To Steelman the topic, Musk’s whole alleged mission is to make humans a multi-planet species that can survive an earth killing event.

To that end, a small data center space isn’t about unit-economics, it’s a bigger mission. So the question we should consider is what can we put into space the further that mission. Can we put a meaningful sum of human knowledge out there for preservation? It sounds like “yes,” even if we can’t train ChatGPT models out there yet.

They are managed by different standards organizations. One doesn't like the other encroaching on its turf.

It’s not them denying it, it’s the LLM that generated this slop.

All they had to say was that the KiB et. al. were introduced in 1998, and the adoption has been slow.

And not “but a kilobyte can be 1000,” as if it’s an effort issue.

I am not anti-LLM by almost any stretch but your lack of fundamental understanding coupled with willingness to assert BS is at the point where it’s impossible to discuss anything.

you have an outdated view on how much it hallucinates.

We need a new word, not "local model" but "my own computers model" CapEx based

This distinction is important because some "we support local model" tools have things like ollama orchestration or use the llama.cpp libraries to connect to models on the same physical machine.

That's not my definition of local. Mine is "local network". so call it the "LAN model" until we come up with something better. "Self-host" exists but this usually means more "open-weights" as opposed to clamping the performance of the model.

It should be defined as ~sub-$10k, using Steve Jobs megapenny unit.

Essentially classify things as how many megapennies of spend a machine is that won't OOM on it.

That's what I mean when I say local: running inference for 'free' somewhere on hardware I control that's at most single digit thousands of dollars. And if I was feeling fancy, could potentially fine-tune on the days scale.

A modern 5090 build-out with a threadripper, nvme, 256GB RAM, this will run you about 10k +/- 1k. The MLX route is about $6000 out the door after tax (m3-ultra 60 core with 256GB).

Lastly it's not just "number of parameters". Not all 32B Q4_K_M models load at the same rate or use the same amount of memory. The internal architecture matters and the active parameter count + quantization is becoming a poorer approximation given the SOTA innovations.

What might be needed is some standardized eval benchmark against standardized hardware classes with basic real world tasks like toolcalling, code generation, and document procesing. There's plenty of "good enough" models out there for a large category of every day tasks, now I want to find out what runs the best

Take a gen6 thinkpad P14s/macbook pro and a 5090/mac studio, run the benchmark and then we can say something like "time-to-first-token/token-per-second/memory-used/total-time-of-test" and rate this as independent from how accurate the model was.

Next year: LeGit !

Gun nut Eric Raymond was cheering when the first printable guns came out. Checkmate gun grabbers, you'll never prevent us from having our shooty-shootys now! Haha! I thought, well the answer to that is simple: simply declare 3D printers to be weapons. You know, like how the Feds declared encryption to be "munitions".

I like how the GNU coreutils seem to have done. They use real, 1024-byte kilobytes by default, but print only the abbreviation of the prefix so it's just 10K or 200M and people can pretend it stands for some other silly word if they want.

You can use `--si` for fake, 1000-byte kilobytes - trying it it seems weird that these are reported with a lowercase 'k' but 'M' and so on remain uppercase.

> Sadly, instead of having better laws we get fallacy rhetoric by people who probably have never touched, much less fired a gun in their lives.

Why is this the litmus test for being qualified to write gun legislation? Do we also expect our lawmakers to have tried heroin or downloaded child porn so that they can regulate those activities?

A full rebuild might be throwing out the baby with the bath water. As someone who’s been using it since it was known as Project Builder, bugs seem mostly concentrated in the XIB/Storyboard editor (formerly known as a Interface Builder), SwiftUI live preview, and SwiftPM package resolve.

In a project with code-only UIKit, only a smattering of SwiftUI for small components, and minimal dependencies, Xcode isn’t too bad of an experience and I’d say comparable to and in some ways better than Android Studio (that localization XML editor, not mention Gradle… ugh).

This is why I come to this site. Obviously, Twitter's financials are struggling and theres more than a few people rich people who don't want to take the hit... but we can all drop that for a second to discuss the plausibility of data centers in space. Some links and comments I enjoyed:

  * https://taranis.ie/datacenters-in-space-are-a-terrible-horrible-no-good-idea/
  * https://en.wikipedia.org/wiki/Tiangong_space_station
  * "Technically challenging", a nice way to say "impossible"
  * "I’m not that smart, but if I were, I would be thinking this is an extended way to move the losses from the Twitter purchase on to the public markets."
  * "ISS radiators run on water and ammonia. Think about how much a kg costs to lift to space and you'll see the economics of space data centers fall apart real fast. Plus, if the radiator springs a leak the satellite is scrap."
  * "5,000 Starship launches to match the solar/heat budget of the 10GW "Stargate" OpenAI datacenter. The Falcon 9 family has achieved over 600 launches." [nerdsniper]
  * "No, we just "assume" (i.e. know) that radiation in a vacuum is a really bad way of dissipating heat, to the point that we use vacuum as a very effective insulator on earth."
  * "World's Best At Surfing A Temporary Hyperinflation Wave is not a life goal to really be proud of tbh"

This has nothing to do with the 1024, it has todo with the 1200 and the multiples of it and the 14k and 28k modems where everyone just cut off the last some hundred bytes because you never reached that speed anyway.

