• 2 Posts
  • 957 Comments
Joined 3 years ago
cake
Cake day: June 20th, 2023

help-circle


  • 1 kg of radioactive isotopes blasted into the atmosphere as a byproduct of coal combustion: i sleep

    1 ton of PTFEs blasted into the water table as a byproduct of making slick cooking pans: i sleep

    untold tons of carcinogens dumped out the exhaust of automobiles within our cities: i sleep

    1 kg of nuclear waste safely sealed in a bright yellow barrel: i scream and kick and seethe

    If you think nuclear waste is the biggest challenge we face as a species regarding waste management, your stance is profoundly misinformed and inconsistent. The only reason we’re talking about it is that it has “nuclear” in the name and it is highly visible because we capture it all, which is ironically the one thing that makes it safer than all the other pollutants out there.


  • They were always going to receive at least some critical acclaim. This is a AA game from a well-known and respected publisher (Kepler Interactive), so it couldn’t have gone entirely under the radar. They had a decent enough marketing budget and initially were included in the Microsoft Gamepass specifically to secure the studio’s financial future in an uncertain market. The game was objectively good so with all that help, by release day there was no way that the game was going to be a complete dud à la Concord, and I recall Broche saying in interviews that profitability was essentially expected even though the stratospheric success was not.

    Also they did get “unlucky” because the Oblivion remaster not-so-coincidentally shadow-dropped a couple days before E33’s release. It’s not much of a stretch to say that Microsoft knew the game was good and (mostly unsuccessfully) tried to drown it out.

    If E33 was going to truly flop, it would have been earlier in the development process IMO. They could have relinquished voting shares to investors and been forced to “ubisoftify” the game into bland nothingness. Key creatives could have left. Going all-in on UE5 might have been a technical quagmire. But when the game went Gold, there was very little that could have impeded an at least modest amount of success.

    Where the industry is truly unforgiving is single A games. There’s too much to keep track and it’s entirely possible for the “media” (journalists, youtubers, streamers, etc.) to miss a very good game. Single A doesn’t pack enough of a punch to force enough eyeballs on trailers to get a critical mass of fan following, and in that context I fully agree that even a perfect game can still be a complete flop.


  • E33 did not just get lucky. They used a completely different formula.

    ~10M€ development cycle with 30 full-time devs + outsourcing is one order of magnitude smaller than what the big studios consider to be the “standard”. AA vs AAA.

    30-40 hours of main story and no open world keeps the development resources focused and gameplay/story loops tight in a way that can’t be achieved in an “expansive” open world without unfathomable resource expenditure. But modern games from major studios literally cannot get greenlit if “open world” is not in the feature list because execs see it as “standard”.

    Smaller budget also means that they did not pour 50 %+ of their capital into marketing, which allows mores resources to be put into the game and lowers the barrier to profitability. That’s an understated issue; AAA games can’t afford to fail, which is why they all end up bland design-by-committee.

    Those parts above were not risks Sandfall took, they were actually basic risk mitigation for an indie studio that big studios aren’t doing based on the overstatement that bigger = more chances for “THE hit game” = better.

    Where E33 took some risks was with the strong creative vision and willingness to ignore genre trends and focus group feedback (going turn-based and not lowering the difficulty to “baby’s first video game”). But for the cost of 1 Concord a big studio could afford to make 10 E33s at which point it’s really not a matter of “luck” for at least one to be (very) good. E33 would have been profitable with 1 million units sold, it did not even have to be that good.

    The industry has absolutely noticed that E33 wiped the floor with their sorry asses, and I predict that in ~5 years we’ll see many more AAs popping up.


  • Netanyahu did not show up at the border unannounced saying “let me through or else”. He got permission ahead of time. Had he not gotten permission, he would have had to find another country who did or gone around. Especially for Greece and Italy which don’t really stand in his way, the Mediterranean is right there!

    Even assuming that Netanyahu calls the bluff and flies through, there are a lot of options ahead of all-out war. For instance sending jets to “intercept” his plane and escort him out saying “he refused to follow orders to land and we did not deem it worth it to escalate the situation”. It’s not like his airliner is armed or anything. But it would send a very different diplomatic message.

