I watch a stupid amount of YouTube so I pay for YouTube Premium. I wish that meant I could have premium features like disabling shorts, disabling those annoying themed sections that keep popping up (right now it’s the olympic games), a search function that actually searches for what I want instead of shoving more suggested videos in my face, and changing every “not now” button into “don’t ever ask again”.
Despite the fact that I’m a voracious consumer of YoutTube videos and a long-time paying customer, I have to accept that I am not the target audience. They want passive users who endlessly watch whatever gets put in front of them so that they never leave the app. If there was a respectful alternative that worked well with iOS and AppleTV I’d gladly pay for that instead.
I also pay for it and watch it on my android TV. And I completely agree with you. But for now it’s my main source of TV and I mainly follow channels I subscribe. I don’t listen to music on YT, i do follow some live music channels
Yeah I have paid for YouTube Red since it was YouTube Red.
I’m pretty sure they don’t want want me anymore cause they have made my whole experience worse and trying to get more and more money and attention out of me.
I really should stop paying too but they are literally making it so there is no alternative and I hate paying on patreon but watching on YouTube with ads
Any S3-compatible object storage solution would do, plus it’s immensely used in enterprise so a lot of software supports backing up to S3 objects. Operates entirely over HTTPS.
They said something similar with detecting cancer from MRIs and it turned out the AI was just making the judgement based on how old the MRI was to rule cancer or not, and got it right in more cases because of it.
Therefore I am a bit skeptical about this one too.
Using AI for anomaly detection is nothing new though. Haven’t read any article about this specific ‘discovery’ but usually this uses a completely different technique than the AI that comes to mind when people think of AI these days.
The problem is that it refers to so many and constantly changing things that it doesn’t refer to anything specific in the end. You can replace the word “AI” in any sentence with the word “magic” and it basically says the same thing…
It’s literally the name of the field of study. Chances are this uses the same thing as LLMs. Aka a neutral network, which are some of the oldest AIs around.
It refers to anything that simulates intelligence. They are using the correct word. People just misunderstand it.
According to the paper cited by the article OP posted, there is no LLM in the model. If I read it correctly, the paper says that it uses PyTorch’s implementation of ResNet18, a deep convolutional neural network that isn’t specifically designed to work on text. So this term would be inaccurate.
or a pattern recognition model.
Much better term IMO, especially since it uses a convolutional network. But since the article is a news publication, not a serious academic paper, the author knows the term “AI” gets clicks and positive impressions (which is what their job actually is) and we wouldn’t be here talking about it.
Well, this is very much an application of AI… Having more examples of recent AI development that aren’t ‘chatgpt’(/transformers-based) is probably a good thing.
Op is not saying this isn’t using the techniques associated with the term AI. They’re saying that the term AI is misleading, broad, and generally not desirable in a technical publication.
Op is not saying this isn’t using the techniques associated with the term AI.
Correct, also not what I was replying about. I said that using AI in the headline here is very much correct. It is after all a paper using AI to detect stuff.
Haven’t read any article about this specific ‘discovery’ but usually this uses a completely different technique than the AI that comes to mind when people think of AI these days.
For the image-only DL model, we implemented a deep convolutional neural network (ResNet18 [13]) with PyTorch (version 0.31; pytorch.org). Given a 1664 × 2048 pixel view of a breast, the DL model was trained to predict whether or not that breast would develop breast cancer within 5 years.
The only “innovation” here is feeding full view mammograms to a ResNet18(2016 model). The traditional risk factors regression is nothing special (barely machine learning). They don’t go in depth about how they combine the two for the hybrid model, so it’s probably safe to assume it is something simple (merely combining the results, so nothing special in the training step). edit: I stand corrected, commenter below pointed out the appendix, and the regression does in fact come into play in the training step
As a different commenter mentioned, the data collection is largely the interesting part here.
I’ll admit I was wrong about my first guess as to the network topology used though, I was thinking they used something like auto encoders (but that is mostly used in cases where examples of bad samples are rare)
They don’t go in depth about how they combine the two for the hybrid model
Actually they did, it’s in Appendix E (PDF warning) . A GitHub repo would have been nice, but I think there would be enough info to replicate this if we had the data.
