Gregory's Server|
14 posts total
Show previous comments
> While reviews note that you can run PUBG and other 3D games with decent performance on a Tecno Spark 8C, this doesn't mean that the device is fast enough to read posts on modern text-centric social media platforms or modern text-centric web forums. While 40fps is achievable in PUBG, we can easily see less than 0.4fps when scrolling on these sites. If your website is 100x slower than PUBG then you may want to reevaluate what you've done :neofox_woozy:> > Google: Are 100% of users on iOS? > Discourse: The influential users who spend money tend to be, I’ll tell you that ... Pointless to worry about cpu, it is effectively infinite already on iOS, and even with Qualcomm’s incompetence, will be within 4 more years on their embarrassing SoCs as well I... I don't even know how to respond :floofWoozy:"Unfortunately, a recent software update was not successful. Your vehicle cannot be driven. Please call customer support"
Show previous comments
Show previous comments
 @danluu @heathborders I mean, they actually sent a note and gave a reason, which puts them leaps ahead of most tech companies, in all seriousness.  Compared to 10 years ago, in terms of being able to find useful content, do Google search results feel better or worse to you? Anonymous poll
Poll
Significantly worse
882
0%
About the same
138
0%
Significantly better
0 people voted. 12
0%
Voting ended 4 Dec 2023 at 19:22.
Show previous comments
@danluu I find myself using ChatGPT more and more for all kinds of silly questions that Google has no answer for. I don’t know if that’s bad or not though 👀  The new stuff in the Twitter algorithm is wild. "author_is_elon", "author_is_republican", "author_is_democrat", etc., are explicit terms that are special cased. I'm a bit surprised at how many "well, actually, these are just metrics" replies I'm getting that seem to find the spirit of this comment plausible after (just for example), Elon asked why his engagement was dropping and then fired Twitter's most senior remaining engineer for telling him it's not a bug, which was followed by people seeing a lot more Elon in their feed and people also report seeing way more Republican stuff in their feed. Ever since getting burned by Hugo (https://twitter.com/danluu/status/1244023627613274115, etc.), I've made mental note when I see a maintainer complain that it's stupid to waste effort on backwards compatibility and I don't use their software. I don't think this would've saved me from Hugo since I'd never heard of the author before they wrote Hugo, but even in just a few years, these mental notes have saved me from software that constantly breaks at the maintainer's whim multiple times.
Show previous comments
 @danluu Before I went on to write an SSG whose explicit goals are extensibility and compatibility, I evaluated Hugo and found it quite amusingly bad: https://baturin.org/blog/why-not-hugo/ My "favorite" part is that it parses HTML internally in two different ways: in some places with regexes, in other places with a bundled HTML parser, and also calls external utilities for AsciiDoc etc., but gives the user no way to even configure asciidoctor options.  @danluu this is why I clearly document my opinions on what "compatibility" means and what assumptions I try to operate under for what I will do to support things. I can't make everyone happy, but I *can* take steps to ensure that people aren't surprised (negatively) by things. Separating out my opinions, tech debt, and forces ecosystem evolution, is something I've found particularly effective for this. If I think about everything I've learned about the world since I was a kid, the thing that would've most surprised kid-me is how privileged the median or even 10%-ile person with a "good" job is, e.g., what's the rate of tech folks at top companies who were too poor to get dental care and have crooked teeth? 1/1000 seems like the right order of magnitude for someone my age or younger in tech and maybe it's 1/100 but, at a population level in the U.S., seems more like 1/5 or maybe 1/10. Where are all the people who grew up too poor to get dental care ending up? Not in high-end tech jobs, and tech is less highly selected on this than finance, law, etc. I've interacted with quite a few interns and new grads at big tech companies and, among U.S. born folks, I think the rate of people who were on the U.S. Olympic team is pretty similar to the rate of people who haven't had dental care (IME, the Olympic team rate has been higher, but sampling from tails like this is noisy)  @danluu the fact that there are a lot of "wow it's a small world!" moments in these spaces are kinda all one needs to see to understand how homophilous these spaces are too I get that you can substitute dental care for other social markers but I do think such social differences might be less pronounced in countries with better social safety nets like Western European countries. I noticed even in Canada the class composition of those who intern at big tech companies is a lot broader than what you see in the US. My US intern friends within the same companies had a totally different pedigree than my Canadian friends. It really tickles my funny bone to see people make hardware SKUs to match arbitrary software restrictions. It's, of course, horrible, in that it generates massive deadweight loss but, in general, I view the "cause" of the deadweight loss as the software restriction and not the hardware workaround. A recent example is this IBM SKU that works around Oracle licensing limitations: if you have per-socket pricing, of course vendors will sell more cores per socket.
