radicality

On Mac, I recently mistakenly held Shift (or some other key combo?), before hitting Space to open QuickLook as I was gonna select a few image files. It proceeded to open quicklook with an incredibly slow >5s animation, very annoying :S

Whoa, is that right, I somehow never knew this. Why does it do that, does it still make sense if 3rd party cookies are disabled? And is there a way to disable it apart from the add to home screen?

Surprised at your issues with HA. Similarly to others that responded, my setup with HA / zigbee2mqtt and >30 zigbee devices (including some ikea buttons) has been pretty rock solid over many years, including easy migrations from an rpi3 -> rpi4 -> rpi5 (with ssd).

Usually when I had some zigbee issue, it was because of a crappy product (eg some wired air sensor that would spam the zigbee network every 1 second with a lot of data), so then I just stop using such devices and before I buy I check compatibility with HA / zigbee2mqtt.

Yeah I noticed that too, it’s a bit crazy to me that stuff like this is getting upvotes and traction. It feels like it was vibe-coded in one-shot style without perhaps even reading any bit of the code. A bunch of hardcoded values, a `sleep(3)`, bunch of other antipatterns.

Up until recently I tended to “trust” github repos a bit more, now I feel like I need to have my guard up so I don’t fall into a trap of using something like this. Funnily enough a good first metric for me now is # of emojis in the readme - the more emojis the more likely you should stay away from it

Uhm, does this do anything else useful besides shelling out to https://github.com/ryoppippi/ccusage ? Idk, this kind of stuff feels a bit disappointing, also feels like it’s perhaps one-shotted with some AI tool, and in this show hn you don’t even mention that all the actual work is done by some other tool.

Can you elaborate on “paid” ? Because I honestly still have no idea if my usage of AI Studio is used for training purposes.

I have google workspace business standard, which comes with some pro AI features. Eg, Gemini chat clearly shows “Pro”, and says something like “chats in your organization won’t be used for training”. On AI Studio it’s not clear at all. I do have some version of paid AI services through Google, but no idea if it applies to AI studio. I did create some dummy Google cloud project which allowed me to generate api key, but afaik I still haven’t authorized any billing method.

Haha, oh yeah it’s completely fake, just meant to imitate how ridiculous and hard to understand normal git manual and settings are, you’re not alone.

With the ‘pre-AI’, just wanted to give context that I found it long ago and that afaik it’s made with some simple markov generator. Nowadays you could ask an llm to keep generating for you bogus git manpages.

> --combine-cruft-below-size

Couldn’t tell for a moment if it’s a real post, or whether I got linked to the (pre-AI era) GitHub manpage generator :)

https://git-man-page-generator.lokaltog.net/

I even occasionally get audio issues on Netflix (AppleTV plugged into Samsung 4k oled tv), which I assume are due to some kind of DRM, though never dug into it. Sometimes when switching inputs (firetv stick / AppleTV stick), or switching content on AppleTV between different apps, the Netflix content audio just stops working. All app UI sounds work correctly, but no audio once you hit play. Toggling the AppleTV audio settings a few times between dolbyAtmos and standard stereo usually brings it back, so I assume it has something to do drm on the audio tracks, but if anyone has other ideas lmk

Oh nice, thanks, just set this up! Though seems like I needed to do enable it separately for the web version and iOS app.

HomeDepot’s is even crazier. You input just your email and hit Next. Then a button appears to “Send magic link” to login via that annoying method. And then there is a tiny text below: “Want to use a different login method? Wait 10s…9s…8s…”. Only after 10s are you able to select a tiny text link “Use Password” to unlock using the password field

And on top of that, the session is probably gonna expire in less than day. I hate logging in to Anthropic because of this signin-email dance

This has been bothering me for a while now (few years maybe ?) - websites repeatedly expiring my login for who knows what reason. Sure, maybe I can understand for high value trading platform or something. But no - most mundane sites which most people wouldn’t at all consider sensitive, like HomeDepot / BHPhotoVideo / various online shops, will expire my session within like 24h - seriously, wtf. And significantly more sensitive sites, eg Meta/FB, are usually able to keep my auth for months/years.

I haven’t chatted about it with anyone, as I partly wasn’t sure if something in my setup has changed and whether I’m a minority that fell into some constant reauth bucket. Or whether most of site owners have slowly been using lower and lower auth expiration, causing me a bunch of frustration and friction.

