radicality

Nice. At first glance looks cool, but then, oh my, some areas are really butchered :S Poor Old Town and Plac Zamkowy - I'm familiar with the area and at first was scanning trying to find it visually by the features I know are there, but then in this model it's very different (wrong color, many more floors in this model, a 20-story building where at most should be at most a 5-story, the old town castle looking like a gray residential 'blok' and not the prettier structure that it is!).

I look at that page in the last few weeks probably every day. Not sure they themselves know what exactly they are summing up or how much tokens they believe they should be charging - my usage recently went from ~48B -> ~51B -> ~47B -> ~44B -> ~46B. Meaning, they messing around with the tokens consumed math / rewriting the consumption history, surely should not ever go back down, and just making it go up/down however they like. Even the individual per-day counters seem to have changed for me without rhyme or reason, some past ones increased, some decreased.

I forget where, but I remember reading somewhere about how STOP is meant to actually be handled at the carrier level and do something to not allow future messages at all from the sender? Though maybe I’m misremembering?

This is amazing, pretty much every single thing I hate with all the AI-generated documentation nowadays, all in one repo! And conveniently organized across >20 separate files haha

> I’m a building automation and controls tech in Manhattan

Hi - sorry if this is not your expertise, but I thought I might as well try asking. I live in ~central Manhattan, in a ~40th floor apartment with a corner view with large windows meanings lots of glass through which sound will travel. For around a year now there's been an annoying high tone coming from somewhere nearby, I would suspect perhaps some large AC fans, or maybe generator coil whine? It's persistent, around both ~2587Hz and ~2475Hz peaks, so a quite high pitched tone that's been driving me insane. The moment I open a window it's just excruciatingly "there", and even with windows shut it's still faintly there in background.

I don't know how somehow no one else has complained/gotten it fixed in this long. I'm overall use by now to manhattan noise, but this kind of persistent high pitched tone is another level of noise annoyance. I know the general direction it's coming from, and have few ideas for which buildings/devices it might be based on just seeing from my window, but pinpointing this kind of tone is difficult.

Again, might not be your expertise at all, but if you know of how I could proceed with it, to either find the source, and/or report this somehow for an investigation (311 ? some other agency? ) ? Thanks upfront!

My macOS 27 m4 MacBook 128GB ram has in the last 3 days: completely locked up and went black screen (after reboot it had a stacktrace that windowserver crashed and didn’t come back) ; had a full on kernel panic, which later an AI agent analyzing the panic dump said it’s because the iOS simulator was somehow leaking resources and watchdog triggered a the kernel panic.

I think I would be more inclined to get it if it supported 2.5Gbs or preferably even 10Gbs.

I have Infuse app for playback on my Apple TV, and recently tried to play, from my local nas on 10Gbs link, some 4k ProRes footage. I forget if it even has the decoder for it, but just the bandwidth wasn’t even enough to play that. Honestly also from experience, playing from local network a 80GB BD remux is also kinda so-so. Random hiccups for me during normal playback. And just completely unreliable fast forward/backwards seeking.

Also single seat workspace user. Half the time my Gemini forgets I connected Gmail to it and hallucinates false instructions to enable it.

Truly crazy how much Gemini still hallucinates, especially things which its own Google documentation or just a basic web search away. I feel it’s mostly solved problem with chatgpt/claude , but for me, with Gemini it’s almost guaranteed it will just make stuff up

Ha, same experience here. It opened my Gemini and gave me images like “Design a terrarium with Gemini 4” - proceeds to pick 3.1 Pro from the picker because 4 just doesn’t exist in the dropdown. Typical Google.

I really can’t believe how bad the Gemini experience is. I have personal paid workspace account, and meant to have some better AI Google features I think.

So, I connected my Gmail to Gemini. 3.6 Flash kept telling me it doesn’t have any Gmail connnection, and then proceeded to hallucinate instructions for me to enable it, all bogus. At least 3.1 Pro realized that my Gmail is connected and then searched.

I wish it could do something what should be seemingly incredible simple task - sync and download ALL the email. But afaik Mail iOS can’t do it, Gmail ios app can’t do it , and no other mail app I tried seems to be able to, which seems absurd.

All I want is all the mail and attachments downloaded locally the way normal desktop mail clients work. Not the last 90 days of it (Gmail ios max) - all of it, I don’t care if it’s many gigabytes or more, i just want the full mail, so that something as full text local search of the email would actually work…

Have not tried yet, but did see in the readme it supports x402 protocol for payments, that’s nice.

1) double checking whether it actually does, does it? 2) Which chains are supported, and how’s the x402 being handled - your service self hosted or third party?

Not OP, but probably because they have no idea what they are doing or what CLIP even is. And the whole thing is likely just from one llm prompt, and then they dumped the whole thing onto GitHub, undoubtedly without even looking at any of the code or architecture.

And then you end up with image metadata processing code like this, which just by a cursory glance I’m sure has edge case bugs : https://github.com/allenv0/SCM/blob/main/screenshot-probe.js

Oh yeah? I’m absolutely against Ticketmaster and their whole monopoly and the whole platform sucks with all the scalpers, but, for this specific case I never had any issues and it works well for me. I’ve bought 2 tickets multiple times on Ticketmaster, and then ‘transferred’ one of them from the Ticketmaster UI to my partner, and then they ‘accept’ it on their account and the ticket appears there. This transfer is thankfully free.