Yeah it’s not the most interesting paper in the world. But it’s still a cool use IMO even if it might not be novel enough to deserve a news article.
It’s really difficult to clean those data. Another case was, when they kept the markings on the training data and the result was, those who had cancer, had a doctors signature on it, so the AI could always tell the cancer from the not cancer images, going by the lack of signature. However, these people also get smarter in picking their training data, so it’s not impossible to work properly at some point.
That’s the nice thing about machine learning, as it sees nothing but something that correlates. That’s why data science is such a complex topic, as you do not see errors this easily. Testing a model is still very underrated and usually there is no time to properly test a model.
Restic. You just need a s3-compatible object store in k8s to make it work. All else is handled by the client. That’s what I used (not with k8s), with resticprofile.
I also heard Borg is a great alternative, but never try personally, nor how it works. Both are CLI only I believe.
I second restic and i use it with wasabi. Haven’t touched it in years. Do a fire still once a year and it’s worked perfectly. I even basically cloned my proxmox setup the other day.
Even though most people don’t agree with the stats, I think it makes sense because Arch users are never satisfied with their setup. It could cause many of them to choose an average number.
So, part of it is, that if one attractive person finds you attractive, other people are more likely to find you attractive.
Humans are an odd bunch, like that.
The other side of it, is that a truly astonishing number of people go through life being told- and believing- that they need to have a “serious” relationship to be happy/successful/not a total waste, and then there’s the biological hardwiring that more or less confirms this is true.
Sometimes, to the point that they jump from one shitty relationship straight into the next.
It is well known that, when you’re single nobody wants you, but the second you manage to fool someone into relationship with you, you’re the most wanted man around.
I’m just speculating here based on my own experience, but I wonder if part of it is also something subtle in your attitude that others pick up on.
When I met the man who is now my husband, I wasn’t even looking for a relationship. I was just enjoying my life, exploring new hobbies, and was in a good place mentally – that is, I was fine if I ended up single for the rest of my life.
I imagine married people also tend to be in a similar frame of mind where they aren’t looking for a relationship, and maybe there’s a confidence that comes out of that that is attractive to others.
As someone in an ENM relationship, no. If you’re an awkward nerd who happened to hit it off with one person, you’re still an awkward nerd when trying to hit it off with a second.
If you’re a man, put on a wedding ring and go out to the bar and see for yourself. This may work if you’re a woman too but I can not claim to have done the research on that one.
I have my dad’s old wedding ring, I may try this. But the issue becomes I don’t want to have a serious relationship with a woman who only likes me because she can “destroy my family” essentially (whether the “family” is fake or not.) Something seems seriously wrong with those people.
So measuring from full fore to aft might give inconsistent results among other similar variables for the examined species, such as location, age, and weight. Going booper to pooper might give more reliable data, if the tails are often snipped at the tip by other reptiles, predators, disease, rogue mohels, swamp boat propellers, or hastily closed doors.
This is my best guess, I’m not lizardologist or a measuresmith.
I’m also not an expert, but that was my thought, too.
More than that, even if a tail is undamaged, including it is not giving useful imformation because tail size can vary out of proportion to the main body and is pretty standard for other animals as well. For example, no one is measuring a horse to include the tail length, nor a dog, cat, and generally not a bird, either.
That said, I expect an news story about alligators on the golf course or catching invasive snakes to measure the whole body for the NEWS story and let the experts worry about the booper2pooper length in their own space.
I mean, it’s not the exact same, but human fetuses can be measured similarly with a “crown to rump” length. I have to notate this in my reports (but I include other measurements too like hand/foot length and “crown to heel” length).
I’m done recommending stuff, because my use case is not necessarily your use case. I can only tell you that Brave is the sweet spot for me, at the moment.
Thanks for the link! I was trying to figure out the Lunduke complaints (I don’t know if I’ve ever heard of this guy before). This video did a solid job of catching me up.
I used to enjoy listening to him on YouTube simply because he didn’t yell in his videos like every other YouTuber. He had a bit of a Bob Ross vibe to him, but then he went off the deep end.
