Michał Matyas (@nerdblogpl) header
Michał Matyas (@nerdblogpl) profile picture

Tech person

Kraków, Poland
Joined February 2016
I feel you, Gemini... I feel you...
1
52
Michał Matyas retweeted
this is what apps looked like 12 years ago
125
161
60
4,139
299,442
PSA: If you're using @Dropbox, the latest version is overriding the Mac OS floating thumbnail in the bottom right corner with their own version that doesn't support drag & drop. After you disable it, press ⌘+shift+5, go to Options and re-enable "Show Floating Thumbnail".
93
On one hand, this is quite impressive. On the other, if this is the future of programming then I may just as well start looking for a new job because the process sounds absolutely horrible.
Ok, going to vibe code again tonight. This time, I'll try to build a zombie survival first-person shooter in blocky style with Cursor + Sonnet 3.7 + @superwhisper. Maybe too ambitious, but we'll see. Follow along in this thread 🧵
49
Michał Matyas retweeted
Replying to @zack_overflow
Web performance engineer here! Browsers don’t actually render stuff outside of the screen! At least in the narrow sense of “painting what’s off-screen”. Browsers, however, have other “rendering”-ish work that still they have to do for off-screen elements. This work is computing 1) styles, and 2) layout (sizes/positions) of elements on the page, and it’s this work that content-visibility: auto disables. Why can’t browsers skip this work by default? Well, for one, if they don’t compute sizes for off-screen elements, how would they know how tall the page is? If they don’t know that, they can’t even render the scroll bar correctly. This is why `content-visibility: auto` by itself is actually not sufficient enough. You need to also add `content-intrinsic-height: XXXpx` on each element with `content-visibility`. If you don’t, then off-screen elements will render as 0px tall, and when you scroll, you’ll see your scrollbar jump worse than a toddler
8
33
2
1,068
19,958
I dug out my Meta Quest 3 recently and it's fascinating how the company that renamed for the sake of VR world, came up with such an amazing piece of technology etc. also invented probably the worst possible hands-free user experience possible. The pinch-to-click is so bad.
1
57
One of the major problems when working in web development is that backenders often think of the clientside UI as a weird version of a server and assume certain guarantees. And then they get annoyed or outright angry that the UI/front-end is a very different realm.
32
Michał Matyas retweeted
mfw, when I try to search for a C library on GitHub (multibillion dollar company advancing AI and shit btw)
63
72
7
2,879
81,504
Michał Matyas retweeted
> how many r's are there in strawberry? > whole foods organic strawberries, now just $3.99!
65
22
20
1,250
60,046
average LLM experience:
Ah! I see the issue now. Let me just fix that Ah! I see the issue now. Let me just fix that Ah! I see the issue now. Let me just fix that Ah! I see the issue now. Let me just fix that Ah! I see the issue now. Let me just fix that
25
Does anyone have a good solution to websocket connection dying when tab goes to background considering ActionCable doesn't have message replay support? Or do you just accept that the app may randomly become broken and require a full refresh?
The more I learn how ActionCable and websockets work, the more I'm wondering why anyone would want to use Hotwire in a production level app
1
41
Since I'm integrating it with graphql-ruby my plan is to monkey-patch ConnectionMonitor from rails/actioncable JS package to add an event on reconnect and then trigger Apollo's reFetchObservableQueries() but I'm surprised I couldn't find anyone else mentioning doing it.
1
24
The more I learn how ActionCable and websockets work, the more I'm wondering why anyone would want to use Hotwire in a production level app
1
70
banger @deepfates
151
1,846
118
45,135
1,200,964
Damn it, I was late with a snarky comment by few seconds
1
18
(it's been down for at least an hour by now)
14