There's a good reason that gigabit ethernet is 1000MBit/s and that's because it was defined in decimal from the start. We had 1MBit/s, then 10MBit/s, then 100MBit/s then 1000MBit/s and now 10Gbit/s.

Interestingly, from 10GBit/s, we now also have binary divisions, so 5GBit/s and 2.5GBit/s.

Even at slower speeds, these were traditionally always decimal based - we call it 50bps, 100bps, 150bps, 300bps, 1200bps, 2400bps, 9600bps, 19200bps and then we had the odd one out - 56k (actually 57600bps) where the k means 1024 (approximately), and the first and last common speed to use base 2 kilo. Once you get into MBps it's back to decimal.

> If we beat the Chinese somehow, I don't think they'll just dismantle their space program and focus on Earth.

The Soviet Union won the "space race" of course (or perhaps Germany did if you define it as suborbital space flight), it just lost the "man on the moon race". In any case, after losing the man on the moon race, the Soviet Union did not just dismantle their space program and focus on Earth. They continued to invest a great deal in their civil, scientific, and military space capabilities after 1969.

Will the Chinese Communist Party similarly collapse in the 2050s? Perhaps not, but they will be going through significant demographic decline from the 2030s; they are increasingly in conflict with the west and with their territorial neighbors; they may become involved in significant military conflicts (e.g., over Taiwan); their current leader has consolidated power and succession could be spicy. So who knows? It's not inconceivable. China would surely continue and continue a space program as Russia has.

I built an entire iOS app without opening Xcode UI even once. Why so many iOS engineers prefer XCode?

Upper-K is for Kelvin, so can't be mixed in as a prefix in case someone decides to commit physics crimes and talk about temperature-mass (Kkg).

I'm convinced that >30% of this comes from ideas leaking out of fiction such as like Neuromancer, and percolating through the minds of wealthy people attracted to some of the concepts. Namely, the dream of being a hyper-wealthy dynasty, above any earthly government, controlling an extraterritorial Las Vegas Fiefdom In Space. (Which in the book, also hosted a powerful AI.)

Then they work backwards, trying to figure out some economic engine to make it happen. "Data centers" are (A) in-vogue for investment right now and (B) vaguely plausible, at least compared to having a space-casino.

fighter jets ARE a threat of violence, and it is widely understood and acknowledged.

Again: the threat is so clear that you rarely have to execute on it.

I had a university friend who spent hundreds of hours on his YouTube channel whilst the rest of spent hundreds of hours arguing about politics.

He’s now unimaginably successful at YouTube but at least I’m better at predicting the content of tomorrow’s newspapers.

> As someone who generally stays out of politics, I didn’t know much about the incoming administration’s stance towards tariffs, though I don’t think anyone could have predicted such drastic hikes.

I have an appreciation for very bright lamps, and the project is neat, but that stuck out to me.

I'm always fascinated by people who both feel comfortable ignoring maybe the single most impactful society-determining apparatus but will also say "no one could have seen that coming", where that is whatever they were unaware of because they chose to check out. I find the stance so fascinating because for myself, it would be impossible to not try and understand why the world is the way it is.

Everything is downstream of politics whether people want to recognize that or not, and choosing to ignore it is, in fact, a political choice.

Why, why, why, and whatabout?

Most likely because you are not in charge of all of the money, oil, data, and political power.

> The New York definitions sweep in not just FDM and resin printers, but also CNC mills and “any machine capable of making three-dimensional modifications to an object from a digital design file using subtractive manufacturing.”

...what? This some of the stupidest, most out of touch garbage I've ever read and clearly made by uneducated lawmakers being out of their depth.

Your AI can’t make a UI for 3 platforms? Seems pretty worthless.

let me guess... you don't follow politics either...

"la forge" in french is feminine, it would need to be LaGitette

If we won't stop what he is doing with grok and ai-generated CSAM, he will be completely free from oversight up there.

> And finally surface area, once again, getting quite good here with nanotechnology.

So your hot thing is radiating directly onto the next hot thing over, the one that also needs to cool down?

Realistically, everyone always seems to think everything was predictable but I have maybe a handful of friends who sold the Russell 2000 futures short and then rebounded long who made millions off the various tariff trades. Ironically, Ukrainian and Russian. Ex-HFT but just doing very normal click trading. So I don't get it. Why isn't everyone who can predict the future so accurately a (deca-)millionaire?

Is this bait? XCode has been a mainstay of iOS development ever since iOS was introduced and is a successor to Interface Builder on the Mac.

Why wouldn’t engineers prefer tools they’ve been using (mostly happily) for a decade+?