Except you don’t seem to be able to filter by actual duplicates, it shows you both exact dupes and visually similar matches

Oh yeah? That’s surprising, I’m doing something similar - with esp32s from M5stack, specifically their SCD41 sensor for co2 (same as in the device in the post), with an AtomS3 to drive it, and it’s super responsive and I think very accurate. Perhaps you have the scd40? It’s rated for high accuracy up to 2000ppm, while scd41 maintains accuracy up to 5000ppm.

Using with esphome and homeAssistant, I grab the data every 5 seconds (I did 1s too but seemed unnecessary), and push it to HomeAssistant. The moment I turn on the oven, or open a window, or just enter the room after the night from the bedroom, the graph moves very very visibly and very quickly. I like this setup much more than my uHoo Air sensor.

I use LosslessCut for simple edits like this, works pretty well. Allows you to merge and split without reencoding.

https://mifi.no/losslesscut/

It does look like it comes with a few other perks that would normally cost a bunch too, specifically, 30TB of Google drive storage

Just tried it on chrome incognito on iOS and do hit this 429 rate limit :S That sucks, it’s already bad enough when GitHub started enforcing login to even do a simple search.

Is that kind of refresh rate even supported by current hdmi/displayport standards and graphics cards? Can I realistically expect to - for example - make a 4k hdr video at 500 fps, and play it back on that screen, connected to something like a newest macbook pro or a desktop with nvidia gpu, and actually get the correct output (verifiable by a high fps camera for example) ?

Fyi anything on AIStudio can be used for training and/or review, do not upload anything sensitive there.

If you have a FireTV stick or something Android based for your TV, I can recommend SmartTubenext for browsing/watching YouTube.

I still use AppleTV for pretty much everything else, but got a firetv stick just to use that. https://github.com/yuliskov/SmartTube

Damn, 2k tabs, I’m not familiar with sideberry but with 2k of them are they even right to be called native ‘tabs’ or do they function more like bookmarks?

> computational costs are close to zero

Is that true for normal Firefox tabs (I usually use chrome/safari) ? Wouldn’t each one still use up some memory, keep any background tasks running etc. If some tab starts playing audio how do you even find it?

AFAIK, The newer models for image gen like this OpenAI one, don’t actually use the normal diffusion process (image generates all at once from blurry to finished), but use transformer architecture where the full final image is generated from top to bottom, as a stream of ‘tokens’.

That’s why when you generate an image in chatgpt nowadays, it will start displaying in full resolution from the top pixel row and start loading towards the bottom.

Was just browsing earlier today for some potential upgrades from my GoPro10 and was looking at these too, that’s disappointing, guess wont be looking at these.

I have both too! It’s a really fun talking piece when I have friends over.

At first thought you are gonna talk about how various LLMs will gaslight you, and say something is true, then only change their mind once you provide a counter example and when challenged with it, will respond “I obviously meant it’s mostly true, in that specific case it’s false”.

Probably better to put in a separate restricted vlan. Who knows, maybe one day some piece of firmware code will kick in to start rotating MACs to lie who the device is to get a connection.

I used atop sporadically at Facebook to debug performance issues. I actually learned about it there, was I think on all the machines. This was bunch of years ago, so not sure if it still is there fleetwide, but it was really helpful to get a past granular view of what happened on the machine on some exact second few days ago where error rate metrics indicate a particular host was struggling.

I’ve started using Atuin based on a recommendation from here a few weeks ago. I generally like it, but feel like the fuzzy matching is bad - maybe I either misconfigured something or nobody else cares about it?

If the command I want to match against is within the last few days or so, the fuzzy match performs ok, but anything from months ago which I know for a fact is in the history, it just doesn’t match it and instead shows irrelevant items (which technically I’m sure according to some algorithm are correct). Another one is I often have to put a ^ to ensure the ‘curl’ I’m searching for starts looking for matches at beginning of line.

Often times I still fall back to using fzf on ~/.zsh_history which seems to get what I want. There is this thread which is same issues I’m seeing, and I did try smart_sort=True but it’s still not great.

https://forum.atuin.sh/t/understanding-atuin-history-search/...

Ah I see, yeah makes sense. I wasn’t aware that this is a feature in Passwords, (and even if I did I probably still would rather go manually to the website).

Since it pops up a web view which I presume is webkit/safari, I wonder if the Safari setting “Not Secure Connection Warning” (which you should set to on), is correctly applied to the view. Obviously it’s a bug they used http in first place, but this would have helped.