IHeartBadCode

Ha ha, but for real. They’ll just turn people’s power off, tell them to ration, and/or jack the price per kWh to 500% what it was.

Team Orange let’s no good calamity go to waste. Everything is potential profit if you have no moral compass.

Their creditors are going to be looking for recompense. So I feel like this is one of those times when we use the answer of “yes”.

Yes, yes, now what about the rest of the stock market?

To say, “oh this boycott is self injury” is akin to worrying about one’s stubbed toe all while bleeding out from a severed arm.

Additionally it’s typical American only thinking to believe it’s just the US citizens boycotting the company. You easily forget that sales are down globally, not just the US.

Yeah, China sure as shit isn’t going to lose sleep over a US Copyright case.

Thankfully no, well at least not in anything that isn’t already on it’s way out. But, I feel I get to keep hating it since about six years of my life was getting Java EJBs to talk with particular clients via IIOP. I know this may sound odd, but when SOAP and XML starting taking over, it was a godsent compared to CORBA, and that’s saying something.

From the story.

Cursor AI’s abrupt refusal represents an ironic twist in the rise of “vibe coding”—a term coined by Andrej Karpathy that describes when developers use AI tools to generate code based on natural language descriptions without fully understanding how it works. While vibe coding prioritizes speed and experimentation by having users simply describe what they want and accept AI suggestions, Cursor’s philosophical pushback seems to directly challenge the effortless “vibes-based” workflow its users have come to expect from modern AI coding assistants

Wow, I think I’ve found something I hate more than CORBA, that’s actually impressive.

Instructions unclear, jerking all body parts off human and placing them into a neat bloody pile in the corner.

— Future Robot probably

Oh yeah, but the thing that usually offsets the intrusive thoughts is a lot of courts treat this as the crime of “hurting rich people” which comes with like 30 years in pound you in the ass penitentiary.

WHY ARE THEIR ROCKETS NOT EXPLODING!!?

— Rocket X. Hitler

I’m curious, how repairable? Like comfortable with a solder iron or slots and what not like a PC?

Repairable phones would be great but the demand for them hasn’t undone the cost of design for them. There’s a lot of tech in an incredibly small package, so repairable phone would still require people to have specialty equipment to repair.

Like very few people own an oven for working with BGA chips. And if we go with socket based chips, the thickness of the phone has to increase or the battery has to decrease.

Don’t get me wrong, I think an open and repairable phone would be great. But having one is an engineering challenge that most phone makers have opted to just skip putting dollars into because the demand for one doesn’t justify the cost. Your average buyer is just chasing shiny and doesn’t see repairing their dinosaur as valuable.

But yeah, I’m sure there’s plenty here that would love such a device. Sadly we are not the majority.

Fucking CyberTruck like fucking pile of shit website. What kills me the most is that the fucking things they’re screenshoting, those pages have literal “export to XML” buttons that they could fucking export, save the XML to some shared drive that gets swept, and the put it in some actually secure database.

This whole fucking thing reeks of some fucking weeb ass Roblox hackers whose last project consisted of Lua Script emulating some fucking redstone calculator they wrote in Minecraft. And the export fuction on the thing? It’s just one dimension SUM function CSV exports. Literally no other dimenstions of values to add, shit I would be fucking surprised if a single one of the people writing the goddamn have ever heard of OLAP.

And to top it off, we already have a fucking website that does what this fucking place does, but 846 decillion times better. And it doesn’t have a fucking Instagram esque reel of Tweets of people taking fucking screenshots of an open database.

I can’t wait till the next dumbass gets into the White House and turns this pile of grabage off. Paying these idiots millions to power and run the hardware this pitiful excuse of a website runs on. And all we got for that money is some shit that is about on par as the shit you get from some O’Reilly book called “Building a Government Website Crash Course” with a Bald Eagle dying of bird flu on the cover.

This fucking idiot maybe wants to fucking learn what the hell SQL is.

Distrowatch doesn’t research anything and cries foul without second thought because Meta is evil

But it’s not Distrowatch’s job to verify an opaque censorship process. Meta is the one who created their filters and their filters are not open for public review. So it is entirely incumbent for Meta to handle the matter.

Distrowatch is correct to cry foul because that’s literally all they can do. It’s not like they can suggest a patch on github or something.

We have to remember that black box logic is wholly owned by the author of the logic. If it ain’t working, then yeah, cry foul, there’s no additional research to be done. That’s literally the entire point of obfuscating logic in a service, to ensure that nobody else can review the internals.

