abhibeckert @ abhibeckert @beehaw.org

Posts

0

Comments

321

Joined

2 yr. ago

I didn't even know it was happening until the news stories about interest rates potentially rising on the day of the race.

ChatGPT 4 is estimated to use 700GB of “High Bandwidth Memory”.

… which will set you back about half a million dollars at current prices (which are high, because the manufacturers can’t keep up with demand). Or, you could just pay 20 bucks a month.

I don't think it's an issue. If your content is good, you should be able to find an audience and if you have an audience you'll be able to find sponsors. That doesn't have to be by directly reaching out to sponsors themselves, you can work with intermediaries.

Youtube obviously dominates the space right now but it's hardly the only viable business model. In fact I think it's better if content creators have more control than YouTube provides.

It shows a QR code on the screen, and 'sir' scans the code on their own phone without touching yours.

Also - you can customise the amount of data revealed, for example you might just provide proof that you're over 18 (an therefore allowed to buy beer) without letting them see all the other details on your license.

The brand I linked is sold nationwide - have you tried it? I don't think there's any other dairy like it in the world.

The cows are 100% grass fed, and the grass is incredible. The soil on those farms was built up by being rainforest for over a hundred million years - unfortunately the land was cleared before it could be protected under Unesco World Heritage laws, and the paddocks get rain almost daily. Sometimes they are literally in the clouds.

They do a decent job at the factory as well, but mostly it's the quality of the milk that makes it so premium.

Dunno about "the last update" or the current state in each region but as far as I know the default search engine in FireFox has varied over the years and has always depended what country you're in.

Baidu, Yandex and Yahoo are / have been the default in some countries. They made Bing the default for "1%" of users in a bunch of major countries recently to test the waters (and didn't take it further than that).

Google blocks traffic from Chinese IP addresses as a protest against censorship there, so nobody has Google as the default in that country.

They made a $40b profit last year. More than half their profits is a "drop in the bucket"?!

It's non negotiable. I'm always going to use every tool I can to block tracking networks.

If they respond by not letting me use the service... that's fine I'll find a competitor. But in my experience that's pretty rare. Usually they're happy to let people use their service even if they can't track some people.

The only real road block I've ever seen is occasionally a service will ask me to prove I'm not a robot. I rarely bother with proof - just close the tab and switch to an alternative that doesn't do that garbage.

As far as I know, patrol cars routinely scan the number plate of every car near them and do a license check automatically. The system will notify the officer if someone's license is expired/suspended/warrant for their arrest/etc. And if the officer does pull you over, they've often already checked your license by the time they walk up to your door.

Even before those existed, it was common practice to use a radio to ask someone back at the police station to run the number plate of a car before pulling them over.

I've had a cop refer to the owner of the car I was driving (e.g. my partner's car) by name when they've pulled me over for a random breath test/etc.

The M1 is around 25% more efficient than the M2.

No that's not right. The M2 is far more efficient. Third party tests report he M2 MacBook Air lasts up to 3 hours longer than the M1.

Yes, it draws more power under peak load... but it more than makes up for that with better performance allowing it to return to an idle state more quickly. Give an M1 and an M2 the same task, and the M2 will draw less power to get the task done.

Well, except Feta - the Greek stuff is amazing.

You can buy Australian made "Greek style" Feta. Try this one: https://www.mungallicreekdairy.com.au/products/organic-fetta-in-brine/

I'm on an M1 MacBook Air - Anandtech measured between 11 and 17 watts with an M1 Mac Mini.

However, the Mac Mini has an excessively large cooling system for the chipset it runs (before Apple Silicon, they sold the same Mac with an Intel i7 that turbo boosted to 4.6Ghz).

The MacBook Air has basically no cooling at all and it definitely throttles under high load. It's still fast enough to get 60fps with good graphics settings while throttled for the games I play - I'd say it's about on par with my gaming PC that has an entry level Nvidia GPU, but there's no way it's drawing as much power as in Anandtech's testing on an actively cooled chip.

Based on the battery life I'm getting, I'd guess it's drawing somewhere around 8 watts on average while playing games. It's a very efficient chip... it draws 0.2 watts while idle according to Anandtech testing. Remember, this family of chips started life on devices with a 10Wh battery and the MacBook Air isn't much faster than an iPhone.

Likewise even macs will start to drain quite a bit when say watching an hd video 1.75x speed, or playing a video game

That's not my experience. I can play demanding games (CPU/GPU flat out) for several hours on battery on my Mac, and it only has a 50Wh battery.

With "normal" use I get about 18 hours on a charge.

I generally charge it overnight, like a phone, except I don't do it every night. I often don't even have access to a charger for days at a time, a laptop charger isn't part of my normal travel kit. If I notice the battery "running low" that means I need to find a charger in, like, five hours time.

The high end MacBook Pros, with a 12 core CPU and 38 core GPU... yeah those can draw a lot of power. In fact they even drain the battery while plugged into a charger if you really push them. But I don't think of those as "proper" laptops. They're more like a portable desktop.

Bing (and therefore DuckDuckGo, which is what I generally use and is a wrapper around Bing) is definitely worse than Google especially for dev research, but it's not as good as it used to be.

I do use Google for a lot of my dev research, and they seem to be losing the ongoing war against spamers flooding the internet with garbage content.

