Depending on the size of the blob you might not want to use `WITHOUT ROWID`. I had a similar table with uuid keys and blobs of 40kb size on average and `WITHOUT ROWID` really tanked performance.
It sends very weird signals when the EU will fine an American company over some data moving in a direction they don't like while at the same time EU governments will allow home grown companies to de facto force people into using products from those same American companies all while lecturing us about duopolies and privacy only to re-enforce those same problematic patterns. It is absurd.
All countries are more lenient with their own companies. Remember who started grounding the 737 Max? It wasn't the FAA, an otherwise highly respected organization.
Who is paranoid about Chinese routers while spying on everyone? Etc.
No, it's not weird at all. They're each just an outcome of two different regulatory philosophies about how to protect users. We might want both because we're coming from an individualist hacker mindset, and thus see them as similar issues. But this is not how regulatory environments work (unfortunately).
From what I can surmise, the German/EU philosophy is more of a closed world approach - accepting that companies will keep control, then government regulates the companies to stop the companies-with-control from causing harm. If you don't like the harm, your recourse-focus is to petition the government to stop it (eg GDPR). Whereas the US philosophy is more open world - once someone "chooses" to patronize a company, then the company is free to do whatever they want. Your recourse-focus is to stop using that company.
They both have shortcomings and glaring loopholes, of course.
In Germany I spent exactly 0 dollars/euros on books in university. We got access to a huge amounts of ebooks through the university network and profs never required a specific one and would just recommend a few. One of these was always available. This free access was of course covered by the 250 euros per semester tuition…
But yes, standards and certifications are horrible.
I would understand if they didn't backup anything with FILE_ATTRIBUTE_RECALL_ON_DATA_ACCESS set (on Windows). Everything else makes their product worthless.
Since all input is run through a tokenizer, I would expect the tokenizer space doesn't change a lot between one trained on uncompressed vs one trained on compressed data.
Didn't read the paywalled article. But imo if it's not worth listening at 1x speed, it's not worth listening at all.
If it's entertainment like audiobooks, it's for my enjoyment, so why should it be faster than 1x? If it's for education, everything faster than 1x makes it hard to understand, and if there is too much filler content, I will find another source to learn about it.
I use 1.5x-2x for stuff I have to listen too, but don't enjoy.
It would be a non issue, if the options are not just allow/disallow, but also offer to provide random/fake data, so bad apps would not actually know if the permission is allowed or if the data is fake.
