Wow. The user leaderboard is crazy. It's literally like, three guys getting 200 likes combined, down to two guys with 70 likes combined over the last quarter, then now a guy with 35 likes over the past month.

My gut is telling me that THE lesson learned from Twitter buyout is that, you could trim off 99% of devops out of Twitter(2019), it just leaves you with ~100% opex and 1% as much sales, and you don't want that.

  1: https://devcommunity.x.com/u?order=likes_received&period=quarterly

What makes general OS difficult on a SoC is usually the lack of hardware virtual<->real RAM address translation accelerator, the MMU.

I guess it's not that important if developers had no raw pointer access so that object can be anywhere dynamically assigned(Java!!! also JavaScript and all interpreted langs). But basically all desktop OS apps are written in C and they always want single consistent virtual addresses to jump around in and none of apps work without the hardware dynamic translation.

C semantics account for segmented storage, that is why two pointers holding the same bitwise address are not considered to point to the same object, when they are derived differently.

Practically ALL course introductory materials that regard robotics and AI that I've seen began with "you might imagine a talking bipedal humanoid when you hear the word `robot`, but perhaps the most commonplace robot that you have seen is a vending machine", with the illustration of a typical 80s-90s outdoor soda vendor with no apparent moving parts.

So "maybe cars are a bit of robots too" is more like 30-50 years behind the time.

Real time as in at >1x speed? Probably?

Real time as in per-word basis? Probably not?

If you are thinking about section 230 it only applies to user–generated content, so not server–side AI or timeline algorithms.

So if a social network tool does the exact same thing, but uses the user's own GPU or NPU to generate the content instead, suddenly it's fine?

If a user generates child porn on their own and uploads it to a social network, the social network is shielded from liability until they refuse to delete it.

Changes who's technically the perp. Offer managed porn generation service -> service provider is responsible for generating porn, like, literally literally

Grok do seem to have tons of useless guardrails. Reportedly you can't prompt it directly. But also reportedly they tend to go for almost nonsensically off-guardrail interpretation of prompts.

I wouldn't know but the exploded view do look a little sus even from a printing hobbyist's perspective

There are apparently rad-hard DDR4 chips these days so this is patently false. SpaceX used to talk a lot about substituting rad-hard components with triple redundant regular x86 years ago, that's true.

I think I've also seen someone mention that the cost and power benefit of substituting rad-hard chips with garden variety wean off fast once the level of redundancy goes up, and also it can't handle deep space radiations that just kill Earthbound chips rather than partially glitching them.

no, no, no, the key enabler to space datacenter is complete out-of-world computer 3D printers. You print entire 130nm, hell, even 130 micrometer GPUs and DDR2 VRAMs to go with, entirely on the Moon solely from Moon dusts, and shoot the complete satellites out into Earth LEO using maglev sleds. PUEs, opex, nothing matter because the Moon factory is self contained and don't interact with Earthian economy at all. The ssh key into the factory will be the source to free money to whoever holding it.

Is that possible in our lifetime? I'd be optimistic about that. Can SpaceX pull that off? Space what? ...

> Due to a miscommunication with the factory, the injection pins were moved inside the heatsink fins, causing the cylindrical extrusions below.

Never done casting let alone worked with Chinese factory to ship hundreds of units but: this sounds potentially intentional at factory's end. It's plausible that these fins didn't pass factory guy's manufacturability gut DRC who would made changes thinking the customer would just give in. IIUC molding people in general don't like corners and narrow channels with no sprues and/or gas escapes. Especially the outer ring of pins appear to be exactly where they like to place sprues.

I wonder how this problem was eventually solved. The final product seem to retain the number, height, draft angles of the fins, but fillet radii appear to have increased(2mm -> 5mm?) and the entire body shows more material shrinkage at the outer edge of the body as well. Was it just no pins and higher defect rate, or something else entirely?