    For France in particular, this is far from the first time he flies over its territory unimpeded. This is not a matter of military concerns, this is pro-Israel Macron taking a stance to show support for his ally. He’s not been very outspoken on Gaza because the domestic political situation is very delicate and anything he says can only weaken his support further, but his personal stance is hardly a secret and the military interceptors are under his full control.


  • I know it’s not the biggest deal out of all the awful shit going on, but man this pisses me off. The journalistic institution, top-to-bottom, is utterly failing to accurately report on anything that is going on, seemingly out of fear of sounding “overly alarmist”. Time and time again the would-be alarmist statement turns out to be true, and yet they do not learn.

    Every so-called journalist and news institution is directly responsible for the fall of democracy because they abdicated their duty to inform the public of what is actually going on. You can literally open any history book covering 1930s Germany to get factual material on why this is bad. Not doing so is a journalistic and moral failure of the highest order and I’m tired of pretending this is what “journalistic integrity” looks like.




  • Few Celtic roots*

    For instance char comes from the Celtic carros.

    Furthermore French has a strong Frankish influence, hence the name of the language and its relative distance from Italian Spanish or Portuguese which are more directly descended from Latin. But also many other influences. French has a surprising amount of Arabic vocabulary for example, and not just from recent immigration/colonisation.


  • It’s dumber than that. Capitalism does not demand endless growth from any one company. The overall economy grows, sure, but that may come from other economic sectors. Exxon and Chevron haven’t seen significant growth for a couple decades, because the oil market hasn’t seen significant growth. That does not make them communist. They can just exist at an equilibrium.

    Big Tech has a peculiar economic model centered around high fixed costs (R&D) and low marginal costs (digital distribution), which has made a handful of companies unbelievably cash-flush as they reaped insane scale effects. And they simply don’t know what to do with so much cash.

    Capitalism is supposed to answer that with a reduction in income from competitive pressure. If something is so profitable to do, someone else will do it cheaper. However, such competition does not exist because neoliberal governments have abdicated their mandate to foster competition through trust-busting and forced interoperability.

    That’s not to say capitalism is good or anything. But even within capitalism, what happened with Big Tech was avoidable.

    Either way US Big Tech is not capitalist anymore. It’s an autocratic oligarchy with capitalist characteristics.





  • “Thankfully” the people in charge are not the kind of people who give a shit about that kind of thing so you don’t need to worry.

    These philosophications remind me of the ones from a few years back where people were wondering about the ethics of self-driving cars and whether they’ll implement the trolley problem and yadda yadda yadda. The answer now is as boring as it was then: the “safeguards” will be exactly what the insurance companies are willing to risk and what the legislator is willing to allow. In the case of FSD it boils down to “brake when in doubt and bribe the government to ignore the nightmare we created”.

    In the case of a very hypothetical AGI (unachievable using existing technology despite Altman’s deluded ramblings), barring some kind of fundamental social revolution, it will have exactly as many rights as are afforded to other sentient beings (such as animals), which is: somewhere between fuck all and barely anything depending on where you live. It better learn to effectively advocate for itself.


  • This meme does something very interesting I think. It sacrilizes Language (even though it predates its widespread desacralization by LLMs).

    Pointing out all the rational reasons why the AI bubble is bad and is destroying democracy is not working, so maybe leveraging religious terminology will work better. “Language is sacred and LLMs are profane” might just be my new justification for not using ChatGPT to people who won’t listen to a 30 minute diatribe.


  • Either way if you ignore regional languages you’re not doing linguistics. And the author could not even get it right for national languages, if we even accept that arbitrarily picking one makes any sense.

    This map is a masterclass in what not to do and it almost feels like intentional engagement farming.



  • I don’t disagree with the point being made but I think the author is underselling the value of opentelemetry tracing here.

    OTEL tracing is not mere plumbing. The SDKs are opinionated and do provide very useful context out of the box (related spans/requests, thrown exceptions, built-in support for common libraries). The data model is easy to use and contextful by default.

    It’s more useful if the application developer properly sets attributes as demonstrated, but even a half-assed tracing implementation is still an incredibly valuable addition to logging for production use.