Meta was completely in the wrong. Distrowatch called them out on their fuckery. That is exactly how the system is supposed to work when a company blackboxes their shit.

Facebook support (person making around $3.5 per month in some third world country) doesn’t know difference between specific Linux distro and Linux itself, tells Distrowatch that Linux is now banned

Very likely. Run a shit company, get shit results. Distrowatch running with “Linux is now banned on Facebook” is not a result of lack of research, it’s the result of a company that just gave up on giving a fuck.

Our standards for companies have really fucking eroded over time and boy oh boy do CEOs eat that shit up. Meta has a systemic failure on-going in their company to which the C-Staff do NOT care one bit about. This episode is a manifestation of those failures. Meta has to up their fucking game here.

This notion that Distrowatch should have… That’s like saying someone who cut their mouth on glass in their McDonald’s burger is to blame for not first checking their burger for glass. The glass shouldn’t fucking be there in the first place. Customers have a reasonable expectation that a company isn’t a dumpster fire and it isn’t incumbent on the customers to ensure they aren’t stepping into a goddamn disaster zone.

I just really need people to understand, we have got to upper our expectations of companies. Because every time we let something like “oh well Distrowatch should have known they were talking to a complete moron”, we are letting these asshats who are currently enriching themselves on the United State’s taxpayer’s dime, get away with it.

Meta was fucking up badly and Distrowatch was letting everyone know in medias res how Meta was fucking up. This is 100% Meta fucked up. Don’t want $3.50/mo employees giving shit answers? Likely a good way for Meta to solve that is to NOT have fucking $3.50/month employees. It’s a pretty clear strategy for them to consider. Till then, they’re likely going to be handing out bullshit answers that contain no sense of logic and we ought to fucking call them out on it.

We live in everyone is dumb timeline

I’m not going to have this, “well people should have known better.” We ought not excuse Meta for this monumental fuck up. This is theirs to own. Everyone cannot randomly research every line of bullshit that’s pandered off by companies. There is just no time for that non-sense. If Distrowatch says “Meta told us Linux is banned” and provides the email to back it up, then until Meta says otherwise, that should be taken as the gospel of the company. If Meta thinks that it shouldn’t have gone the way it went, then Meta needs to fix their fucking hiring policies.

WE HAVE GOT TO STOP EXCUSING THESE PEOPLE. They are NOT going to act better if we give them even a single centimeter.

I get what you are saying, but this isn’t Distrowatch’s thing to reevaluate how stories hit the front page. Meta fucked up every step of the way. And for a company that’s pulling down a seven digit multiple value to what Distrowatch pulls in a year. Meta can fucking figure it out because they have access to a ten million fold more resources.

That’s just my two cents on this explanation.

this means deepseek is based on an openai model?

It doesn’t sound like it is. It sounds more like it’s hallucinating which DeepSeeks has a really light end fine-tuning. But who knows? While their stuff is Open Source, no one has yet to test it and see if they can reproduce the results DeepSeek got. For all we know this is just a Chinese con or the real deal. But not knowing how you landed into this point of the conversation it comes off as a context aware hallucination.

It knows about openai and it being a LLM but it’s mixed up self identity in specific with identity in general. That is it is start to confuse LLMs and ChatGPT as meaning the same thing and then trying to wire back this bad assumption to make sense again.

Again, who really knows at this point? It’s too new and it being in China, there’s likely no way to verify these people’s claims until someone can take what they’ve published and made a similar LLM.

Exactly. What the banks are doing are selling “loans”. Musk has to pay those loans back quote/unquote someday. If the loan is good, you hold on to it as a bank because the interest makes you money. If the loan is bad, you sell it so that you can get some of your money back and make the collection of the loan someone else’s problem.

Banks will do this for a number of reasons:

• To manage their balance sheet. Every loan not paid in full is bad and you need to balance good (income/good loans/etc) and bad.
• Generate immediate liquidity. Banks need to have some hard cash on hand, sometimes they sell to do just that, have hard cash.
• Free up credit lines to lend to new borrowers. Banks only have so much resources, sometime you cut losses to get new gains.
• Diversify the risk pool. You want a nice balance between “loans that might default” and “loans likely to not default”.

Now for everyone else, what the parent to this comment is indicating is the second option in that list. Having to create some cold hard cash suddenly. Usually, there’s a cyclical nature to needing greenbacks by the fistful, but like everything that’s not always true. Something can “happen” and you have a sudden need to have cash in hand pronto. Good way to get that cash is to start selling low hanging fruit if you have it.

