derangement Syndrome and nothing he does will be good enough for you.

The unfortunate problem is that what you just said is objectively very humourous to me, which I can feel actively improving my sentiment towards the slightly worrying volatility of the whole situation at present.

GP means US casualties. As in the other guys lives don't count.

No I did not. Moreover the bodyguards were Cuban mercenary, not civilian nor Venezuelan military.

My mistake, you're right, Cuban mercenaries don't count.

For a minute it felt like we were gunning for Canada and then chickened out.

They're just clearing their six.

Same goes for a human and doubly so when expressed through text.

If you squint your eyes it's a fixed iteration ODE solver. I'd love to see a generalization on this and the Universal Transformer metioned re-envisioned as flow-matching/optimal transport models.

This makes me think it would be nice to see some kinda child of modern transformer architecture and neural ODEs. There was such interesting work a few years ago on how neural ode/pdes could be seen as a sort of continuous limit of layer depth. Maybe models could learn cool stuff if the embeddings were somehow dynamical model solutions or something.

How would flow matching work? In language we have inputs and outputs but it's not clear what the intermediate points are since it's a discrete space.

One of the core ideas behind LLMs is that language is not a discrete space, but instead a multidimensional vector field where you can easily interpolate as needed. It's one of the reasons LLMs readily make up words that don't exist when translating text for example.

Not the input and output though, which is the important part for flow matching modeling. Unless you're proposing flow matching over the latent space?

The aversion to the unknown is too powerful to allow most people to live in the gray area suggested.

Besides, it's leaving a lot on the table not to leverage this for one's own will to power. Give someone the frame you want and they save the energy otherwise spent finding or creating it themselves. Social construction is too enormous an opportunity to leave up to other people.