Or they dump their entire 6gb SQL database, customer info and all, into a SQL file that you have to load into a mariadb docker container when you just needed a subset that you were going to turn into csv anyway ☺️
It really depends on the machine that is running the code. Pandas will always have the entire thing loaded in memory, and while 600Mb is not a concern for our modern laptops running a single analysis at a time, it can get really messy if the person is not thinking about hardware limitations
Then I guess that the meme doesn’t apply anymore. Though I will state that (from my anedoctal experience) people that can use Panda’s most advanced features* are also comfortable with other data processing frameworks (usually more suitable to large datasets**)
*Anything beyond the standard groupby - apply can be considered advanced, from the placrs I’ve been
**I feel the urge to note that 60Mb isn’ lt a large dataset by any means, but I believe that’s beyond the point
What do you mean not optimal? This is quite literally the most popular format for any serious data handling and exchange. One byte per separator and newline is all you need. It is not compressed so allows you to stream as well. If you don’t need tree structure it is massively better than others
I think portability and easy parsing is the only advantage od CSV. It’s definitely good enough (maybe even the best) for small datasets but if you have a lot of data you need a compressed binary format, something like parquet.
But which separator is it, and which line ending? ASCII, UTF-8, UTF-16 or something else? What about quoting separators and line endings? Yes, there is an RFC, but a million programs were made before the RFC and won’t change their ways now.
Have you heard that there are great serialised file formats like .parquet from appache arrow, that can easily be used in typical data science packages like duckdb or polars. Perhaps it even works with pandas (although do not know it that well. I avoid pandas as much as possible as someone who comes from the R tidyverse and try to use polars more when I work in python, because it often feels more intuitive to work with for me.)
I used to export my pandas DataFrames as pickles, but decided to test parquet and it was great. It was like 10x smaller and allowed me to had the the databases on a server directory instead of having to copy everything to the local machine.
Ah I was trying to point out that CSV is the inefficient format. Reading a large amount of data from a more efficient format like parquet is more likely to cause trouble because the memory required can be more than the file size. CSV is the opposite where it will almost always use more disk space than is required to represent the data in memory.
tert-butyl lithium. Ignites on contact with air. Often used in conjunction with flammable solvents, so large fires and explosions are possible when working with large enough quantities.
As far as safety SOPs go, nearly any chemical spill of a large enough quantity warrants evacuating the area in my chemical safety plans. For some institutions, this is as little as 1 liter or 500g of material. This can obviously be overkill if you spill something that is relatively inert and non-toxic such as water or NaCl.
i’m not a chemist so take my words with a grain of salt -
google says the most dangerous bases can cause skin and eye damage, and be very flammable if they were to dissolve aluminium (two different chemicals). So I guess worst case scenario you open the windows, lock the room, and come back with protective equipment to clean up your mess
you probably wouldn’t handle something, that if dropped, would be dangerous enough to need a whole building evacuated outside of a dedicated room without wearing a full hazman suit and adhering to additional 100 precautions and safety measures
unless you’re the sort of guy to use a screwdriver to play with the demon core, but that’s not a liquid chemical base and hopefully won’t happen again
you probably wouldn’t handle something, that if dropped, would be dangerous enough to need a whole building evacuated outside of a dedicated room without wearing a full hazman suit and adhering to additional 100 precautions and safety measures
You underestimate academia.
Also, going through the old chemicals of some academic labs can lead to having the bomb squad called because they didn’t dispose of an unstable reagent 30 years ago.
Edit: Turns out it reacts explosively with water (so it itself can’t form the aqueous solution necessary for the concept of pH to be applicable) and decomposes into two different strong acids (hydrofluoric acid and hydrochloric acid). So yeah, not a candidate for our mystery base.
Yes, this is the kind of substance that would promptly react to protons, what would be a base-like behavior if it also didn’t promptly react to hydroxyl, what would be an acid-like behavior.
But given that it will consistently turn water into plasma, I guess it technically has a PH of 0.
I was curious as well and in this article the only mention of dangerous bases is tert-Butyl lithium (“t-BuLi is very pyrophoric, it readily reacts with air catching fire, that’s why it has to be handled and stored with very special care, always under a protective inert atmosphere of pure nitrogen or argon”). But in that case you couldn’t just drop it on the ground outside of a vent?
kbin.life
Newest