Websites like reddit (and beehaw) are somewhat of an oasis – actively moderated with absolute garbage content deleted straight away and questionable content at least has replies where people have pointing out if they think it's wrong. If (when?) Reddit goes away, that's a whole bunch of really good content that will suddenly disappear from google results, which will be sad.

PS: If you haven't already, try buying a subscription to ChatGPT+ and use GPT4 as the first place you go for all your LUA/React questions. I find it gives far better answers than Google for most things. You can sort of dip your toes in the AI waters by trying Bing Chat.. but it's nowhere near as good for code as ChatGPT+.

If by document you mean “any kind of data structure”, then yes, those are documents

Yep — that is what I mean by documents, and it's what I meant all along. The beauty of documents is how simple and flexible they are. Here's a URL (or path), and here's the contents of that URL. Done.

But then the term becomes meaningless, as literally anything is a document.

No, because you can't store "literally anything" in a Postgres database. You can only store data that matches the structure of the database. And the structure is also limited, it has to be carefully designed or else it will fall over (e.g. if you put an index on this column, inserts will be too slow, if you don't have an index on that column selects will be too slow, if you join these two tables the server will run out of memory, if you store these columns redundantly to avoid a join the server will run out of disk space...)

Sure, but then finding that document takes 5 minutes

Sure - you can absolutely screw up and design a system where you need to read millions of files to find the one you're looking for, but because it's so flexible you should be able to design something efficient easily.

I'm definitely not saying documents should be used for everything. But I am saying they should be used a lot more than they are now. It's so easy to just write up the schema for a few tables and columns, hit migrate, and presto! You've got a data storage system that works well. Often it stops working well a year later when users have spent every day filling it with data.

What I'm advocating is stop, and think, should this be in a relational database or would a document work better? A document is always more work in the short term, you need to carefully design every step of the process... but in the long term it's often less work.

Almost everything I work with these days is a hybrid - with a combination of relational and document storage. And often the data started in the relational database and had to be moved out because we couldn't figure out how to make it performant with large data sets.

Also, I'm leaning more and more into using sqlite, with multiple relational databases instead of just a single database. Often I'm treating that database as a document. And I'm not alone, Sqlite is very widely used. Document storage is very widely used. They're popular because they work and if you are never using them, then I'd suggest you're probably compromising the quality of your software.

Being 20, I'd assume they have friends who are younger than them, and therefore are well and truly in touch with what "the kids" are doing.

Wait, what? Your comment doesn't read like sarcasm at all.

A patch isn't (yet) available.

But a workaround is. Configure your password manager (or switch to another password manager) so it doesn't automatically fill usernames and passwords as soon as you open a webpage. Set it to fill the credentials when you click a button or hit a hotkey.

And after this security flaw is fixed? Leave the settings like that. Because this isn't the first time autofil has resulted in a major compromise and it won't be the last time either.

PS: this speculative execution bug was reported to Apple a very long time ago and there are experimental settings you can change to test the fix... but they might be buggy. Modifying your password manager's behaviour will not be buggy. The setting is:

defaults write com.apple.safari InternalDebugProcessSwapOnCrossSiteWindowOpenEnabled 1

... they don't pay for the theft. They are businesses not charities.

It's ordinary aussies who pay. Coles and Woolies set their prices high enough to cover the cost of the stolen items.

This is a vicious cycle. Cost of living goes up, theft goes up, cost of living goes up, theft goes up, etc. Eventually theft is so bad the store can justify hiring police officers to supervise the self checkout section all day long.

And who pays for the police officer at every coles/woolies in the country? Yep. Ordinary aussies pay. And the current price is $116.19 per hour (in NSW). And the people who are arrested? We pay for their arrest. We pay for the laywer who argues against them in court. We pay for the legal aid defence lawyer who defends them in court. We pay for the judge. And we pay for the jury. And we pay for their prison costs... fuck. Wouldn't it be better if they just didn't steal that loaf of bread?

These are the things the reserve bank is trying to avoid when inflation is too high. It can lead to the collapse of our entire society. I'm not talking hypotheticals either. Go to a woolies or coles in the northern territory. There's often a police officer standing at the checkouts and stopping people to search their bags. And in some parts of the NT you can pay 70 bucks for a tin of instant coffee (theft isn't the main reason for that, but it's part of the reason - people steal a lot of food when it costs that much).

If you're hungry, there are charities who will give you food. Don't steal food.

The real concern would be adding the watermark to the real thing, to let it slip through the cracks. However, not only would this be computationally expensive if it was properly implemented,

It wouldn't be expensive, you could do it on a laptop in a few seconds.

Unless, of course, we decide only large corporations should be allowed to generate images and completely outlaw all of the open source / free image generation software - that's not going to happen.

Most images are created with a "diffusion" model where you take an image, and run an algorithm that slightly modifies it. Over and over and over until you get what you want. You don't have to (and commonly don't - for the best results) start with a blank image. And you can run just a single pass, with the output being almost indistinguishable from the input.

This is a hard problem to solve and I think catching abuse after it happens is increasingly going to be more difficult. Better to focus on stopping the abuse from happening in the first place. E.g. by flagging and investigating questionable behaviour by kids in schools. That approach is proven and works well.