Yup, that “what can I start in 10min” question really ruins a lot of productivity.

I don’t think that is true. Not much at Google really bought into the UUID hype. At least not for internal interfaces. But really there is no difference between a UUID v4 and a large random number. UUID just specifies a standard formatting.

It is true, don’t do it.

Even at huge companies like Google, lots of stuff was keyed on your email address. This was a huge problem so Google employees were not allowed to change their email for the longest time. Eventually they opened it up by request but they made it very clear that you would run into problems. So many systems and services would break. Over time I think most external services are pretty robust now, but lots of internal systems still use emails (or the username part of it) and have issues.

IIUC Google accounts now use a random number as the key. But there are still places where the email is in use, slowly being fixed at massive cost.

I thought that was the problem at first too. But unless there are fields that are searchable but not visible at all to end users I have definitely found many cases where the term (and no stemmed version of it etc…) was in the listing.

Not to mention that Amazon search is happy to ignore most of the words in your search. So you end up sorting through pages of results that don’t match. Absolutely infuriating and one of the reasons that Amazon is my last choice now. Someone decided that it was unacceptable to show “no matching results” and lost my business.

What I do is take a capture of the page, then if they haggle on the refund I can clearly show that the product is not the one I ordered.

I don’t agree. As a single counter example of many YouTube has a huge wealth of information and content.

Maybe that value isn’t worth the ads, that is much harder to say for certain. But it is clear that there is some valuable information on some sites that are supported by ads.

Most credit card issuers don’t issue credit cards to random apps by solo developers.

I would pay a lot of money to see Nintendo’s conniption over having to allow home brew and non-approved software on their game consoles. I would love to release emulators for older Nintendo consoles for the Switch so that they don’t get to keep charging people again to play old games on newer consoles.

1. I can usually pull out my phone faster than taking a card out of my wallet.
2. Phone-based cards typically have significantly higher limits than physical cards. (I can tap hundreds of dollars with my phone, only about $100 on my card.)
3. The phone needs to be unlocked which is safer than the card which just needs to be tapped with no other authentication.
4. One less thing to carry around.

Because to implement this you need to negotiate with individual credit card issuers. Basically how this works is that your phone is being issued a virtual card with the keys locked inside the phone’s HSM. Then it can be used to make NFC payments just like any physical card. So you need 1. contracts with many card providers, 2. card issuance processes with these providers 3. huge amounts of compliance bureaucracy. At the end of the day it isn’t really worth it unless you are a huge company and expect to have tons of users or see it as an essential feature of your phone OS.

It’s not “inherently insecure” at least not to that degree. (Once could argue that lack of E2EE is insecure.) If you stand up an unrelated instance you shouldn’t be able to access private messages that don’t relate to an account on your instance. So only bugs in your instance, or your conversation partner’s instance, will be able to leak those messages.

They’ll brick your device if a part can’t be verified so that isn’t much different they destroying. Maybe they don’t require repair shops to hand over personal info, but they do require device identifiers so I wouldn’t be surprised if that is basically identical.

Is this worse? It sounds pretty similar.

How exactly does Samsung police this? Surely the repair shop could just… not tattle?

Well there is a contract in place and there would be consequences for not upholding the agreement. Sure, they could probably get away with it for quite a while. But it likely isn’t worth the risk, they would rather just out Samsung as being a piece of shit and go on their merry way.

It would be pretty easy to catch this as well. Samsung can just occasionally submit a phone with a known third party part for repair and see if the expected report comes in.

I’m sure some people will demand it. But for 99.9% of the population you don’t need 1000Hz content. The main benefit is that whatever framerate your content is it will not have notable delay from the display refresh rate.

For example if you are watching 60Hz video on a 100Hz monitor you will get bad frame pacing. But on a 1000Hz monitor even though it isn’t perfectly divisible. the 1/3ms delay isn’t perceptible.

VRR can help a lot here, but can fall apart if you have different content at different frame rates. For example a notification pops up and a frame is rendered but then your game finishes its frame and needs to wait until the next refresh cycle. Ideally the compositor would have waited for the game frame before flushing the notification but it doesn’t really know how long the game will take to render the next frame.

So really you just need your GPU to be able to composite at 1000Hz, you probably don’t need your game to render at 1000Hz. It isn’t really going to make much difference.

Basically at this point faster refresh rates just improve frame pacing when multiple things are on screen. Much like VRR does for single sources.

I’m using Kagi. I find that it does a better job at finding “legitimate” sites rather than blogspam and content marketing. However I’m not sure I will stick with it a long time. I seems like it has mostly stalled and the team is getting distracted by making a browser, non-relevant AI (I have no problem with the few AI experiments tied to searching) and other side projects. We’ll see. I really hope that they pull themselves together and focus or it might not last. But for now they seem like one of the better options available.

Bing’s new “Deep Search” where it has some sort of LLM refinement iteration process has also been helpful sometimes. Probably the best AI search product I have seen, but definitely doesn’t replace most searches for me.

My problem is that I do red, blue, green then can’t think of any more clearly visible colours.

Algorithms are like AI but accurate, predictable and usually much faster.

Oh, flac fixes for HLS. I wonder if https://github.com/jellyfin/jellyfin/issues/8722 was fixed. I’ll have to try it out today.