Something like the Twitter loan is a good pitch for low hanging fruit. Musk is terrible at paying the loan back, Twitter is likely to default one day, but Musk suddenly has direct access to some pretty corrupt as fuck ways to actually pay that loan back. From what I’ve read in the article, the sell price is something like 90 to 95 cents on the dollar. So not a huge discount, this ain’t a fire sale.

But banks might want to offload Musk from their sheets just in case that money is something someone might later investigate. Like that 95 cents on the dollar price is “We think Musk is good for it, but we likely don’t actually want his money.” So you can make that federal investigation in 2033 someone else’s problem, by selling the loan today. The big bank makes about 95% of the original amount back and when Musk goes to pay his loan in Russian Blood Rubles, it’ll be to a bank that get investigated that isn’t <<insert some large bank that would “NEVER” think to take conflicted money>>.

That’s one theory. But there could be something on the horizon. Something that isn’t right around the corner, but coming up in the distance that the banks want to have cash on hand for. Usually you see a much larger discount, like 60 cents on the dollar, for “holy shit, this stuff is toxic but we need to offload it discreetly before everyone else wises up.”

I don’t think point one and three apply to Musk’s particular set of loans. But who knows?! Only the bankers do.

Indeed. A modern Nissan Leaf with a 62 kWh battery can charge in a little over 11 minutes if you have a 2kV 160 amp line to toss into it. Because you know, it’s completely safe and cool to deal with those kinds of values for the average consumer.

Yeah, something got in their crawl and Bluesky is on low key TPOT rising.

Someone did a post on it, you can see the map that was done in the picture there with TPOT in blue.

I don’t know why they started coalescing on Bluesky, but it is what it is. If they get annoying there, I know how to mute them. But as long as that doesn’t also herald the “free speech absolutists” that literally add zero value, I’m cool with TPOT heading to Bluesky.

Besides, I’m more active on Fedi/Mastodon at this point than Threads/Bluesky so it’s not really a pressing thing for me. I mean shit, I wouldn’t have a problem with Musk’s network if he’d clamp down on the utter trash going on over there. Like it’s gotten to pure garbage, I haven’t the time to block everyone who pops into my feeds that how hard the algo there is pushing the trash.

The postrat folk. The deep value Silicon Valley folk. Core Techbro kind of people.

Would have thought they’d be prone to sticking with Musk

Ditto, but at the same time. Being with daddy Musk might be too traditional at this point. No idea the reasons, but you can to see a lot of this popping up in that circle on Twitter.

TPOT → Bluesky is actually an interesting example of what looks like a successful transplantation… quasi-existential concerns about Elon Twitter, vibes have been off leading to big cascades of migration tend to happen after inciting incidents (eg twitter banning substack links being a canary in the coal mine)

You know the “I sound super thoughtful” kind of stuff. Lots of praise from that Group on XTwitter/Bluesky.

I don’t know. Lots of folks pointing out That Part Of Twitter (TPOT) also migrating over.

what long term storage would be the best option for storing digital information

The biggest factor with passive storage, something that’s not refreshed. Optical media that’s made to last. M-DISC comes to mind, but there’s no proof that worse case 100 years is a valid claim. Chemically speaking, a well kept disc should keep 100 years, but that’s chemical composition in an ideal case. Nothing in manufacturing is perfect, so impurities are always going to be there robbing the lifespan of these discs.

Magnetic tape ideally lasts decades if not close to a century, but these are tapes that are kept in incredibly controlled conditions. If you’ve ever worked in the server world you’ll know that any plain Jane LTO magnetic tape can’t be trusted after collecting dust for anywhere close to five years.

We have scrolls, we have books, and we have stone tablets that have endured centuries, but the key in all of those is how well they were kept. The construction matters, but the bigger aspect is the environment they were kept in. For digital media, we don’t know any real way to keep digital data in a passive state for centuries because, well, we haven’t had digital data for that long. We’ve got really old punch cards that are close to that age, but even then, some of the oldest stacks are now sitting in hermetically sealed cases and are actively upkept to prevent UV damage by clear coating those cases on a regular basis.

And that’s the thing with digital media, keeping it in an active storage rather than passive may be the key for centuries of longevity. USB sticks are fine so long as someone remembers to plug them in and allow them to refresh every some many years. Most USB sticks use ceramic capacitors, so leakage there isn’t too much an issue. The bigger thing might be corrosion of the various traces and pins, but if well kept, that might take decades to eventually make an impact.

