I meant your website and resume that you linked there. If this is not some weird thought scenario, pointing her to the public forum you use to discuss alternatives to her simply talking to whoever's responsible for her comsec doesn't seem like the greatest idea. If this is an actual problem it's not yours to solve and certainly has solutions that aren't this... convoluted.
>If this is an actual problem it's not yours to solve and certainly has solutions that aren't this... convoluted.
Yeah, I took that view for a while. For example she gets my physical mail, so in theory I could just send her a burner phone myself, they could be like here this is personal you can't take it inside and she can say no problem and have the phone. But I took the view that this is really not my problem to solve.
Terribly sorry to harp on what the others in this thread have said but it seems like you've either been negligently briefed (due to the implausible scenarios you're describing and that you're posting this tied to your real identity, their first name on your website - assuming that's more real than the ChatGPT Ph.D. and other random entries there -, and your previous HN submission) or you're being lied to.
This is from 21, not really news, and the paper version on arxiv and published at NeurIPS have quite a few citations. No one's suppressing this, people that don't reflect on their datasets or how they use them just either don't care or fail to acknowledge they're actual issues.
IANA AI developer but have been looking into this in detail recently for other purposes. I was puzzled at the lack of info about "books" and when searching for detail (in what I believe was a reasonably diligent manner) found a very surprisingly small amount of it. I assumed there would be more knowledge and did ask for it here. So now I will go look up those papers to get a better sense of things. Thank you for the tip.
I note neither this paper nor any discussion of "BookCorpus" or even "book corpus" has appeared on HN previously.
Addressing "Documentation Debt" in Machine Learning
Research: A Retrospective Datasheet for BookCorpus, 2021, Jack Bandy and Nicholas Vincent
Would you please refrain from creating new accounts to post duplicates of this?
That's in the guidelines as well. HN has an e-mail address down at the bottom if you want to get in touch with them to potentially get this one changed.
There are only official persons/entity accounts on that instance, if you're thinking about the global feed that's just a composition of people those accounts follow.
It's not, especially on the weekend. The freight delays might be somewhat of an issue over the next few days but on the passenger side it's a mild inconvenience at best. It disrupted the rail network country-wide to some degree but the public rail infrastructure in Germany isn't really known for it's reliability anyway.
German rail not known for its reliability... really?! Perhaps not in Germany. But in most other countries, German rail is perceived to be very reliable. And compared to rail in most other countries, German rail is very reliable!
Compared to e.g. Japan and some other countries long distance trains in Germany are _very_ unreliable.
Compared to some other countries it is extremely reliable.
Local train service reliability differs quite a bit depending on region and sometimes in region which exact train you take.
And perception of reliability can differ from reality, e.g. Berlin people always complain about how unreliable their public transportation is, but due to high frequency of the transportation and a tight nitted 3-4 layer transportation net it is in effect really not that bad. Sure there are a few places where you e.g. reliable miss a follow up connection. But you can calculate that in and with delay times included it is still better then what you find in many other places in the word. So uh, perception is it's a catastrophe, but practice is it's far behind Tokio but not that bad at all in general (for most connections).
Amusingly a transit system has to be pretty decent to be bad - if it is actually horribly bad people don’t treat it at all reliable so they don’t really notice late trains, because getting a train at all is a miracle.
That's basically the case here in Australia (I live in Sydney, I believe the standard is about the same in other cities). My train line has been completely closed for the past week, for "routine trackwork" (and it's not even Christmas / summer holidays). Every weekend one or more lines are closed for trackwork. Throughout this year, trains have been delayed or cancelled, often with little notice, due to workers going on strike. And I've lived in Sydney for most of my life, and it's always been like this. Can't rely on public transport.
The non-YouTube authoritative copy of the recording includes the other audio track(s) and sub(s), typically (for at least most main track talks of 33c3/34c3/35c3/36c3) including the live English/German dub (whichever the speaker doesn't use) courtesy of a volunteer at the event.
While the GSM-R system belongs to the rail system which is dubbed critical infrastructure much of it is either publicly documented and/or shared between a rather large number of operators. So far I didn't read anything that would indicate that this disruption required more than somewhat specialized domain knowledge and a bit of experience in the industry.
Assuming you want to stick with the big cloud providers just sign up and test them on your content, they all have straightforward libraries.