Show previous comments
 @danluu This is not really in the same category but you reminded me of the concept of weird IP workarounds. Are you familiar with ARM's workaround for Intel's AESNI patent? ARM's version broke aesdec (and aesenc) into two instructions instead of one and apparently that's enough to avoid infringement at the ISA level. But then their decoders just macro-op fuse each pair back into one macro-op equivalent to the Intel version. Almost reads like a comedy routine. @danluu back in 2007/8-ish at a previous job they deployed an entire extra VMware cluster to get the most out of their IBM WebSphere licences. what goes around comes around! I find this letter from John Carmack interesting in that it summarizes a sentiment I've heard from literally all of the highest impact/most effective people I've talked to at large companies: You can make a big difference, but you're constantly fighting against a self-sabotaging organization. I know some fairly high impact people who aren't bothered by this kind of thing; those people tend to say things like "I try not to care too much" or "I used to get too frustrated, so now I [do less].
Show previous comments
 @danluu My mind is kind of blown away at that thought. A bit saddened too... somehow He was such a pioneer in early GPU design Vague memories of early 3dfx and Voodoo cards coming back now @danluu in a specific sense, I'm glad that Facebook/Meta is less effective than they could be. They don't share my concerns and their current direction seems ill-considered. In general, large organizations suffer from similar problems. They include people with differing opinions and goals. So it's tough to get them all pulling in the same direction. Employees also realize that management is sometimes wrong and often temporary. So they won't rush to implement what they consider to be bad ideas. @danluu reward structures in large organizations also tend to lag behind their current needs. So people promoted because they're really good at X may not be able to quickly and effectively pivot to Y. When interacting with recruiters, are you an engineer | do you prefer phone calls or emails for info updates during each step of the process? For the purposes of this poll, programmers, developers, ops folks, etc., should be considered engineers. Anonymous poll
Poll
am engineer | email
684
89.4%
am engineer | phone
49
6.4%
not engineer | email
30
3.9%
not engineer | phone
765 people voted. 2
0.3%
Voting ended 23 Dec 2022 at 21:15.
Show previous comments
@danluu everyone on both sides should prefer a written record of every step for evidence.  @danluu I should prefer email, but email is swamped with receipts and spam that I dont really read email. WhatsApp would be great
Show previous comments
@EC_Commissioner_Breton @EC_Commissioner_Vestager please, notice it! It is a anticompetitive practice Of course the journalist bans combined with banning Mastodon as "malware" caused a spike in Mastodon growth. Just as that was settling down, a new wave of bans caused yet another spike in growth. This has been a pattern the past couple months. Every time Mastodon growth settles down (generally to a new, higher, rate), Elon causes another spike. I guess Elon only sees the engagement his actions drive on Twitter and doesn't notice that he was and is the biggest driver of Mastodon growth? @danluu at least in 2015 Twitter management was very careful about changes that might alienate users as it was aware it could run out of users to alienate One of the things that's made it easy to find communities whose discussion I want to see is the linear feed. This is a stark contrast to new social apps I've tried, e.g., Clubhouse, where my feed is dominated by high engagement stuff I don't want. I understand why companies push that stuff; looking at Twitter's experiments, you get more growth/$ when you switch users who've chosen linear timeline back to ranked timeline. Fundamentally, this is why most apps don't even offer linear timeline. At a meta level, something I find mildly interesting is how many people are writing stuff on Mastodon about how it's impossible for Mastodon to scale up without using an ad supported model (b/c server costs), it's better to have ranked feeds because most people want them, etc. The thing I think is interesting is that the people writing this stuff, implicitly, seemingly cannot conceive of a model where the organization is not growth and profit maximizing. |
@danluu byob: bring your own buttons
@danluu Cheers to ingenuity, but booooooo on the people that let this design go to production.
@danluu
There, fixed it.