Sometimes, I like to parallel digital long term storage as the Ship of Theseus. If you keep moving the data from one device to another, it’s still the same data. And in that sense, the data can live forever. Even if there’s a gap of say two decades, if you can still get to the data and convert it into something modern, the data lives on. It’s not the original medium, but with digital data, it doesn’t have to be, that’s the neat thing about digital data.

I think people still are working on trying to wrap their heads around digital data versus the way we used to do it. You know, someone might have the family bible and we’ve got to keep it nice and tidy and careful with it, because with analog data the medium and the information are one in the same. And I think sometimes people look at family digital photo collections like that. Like it’s the family bible and that we’ve got to keep it safe. But if it’s a USB stick that you pull out every so often, look over it, and call it day. Maybe move the photos from the Walmart USB stick that you got in 2016 to the new 800TB USB-F stick you just got from neo-Amazon in 2073, those photos can live forever. You don’t have to be careful with them anymore.

I think that’s one of the reasons that open formats matter so much. If you stored all your family videos in Windows Media Format, what happens when Microsoft dies in the Second US Civil War of 2038? That’s not helping you in 2073 to open those files on a format you can never figure out. But say you stored it in some open format. Now all you need is an implementation of that format and a compiler. And poof, now you have a modern codec to read the files of the before times.

It’s one of those fun maybe slightly existential kinds of things. Nothing lasts, no matter how hard we try, nothing will last. All things forgotten decay, we can only slow that decay down, but we can’t prevent it. But things that live, things that pass through the hands of the living, those things endure because there are people who put time, one of the most precious resources we have, into them. Our reward for that investment of time is something that continues beyond the decay.

I like to think of it as the balance of the universe. You get to keep this, but only if you give a bit of time to pay for keeping it. And sometimes it’s crazy to think of how much that applies to. Also I likely shouldn’t reply after having a few drinks. Wooooo!!

Environment makes all the difference for passive storage, sorry I really went out there on the reply.

Best bet is long term optical discs or long term magnetic tape. USB keys are not good for long term storage. USB keys use NAND memory that is a series of floating gate metal oxide semiconductors (FGMOS). These operate by using Fowler-Nordheim tunneling, in where a charge is carried along a regular style fin field-effect transistor (FINFET) and a charge above the transistor’s channel causes some electrons to quantum tunnel into floating gates that are isolated by oxides.

While these floating gates are sealed off from everything, so the charge should stay “indefinitely”, quantum effects cause some of the electrons to “leak” out of the floating gate, causing a degradation of the stored signal. Typically there’s a refresh circuit within the USB key’s integrated circuit that takes care of that and USB data can last seemingly forever. However, that refresh circuit requires a small amount of power, which if you store the USB stick somewhere for years on end, will never get powered.

This is the reason why flash memory only assures data can be retained for about ten years without power. Eventually the electrons “trapped” in floating gate have enough time to tunnel out of the floating gate completely obliterating the signal. The tunnel events aren’t many per second, but give enough time, and all of those events add up. Paired with the whole thing that USB sticks mostly no longer use binary logic levels. Most are now using something like four or eight logic levels. So instead of there just being on and off, there is 0V-0.7V = 00, 1V-1.7V = 01, 2V-2.7V = 10, 3V-3.7V = 11 logic levels. So a small amount of charge loss can create a different bit pattern.

One thing to look at for long term storage is something like M-DISC. The matter by which the burned data onto the optical media is made is via a process that takes about 10,000 years (estimated) to break down. However, the disc itself is in a polycarbonate thermoplastic that has an average breakdown of only about 1,000 years in extremely dry environments and about a tenth of that in your average sealed lock box environments.

Your average spinning disk hard drive can store information for some time, but the storage requirements are pretty intense and even then hard drives loose about 1% of the magnetic strength per year without power. And about 70 years is the max before the various magnetic bits that form the low level format of the disk have degraded without power to the point that the disk has too many bad sectors to be called usable. But outside of that, the biggest fault is mechanical failure. No matter how well you think you’ve stored a drive, it’s never good enough and the spinny bits always fail from becoming too fragile from pervasive oxidation. Basically the drive will spin up only to tear itself apart as some weaken part of the armature flies into the spinning platters.

But USB sticks will only give you about a decade before the stored information fades away into the quantum ether.