• 0 Posts
  • 42 Comments
Joined 1 year ago
cake
Cake day: August 14th, 2023

help-circle
  • The requirements are steep but whether it’s a “good” steep or a “bad” steep will depend entirely on the end result. The pre-release footage looks good so it’s entirely possible that the system requirements are entirely justified. Or that they aren’t.

    We’ll need to wait and see the final result. Or, the handful of people on PC willing to put up with EGS will see the final result and the rest of us will debate that secondhand. Not sure how big the market will be on that front.

    At the least I expect AW2 to have a healthy lifespan in hardware benchmarking…

    (Note that since Epic funded this game from start to finish I don’t begrudge them making it exclusive to their store, unlike the other exclusives. Doesn’t mean I want to buy it, but I don’t fault them.)


  • Cities Skylines sees a fairly decent improvement going to the 3D cache chips from AMD (17% speedup here for the 5800x3D). Whats your ability to increase the budget to go for a 7800X3D look like? If this is a genre of game you like and you want to hold off as long as possible between upgrades, it might be worth springing the extra. The difference the 3D cache provides in some games is rather extraordinary. City builders, automation, and similar games tend to benefit the most. AAA games tend to benefit the least (some with effectively no gain).

    A 7600X should be more than capable of handling the game though. So it's not a question of need but if it's worth it to you.

    You do not want 4800 CL40 RAM though, that's too slow. I'd strongly recommend going for 32GB of RAM as well; 16GB can be gobbled up quickly, especially if you want to use mods in Cities Skylines.

    Going up even to DDR5-6000 is not much of a price increase. I'd suggest 6000 and something in the range of CL36-CL40. There's a lot of 32GB kits in those specs in the ~$90 range. I would not build a gaming system today with 16GB of RAM.


  • I don't think Kotick is at all certain to be kicked out. As easily as I can see MS letting him go with an enormous golden parachute, I can just as easily imagine them keeping him onboard because all they care about is Activision's ability to make money.

    In all likelihood Blizzard isn't going to be managed any differently. Microsoft's modus operandi with gaming acquisitions is to leave the leadership in place and let the dev/publisher run itself. Why is everyone expecting different here? The most likely outcome is MS does nothing to Blizzard and Blizzard continues on more or less the same trajectory as before.


  • That and the EGS seem to be where Epic funneled all their profits from the height of Fornite. That neither has worked out puts them on shakier ground. How many billions of dollars has been spent on EGS with it being way behind their revenue targets?

    As things stand, Epic has very little in the way of a next big revenue source when Fortnite starts to fade as something new takes its place. That (probably) isn't right around the corner but it will happen eventually. Their bet was on running major digital storefronts; that hasn't worked out. UE will continue to make good money but not anywhere near enough to sustain the company as it is. UE is simply far smaller than something like FN.

    This is likely them realizing this in conjunction with what you said. They need a new big revenue source in the pipeline, since digital storefronts won't be it. Whatever that next thing is will need lots of money.


  • Curious why everyone in the comments (as of my own comment) is happy about this?

    Sure, he exudes C-suite personality and doesn't act like he's a gamer. But that doesn't matter. He oversaw Sony's rise to dominance in the console market. That dominance is built on the foundation of their first party AAA games — which is a less than ten year old change for them. Sony porting their big games to PC was a project that was fully embraced under his leadership.

    Point being, as a gamer it seems like he's done a fairly decent job. I don't care how boring his interviews or speeches are or that he looks and acts like he belongs in a board room — they're all like that anyway even if their public persona says otherwise. I care about games and treatment of consumers.




  • I’m planning to upgrade from a 12 mini, which partly influenced my choice of years too (having seen 3 year data was the main part!). If I had a 12 Pro I think I’d have kept it for an extra year, but the battery is just not sufficient for how my phone use has changed.

    I think furthering your extra details here too is I saw someone point out that one of Apple’s slides for the base 15 was comparing its performance to the base 12. Apple knows how often people upgrade. Picking the 12 as a comparison point wouldn’t be an accident — we’re the single largest target audience for the 15. And in a year, they will in all likelihood compare the 16 to the 13 for the same reason.


  • This year’s new phones are for people that last bought a phone in 2020 or earlier. If the average user is on a three year upgrade cycle (what the data shows as I recall) then you’d expect roughly 1/3 of people to upgrade every year.

    This is better for Apple, as it keeps their revenue more spread out instead of heavily concentrated in year one of a three year cycle.
    This is better for consumers, as it means new features and upgrades are constantly being made. If they want to upgrade early they can, and they’ll get new features even if it’s only been two years.
    This is also better for both Apple and consumers because there’s more opportunities to course-correct or respond to feedback over issues. If Apple only released a phone every other or every three years, it’d take that much longer for the switch to USB-C.

    Just because a new product is launched does not mean you need to buy it. Nvidia released a new GPU last year, but I didn’t buy it even though it’s newer than what I currently have. Arguing that new phones shouldn’t come out each year is like arguing that new cars shouldn’t come out each year. It makes no sense.




  • This doesn’t need to immediately lower housing costs to have a positive impact.

    Hypothetical numbers… If housing was going to go up 5% in the next year and this change causes that to go down to a 1% increase, it will have made things better. Of course, we’d all like to just go straight to lowered housing costs. But individual changes can still do good and bring us towards that goal without strictly accomplishing it.


  • That depends on how long FCC is able to keep it implemented for, IMO.

    Something that gets lost a lot in policy discussion is that once you implement a business regulatory policy like this, you create a constituency for that policy. It’s an advantage in preserving hard fought gains but that also means the timelines need to work for it. The problem net neutrality faced the first time is that it was (a) late in Obama’s presidency, (b) held up by court cases, and (c) reversed early on by Trump’s FCC. There wasn’t much time for the internet business community to build a business model around it.

    If net neutrality is regulated into existence for 5+ years, at that point businesses will have come to rely on its existence. Taking it away will be harder, especially for a big pro-business party if it’s getting an earful from megacorporations that want things to stay as they are.

    Of course, I do agree that legislating it is the most robust option and would be the best course of action. I just don’t see legislation as the only option with any longevity. FCC rules can be that if the timelines work.




  • We still live in the same society as others. People often adopt the cultures and ideologies of where they end up, or at least move closer to it than they were before. If reddit shifts its userbase to the right — even if the net effect is from “very very left” to “very left” — it will impact the lives of all of us that live in societies with large numbers of people using that site, as it’ll filter down into our politics. Even if we don’t interact with them.

    For a long time, the “default” ideology of the internet was on the left. As internet usage has become dominated by a handful of sites owned by megacorporations, there has been a not at all subtle effort to nurture a conservative ideology on those sites. Stuff like reddit holding off on banning the Trump sub for however long or twitter refusing to implement their hate speech detection because it correlated too strongly with conservative politicians (not to mention what Musk has done there). I don’t think this is an accident.


  • The practical performance differences between N3B and N3E should be more or less immaterial to the end user. N3E just has a lower defect rate, meaning a greater portion of chips will be valid when made under that process versus made under N3B. There was a fairly credible rumor a few weeks ago that Apple was paying TSMC per valid chip instead of the industry standard per wafer. So for us, the end users, the cost won’t even be passed down — that’s just a cost that TSMC has to bear.

    That said, if you don’t need a new phone now, waiting is good in general. Whatever is out today, they’ll have something better next year. Wait as long as you’re willing and able between upgrades. Unless you’re absolutely loaded with money, I guess.


  • In this case it’s not truly a result of limited fab availability.

    TSMC has two main variants of their 3nm node. The original one, that Apple is using, is N3B. It has worse yields, so TSMC started work on another variant, N3E. N3E has much better yields but will not be ready until late 2023 or early 2024. Everyone else besides Apple opted to skip N3B and go for N3E. Apple, with their very consistent release cadence, didn’t want to wait for N3E. So Apple — and only Apple — is using N3B.

    Thus, we have:
    (1) TSMC only has one 3nm node in 2023: N3B.
    (2) TSMC only has one customer for N3B: Apple.
    (3) TSMC will never have any other customer use N3B, and have no incentives to build capacity beyond what is needed now.

    It’s effectively tautological that their entire 3nm allocation will be sold exclusively to Apple in 2023.