Hacker Newsnew | past | comments | ask | show | jobs | submit | Hexcles's commentslogin

Now that OpenAI basically abondoned Atlas, which had a truly novel architecture, let's see if this more obvious technical route (not saying this to diminish the achievement!) would gain traction.


It's nice to see wpt.fyi was used to help with the verification of this new browser! Also BiDi can't come fast enough.


As we explained in the blog, tests — and WPT tests specifically — were extremely important to get Kitesurf off the ground so fast, and still are as we move to higher API coverage.

We have been keeping a close eye on BiDi as well.


89% chance the thing you want is among the 10 items returned by the system


So over 10% of the time, it fails? That's not a great search engine


It's sort of the whole tension when you query vectorized data/embeddings -- you need to balance accuracy and recall against the performance you need.

It took me a while to wrap my head around the two terms since they seem similar -- but Accuracy is basically "did i get mostly good results" and Recall is "did I get most of the good results" and they're subtly different. :)

Those two terms, though, will unlock as deep a rabbit-hole as you'd like on the subject.


Sounds like you need to use IO optimized storage billing mode.


No test added/changed in this commit? Is there any test for this area?


Looks like tests were added: https://github.com/apache/httpd-tests/commit/48a85e34051959c...

Shame it wasn’t caught by any existing test though.


Did they check the tests passed on the old code and would that have caught the issue? That's an extra step I often do.


Well, that should be the first thing you? If it doesn't catch it, then what are you testing?


The tests include the comment:

  # Seems to have a side affect for any subsequent GET's
Well… does it? And if so, is that fixable?


Could you give some examples of "bad" pre-clearance offices?


The one that I frequently hear of as a layman (not a lawyer) is Toronto Pearson Airport's Preclearance, which has a very strict and narrow view of the TN Status eligible occupations.


Obviously anecdotal but I've had zero issues with preclearance at Pearson as a TN holder, both crossing and when applying for TN. They usually just ask me some combination of where I work for, what I do, and if I know when my TN expires, and that's it. My experience applying there has also been incredibly uneventful. I guess engineering is a bit easier to do though, especially if you have a strong application that's been done by lawyers.

I've been given hell by immigration about my TN when flying directly into US airports though.


Conversely, preclearance at Pearson royally screwed up my TN visa after claiming that the other ports of entries were doing it wrong. I have received multiple TN from HOU and, in general, I like the CBP at HOU. EWR was always friendly as well in my opinion. I personally prefer getting my TN at land border crossings because they tend to take the "gather all the facts" approach vs. airports have a procedural approach that feels highly dependent on the agent you get. My most recent TN was I-129 which I think is a waste of money but if someone else is paying for it and you have the luxury of waiting in the US, then it's a seamless experience.


Which part of this is a fact? That there are many Asians in California? And that they make up a disproportionately big part of Tesla drivers?


We use https://github.com/1debit/alternat in our EKS cluster to save the AWS tax (surcharge for NAT on top of the egress fee)


Discovered this when :PluginUpdate failed in my vim. It's still in Google index so must have happened very recently. Can't find any post discussing/announcing what happened.

P.S. Hopefully the username is not available for someone malicious to squat...


It was still there yesterday: https://web.archive.org/web/20230220153307/https://github.co...

You might want to ping the original author of the plugin: https://twitter.com/gmarik


You're over generalizing. Not all enthusiasts value the same point. For some including myself, it's not about acceleration or speed. Automatic (DCT and the new ZF8) has long surpassed manual transmission in almost all aspects of performance. It's similar to how some people like manual espresso machines, film cameras, etc. It's a hobby and you enjoy the process. See how the newest Koenigsegg goes great length to make this weird automatic manual.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: