MPSimmons

<3

It was inevitable, though you always hope for an exception. I don't think I've ever seen any work of his I didn't deeply appreciate.

I'm afraid that technological sovereignty is secondary to military superiority when the governments in question value might over justice. It seems that you can apply financial leverage to avoid violence but I think everyone has to play ball.

Agreed, this is the kind of steganography that is meant for humans to read. There's no reason a computer couldn't encode meaningful information in the least significant bytes of every 16th pixel or something like that.

interesting, wonder which class promulgated that myth

sshd

Modern consoles also compile shaders? Ugh, that's the most annoying part of playing RocketLeague right now. The game updates like every day and every time it's updated I need to spend 5ish minutes recompiling (if not more).

Seems like a ToS violation more than anything

Oh shit, this looks incredible. Thanks!

I don't really use plan mode, but I do have Fable write a lot of docs, particularly when the work involved is long and multistep, because I may need to do it across multiple sessions.

Sorry, I think I'm being dumb, but where is the "first principals thinking" in your example?

Yeah, I still remember the first time I ran glxgears on accelerated hardware and I could not believe my eyes

Couldn't you just make a CLAUDE.md that says, "Read AGENTS.md in this same directory"?

I don't understand why the system would evolve to this state if it didn't impact reproduction.

I still remember the day, circa 1997, that I found an anonymous-enabled FTP site on a uunet domain that had been used as a drop site by various couriers, and let me tell you, this high schooler's 540MB hard drive was not large enough.

It seems intellectually dishonest to blanket-label any principled stance that you disagree with as "virtue signaling".

That was my first thought too, but it seems like they're using two of them in the wireless version, with a C5 for wifi

> The asset of being able to quickly understand and produce a solution appropriate for a large technical system is now _significantly_ less scarce for businesses.

Certainly, it's easier to produce solutions now, but I wouldn't underestimate the knowledge you have, and say that others who are capable of using AI to produce the same output have the same understanding that you do.

Personally, I find that I am able to produce excellent outcomes with AI with respect to the topics that I have the most expertise in, and passingly acceptable outcomes in things that I know very little about.

IMO there's more than one valid reason to make things.

Some people make things from a pragmatic standpoint - they need a solution in a space where the existing solutions don't fit for some reason, and they decide to make their own. This is, I think, the majority of generative AI coding. A solution needs to be filled, and even if you are an amazing coder, for most of the solution space, AI is going to be faster than you. Your code might be more elegant, but not every problem requires elegance. Sometimes, it just needs to work.

Other people do it for the love of the game. The enjoy solving problems, crafting code, and finding optimum and elegant solutions to problems. This isn't pragmatism, or even close. It's craftwork by masters, or apprentices who want to be masters, and they wouldn't use AI to produce the work any more than they'd ask another human to do it - that would be cross purposes to their goal, which is to do the making.

Both are valid and there should be no shame in either, but I think people confuse the two and the disparity between them causes friction. When you judge a person's work in one paradigm by the other, you're bound to offend and distress them.

oh wow, I'd never heard of this. Reading https://en.wikipedia.org/wiki/Flicker_fusion_threshold - now I think I understand why I can always tell GM vehicles with LED taillights at night. I can spot them a mile away, literally, because their taillights blink slower than others and I can see it, and it drives me crazy.

Thanks for opening me up to this!

A lifespan of 2 years for intelligent creatures is heartbreaking

I assume someone has considered whether the apparent expansion of the universe is spaghettification as we enter an event horizon?

I miss the time when smaller was considered more advanced

FWIW (and this is not directed to you specifically but anyone reading this) the police can use biometrics to compel you to open your phone without a warrant, but they cannot force you to give up your passcode.

https://www.zdnet.com/article/biometrics-vs-passcodes-what-l...

Why do they insist on every new phone being bigger than the phones before them? I don't mind folding. I think it's cool, and a nice way to get a useful space. But I have big hands, and I can't use my iphone 17 now without two hands. This is even wider before unfolding.

Who wants bigger phones? Who is like, "damn I wish I had a tablet that I could fold up and still barely fit in my pocket"?

I want smaller phones. The width of the iPhone 4 was nearly perfect imo. If they would take modern display technology with approximately that width and expand it to 16:9, I think it would kill, even if it were relatively underpowered compared to the modern top of the line.

Why do they insist on every new phone being bigger than the phones before them?

I don't mind folding. I think it's cool, and a nice way to get a useful space. But I have big hands, and I can't use my iphone 17 now without two hands. This is even wider before unfolding.

Who wants bigger phones? Who is like, "damn I wish I had a tablet that I could fold up and still barely fit in my pocket"?

I want smaller phones. The width of the iPhone 4 was nearly perfect imo. If they would take modern display technology with approximately that width and expand it to 16:9, I think it would kill, even if it were relatively underpowered compared to the modern top of the line.

Awesome comment, thanks for the info!

Weather 2 doesn't seem to have been an ensemble model. Weather 3 is, so theoretically it can get more accurate outcomes by taking the probabilistic analysis of several models concurrently to determine the most likely weather conditions.

I'm building a tool right now that uses an ensemble to determine wind gust likelihood, which is useful for safety critical work on construction sites and the like.

Do you have the cheap models summarize the files? How do they get the relevant information to the bigger models?

You can also just grow potatoes in a cardboard box - https://www.youtube.com/watch?v=HyW5QciFAhc

Out of curiosity, what do you do to distribute the weights and inference across the multiple machines?