On my paid personal Workspace account I’m even more behind. I see 3.1 Pro, 3.6 Flash (New). Can’t not laugh at that ‘New’.

Right yeah, I saw this announcement, opened my Fdroid app and clicked around very confused on how to get it to update :S

Oh, interesting, I feel like a year so ago I read something posted from openrouter team that they write that credits expire, but that they in actuality don’t expire them. But maybe I’m misremembering, or perhaps that’s changed for the worse in last few months :S

It has been at least a few months since I did some testing on my MacBook Pro m4, where I converted a bunch of camera jpegs to jxl losslessly, and yeah the jxl files were smaller, but the decoding/viewing experience on macOS was worse than with jpegs. Thumbnails were missing, and just browsing through a folder of them with QuickLook was unbearably slow compared to jpeg, perhaps at least few hundred milliseconds or even longer than a second, when the JPEGs were instant. Idk if that’s just Apples implementation not good yet, but that’s why I at the time abandoned converting all my JPEGs to lossless jxl

That’s great for the team. Though as a user of openrouter, I am worried that the much larger company will soon find ways to enshittify it :S

I’m not sure if it’s been redacted from there, but try to do a request for a full data export, it might have the detailed data. I overall try to do that for all online services I use around once every six months or so, lots of interesting data in them.

https://www.amazon.com/gp/privacycentral/dsar/preview.html

Scroll down on that page to “request all data”

Aliexpress is also surprisingly impossible to search, you gotta use external search engines, I guess it increases there sales :S More than half the time for me on AliExpress a very specific search term will just return completely unrelated things which are either trending or related to my past purchases.

Is the Costco shipping good / good inventory / similar or same as in store prices? Should I consider it if living in a city like ny?

I live in central nyc, been only a few times to a Costco, and had no idea they even do any shipping! I assumed that part of all those good prices is the smaller selection, large minimum quantities, and that you have to get yourself there with a car to get it in person.

Which black mirror episode are you talking about? White Christmas? S02E07 afaik does not exist - Series 2 had 3 episodes - https://en.wikipedia.org/wiki/List_of_Black_Mirror_episodes

Openrouter has video generation also. They do have Seedance which afaik routes via something called ‚atlas cloud’, but it did work.

Thanks, I’ll try Polycam too.

Did they literally just leave the water supply plant management software out available on the open internet? Hard to even call this a hack!

Ha, for a brief moment I thought VM=virtual machine and not voice mail and was scratching my head how you set that up and what the VM is doing with the call

At Facebook (i left >4 years ago) we used Phabricator where shipping stacked changes like this was table stakes. I still miss it. Some few random points, not really organized, i'm sure i'm forgetting bunch more benefits:

* I often used it as a way to make reviews significantly easier for the reviewer. If I split something more complex into smaller units where the small change seems obviously correct, it's much more likely i'll get a faster accept than with a very large change.

* Critically, once a unit of the stack has been approved, the approval will still stay with rebases of the whole stack.

* You can keep building up a stacs and once you think you have enough for whatever you are doing, ship the whole stack.

* Suppose you have a 'hole' in the reviews and you're missing a review for a middle part of the stack. You can still merge the bottom part up to where all are reviewed.

* Let's assume you're the reviewer: smaller changes is much easier to review. Also, each part of the stack builds incrementally on the previous one. With your example, suppose you're reviewing 'api changes' part of the stack, but it has failing tests because something is broken with the lower 'db schema changes' change - but otherwise, the 'api changes' diff looks fine - you can then accept it and this part of the stack is already reviewed. The author can now whenever fix their bottom-most change (db schema change) to have tests passing, and get someone to review that (perhaps someone completely different more familiar with db), and then rebase the 'api changes' diff, and _it will not_ require review because you haven't changed anything, just updated the diff underneath it. In this way you parallelized your approvals. Right now at work, using github, I have to serialize my PRs and can't effectively get things reviewed until the previous PR is approved.

It's been >4 years since I've been out of Facebook and now have to use Github, and to this day I miss phabricator and landing stacked changes. When I first left and had to work with github I hadn't realized that this, to me, very basic, workflow is _not_ a feature!

How did you actually end up doing it, like the full pipeline ? For example what app did you use to scan your apartment and orocesss it into a 3d model (which kind?) and then how do you actually do the planning (like which tools/software to plan your new space?)

And I ask for details because I have attempted something similar but wasn’t getting the best results. I was using ‘3d scanner app’ (which was good but was sold and is not recommended now), and that created pretty decent models, but like, they were still I think based on lidar and not super precise. I also wasn’t sure how to reconcile my actual floor plan with that 3d model the right way - like if the 3d model got some error built-up, the wall which I know to be 100 inches has to be that, and not 101 or 99, since for furniture planing an inch definitely matters. So I kind of got stuck here on how to actually generate some useful and as correct dimensionally as possible 3d model, into which I can later put furniture 3d models into. Thanks for any info!

True, I didn’t do those in detail. I vaguely looked at the output and thought “let’s see if that works at all”.

I do such more careful planning and AI handholding at work where it’s important the thing doesn’t break in prod, or doesn’t forget about some important use case.

But this was just purely for “fun” / homelab style type of vibe coding where I don’t want it to feel like work and write intricate specs and handhold the agent - point being the current SotA models still currently do need a bunch of guidance to not devolve into a code spaghetti, and it does end up feeling like ‘work’.