I did a course in photolithography, including a lab session where I literally made an unpackaged ASIC by hand, using technology that was current in the mid 90s. Etching silicon and doing photolithography is completely out of the reach for a hobbyist, even from scratch, even with 30 year old technology. Reusing a random chip and adding to it is simply impossible.

Essentially, you are looking at the following problems with simply creating a new chip:

1. Need access to a large clean room
2. Need a furnace that can heat up your silicon wafers to silly temperatures in order to do doping via sputtering.
3. Need chemical handling equipment, including safety gear required for handling hydrofluoric acid.
4. Need a photolithography device with your designs.
5. Need to cut, test, and package chips after they are created.

And the additional problems with using an existing chip:

1. Need to decapitate the existing chip and remove it from the package.
2. Need to work with a single already sliced chip instead of a wafer.
3. A completed chip will have metal vias that cover a large portion of the silicon, you would need to perform your chemical etching without harming these or any silicon that has been used.
4. Generally, Asics don’t have wasted space, but simply larger components. There would be no footprint available for your design, even if your hobbyist tooling was smaller than the one used for the components, which is probably wouldn’t be.

Quest 64 exists despite Quest 63, Quest 62, … Quest not existing.

Not to be contrarian, just running a Matrix server is awful, and Matrix has been plagued with security issues in the past, including basic crypto gaps due to lack of domain understanding in the implementation, which is shocking for what is touted as a security first project.

Moreover, I don’t know of any acceptable alternatives.

I’m generally one of the first people on any FOSS bandwagon. I’ve been using Linux as my daily driver since 1999. Matrix is not simply “not as good” or “not up to feature parity” as alternatives. It is in my opinion unacceptably bad and the project leads seem to be actively hostile to efforts to make it better.

Every simple numeric metric you use to measure a programmers productivity will be gamed and defeated. This was a lesson learned in the fucking 90s. Amazon should know better.

Not sure I care overly much about the fate of the amoral corporation getting fucked over by the fascist regime. They’re both juggernauts, and I would love to see them damage each other.

Remember, the falling out wasn’t about the morality of the unsupervised spicy autocomplete killing people, it was about who had the liability when the AI went inevitably wrong. Had the DoD accepted the liability, I’m certain Anthropic would have sold the the stupidest version of skynet imaginable.

Oh man, I completely didn’t think about maintenance. Yeah, a data center will typically have several hard drives swapped per day. You’d have to have life support and a staff up there, as well as frequent resupply trips.

I am a bit late to this party, but I thought I’d piggy back on your comment to halfway address it using math.

We want to run data centers cool. This means keeping the center itself as close to 20°C as possible.

If we lose our convection and conduction then our satellite can only radiate away heat. The formula governing a black body radiator is P = σAT^4. We will neglect radiation received, though this is not actually a negligible amount.

If we set T = 20°C = 294K. Then we have the relationship of P/A = 423.6 W/m^2

According to an article I found on the Register from this April:

According to Google, the larger of the two offered pods will consume roughly 10 megawatts under full load.

This would imply a surface area of at minimum 23600 m^2 or 5.8 acres of radiator.

I don’t know how large, physically, such a pod would be. But looking at the satellite view of a google data center in Ohio that I could find, the total footprint area of one of the large building of their data centers is ballpark in that range. I don’t know how many “pods” that building contains.

So it’s not completely outside of the realm of possibility. It’s probably something that can be engineered with some care, despite my earlier misgivings. But putting things in orbit is very expensive, and latency is also a big factor. I can’t think of any particular practical advantages to putting this stuff into orbit other than getting them out of the jurisdiction of governments. (Not counting the hype and stock song and dance from simply announcing you’re going to set a few billion dollars on fire to put AI into space.)

https://youtube.com/watch?v=BKJ90uXkGdE

Is… that a reference to Frankie and Johnny’s???

Um… Hello, fellow person who grew up in or around New Orleans in the 1990s. How are things? Where did you wind up after Katrina? I have no idea how to handle someone just randomly referencing commercials I had completely forgotten about.

Twilight princess got a HD remaster though.

Having played all of the 2d Metroids, Super Metroid still holds up very well, even compared to the newer games. There’s less direction than what you might be used to though. Newer entries in the series tend to try to funnel you places. But if you are simply concerned about the gameplay being clunky, don’t be.

Super Metroid way ahead of its time, and the controls are extremely responsive and tight. It’s one of the reasons it’s such a popular speed game. It’s just fun to play. While basic controls are serviceable, there are many tricks you find yourself learning about as you play the game, which you just have access to, that give the movement system quite a bit of depth. (For hints, let the attract move run at the beginning of the game.)

Same tbh. Framework just works. And they do contribute back, just perhaps not as much specifically to Linux. Open source hardware is so incredibly valuable and important and rare. Honestly, if I had to choose only based on how much is contributed back, I’d still pick Framework.

Damn dude. Is your life really so pathetic that you are honestly deriving please from bad news being delivered to people who are on the opposite side of a fucking web browser debate? That’s sad. I hope things start looking up for you in the future.

“You mean I hitched my wagon to Grabthor The Wagon Destroyer, and he destroyed my wagon? This is an outrage!”

Or instead of targeting tiktok specifically, they could have chosen to pass a data privacy law and actually did something worthwhile instead of pointless, unpopular grandstanding. Haha just kidding, they would never do anything to reduce even slightly shareholder value.

Well, to play Devil’s advocate (ironically), I’ve spoken with some conservative Rabbis on this point when I worked in a Kosher kitchen at a Jewish summer camp, and their argument was that God, of course, knows about and intentionally inserted these loopholes into scripture and intended them to be found and used. However, the follow-up question of why bother doing that instead of just not disallowing the behavior in the first place did not yield a satisfactory answer.

It caused my brother to stop talking to me. He doesn’t understand how ChatGPT works, so he’s trying to woo his way to GenAI by layering some sort of fake ass natural language computation system on top of the spicy auto complete.

deleted by creator

Which is why the slang for diamonds is “ice.” They feel quite cold when you touch them because they have such high thermal conductivity.

LLMs are bad for the uses they’ve been recently pushed for, yes. But this is legitimately a very good use of them. This is natural language processing, within a narrow scope with a specific intention. This is exactly what it can be good at. Even if does have a high false negative rate, that’s still thousands and thousands of true positive cases that were addressed quickly and cheaply, and that a human auditor no longer needs to touch.