Folks, just live your life how it fits you. We're all born into a different set of circumstances that dictate many of the choices we make. We all have different ideas about what life means to us and what levels of risk we're comfortable taking. Listen to your gut and focus on what you'll be smiling about as you lie on your deathbed.
Korea and Japan also have these. The Japanese ones are covered extensively on youtube, however, my favs are by Tokyo Lens, I like his style and his voice is very calming. Here are some of my faves:
I was one of the first backers of the Oculus Rift Kickstarter. When I got it I decided eye tracking was going to be huge for VR, so as a side project I cut a hole in my Rift and built my own eye tracker. I posted it on Hacker News: https://news.ycombinator.com/item?id=7876471
A few days later the CTO of a small eye tracking startup gave me a call. I quit Google and joined them. I built a (novel at the time) deep neural net based VR eye tracking system for them, and less than two years later Google acquired us.
I think this is "how to think about coding assistants and your task" but none of this is "tackling" their unreliability.
While coding assistants seem to do well in a range of situations, I continue to believe that for coding specifically, merely training on next-token-prediction is leaving too much on the table. Yes, source code is represented as text, but computer programs are an area where there's available information which is _so much richer_. We can know not only the text of the program but the type of every expression, which variables are in scope at any point, what is the signature of a method we're trying to call, etc. These assistants should be able to make predictions about program _traces_, not just program source text. A step further would be to guess potential loop invariants, pre/post conditions, etc, confirm which are upheld by existing tests, and down-weight recommending changes which introduce violations to those inferred conditions.
ChatGPT and tab-completion assistants have both given me things that are not even valid programs (e.g. will not compile, use a variable that isn't actually in scope, etc). ChatGPT even told me that an example it generated wasn't compiling for me b/c I wasn't using a new enough version of the language, and then referenced a language version which does not yet exist. All of this is possible in part b/c these tools are engaging only at the level of text, and are structurally isolated from the rich information available inside an interpreter or debugger. "Tackling" unreliability should start with reframing tasks in a way which lets tools better see the causes of their failures.
Does the effect work with an animated depth image?
I always have a hard time with these. I get the 3d effect but can never figure out the image. I suspect based on reading some theory that I am viewing them cross eyed instead of slack eyed. But have no idea how to do it the other way around.
That's an amazing piece of work. Assuming accuracy: what strikes me is that certain features of the ancient city apparently carry over into the modern day and how ecologically balanced the older city looks compared to the new one. Another interesting feature seems to be that 'places of power' back then are still places of power today but different kinds of power.
The article says they’re doing “showings” of the film, which makes me think that they don’t want it digitized, probably because the copyright is expired.
If they control all the copies, then they can charge people to watch it, even if the copyright is expired. If a copy leaks, though, then that revenue stream will dry up overnight.
Just for funsies, I'll throw in the latest remote desktop that I have tried for Linux, the Sunshine server. It uses the Moonlight protocol, so you can use any NVIDIA GameStream client, like I have used Moonlight from f-droid to stream my PC to my Android phone. Clients exist for Windows and Linux too, and a lot of other platforms as well. Sharing the desktop works out of the box, just download the server, run it, very minimal config to set a password, pair it with a client, and then next time, you just need to run the server and connect with the pre-configured client.
It works by converting screen or windows's content to video. It's low latency by design, bandwidth is up to you, and sound also works by default.
I'm clueless -- how do these things work? Are they standalone, or do you also need a gaming PC? Do you need internet connection, or is there stuff that works offline? Do you need Facebook account? Are they hackable, e.g., are there custom firmware options?
This is what I wish Dropbox was, a simple layer that make interacting with your FTP server easy so nobody has to own your data. The end game is both to be feature complete with Dropbox and be able to change every aspect of the application through plugin so everyone can get out what they want from it. To do that, I’m making the core a sort of framework to build file manager like application.
I have been tinkering with this word game for something like 6 years. It is intended for mobile, though it works adequately on desktop. It is a virtually endless word-finding game that can be played solo or taking turns with a friend. I still need to add some game modes and more interaction, but the bones are there:
Hey I built getdispute.com to do literally exactly this: sue big companies in small claims court easily.
First, I strongly recommend sending them a demand letter. You can generate one for free on getdispute.com/products/demand-letter and/or pay us a few bucks to mail it.
Second, physically mail the letter to their legal address. The mail goes to someone in legal tasked with reading the mail.
90% of our cases end here. A correctly written demand letter, sent by physical mail usually gets noticed and resolved very quickly.
If nothing happens after a few weeks (give it 3-4 wks) then you could file a case in court. This will require you to file the case, pay the filing fee, and then serve the defendant via a 3rd party process server. Could all be done in 1 day if you know what you're doing. If you serve the defendant (FB) correctly, they are now obligated to show up to court or otherwise risk a default judgement against them that cannot be reversed later.
Another 50-75% of cases end here as most defendants prefer to settle out of court.
If they don't respond, you will need to go to your court date and explain the case to the judge. Usually, without a defendant appearing, the court will grant a default judgement in your favor, assuming the judge decides they do in fact have jurisdiction. To be safe on jurisdiction, file in the county FB does business, or otherwise has a legal entity registered.
All of this said... it's probably not worth the time/money to do more than send a demand letter. It is probably worth your time to send a physical demand letter.
Good luck. 2.7m small claims court cases are filed each year in the US, approx 50% by individuals. So you're in good company.
I've found elinks to be really good. It preserves colors and has really nice navigation controls for interactive use (you can press dot to toggle numbered links). Here is a screenshot of this page: https://i.imgur.com/mP1T0H3.png
If you're just looking to format a page, it has a -dump option.
If you want to preserve colors as well, use -dump-color-mode 4
You probably won’t find exactly the video you’re looking for, but if you’re interested in creative coding and shader based animation, IQ (Shadertoy author) has some amazing videos on his YouTube channel. Shaders are tricky!
It’s a deep and fascinating development world, one much less thoroughly documented than the world of web apps. Good luck!
HIGHLY highly recommend checking out Stephen Marz's OS blog, where he walks through building a basic RISC-V OS in Rust step by step: https://osblog.stephenmarz.com/
I seem to recall IBM launched z/os in the cloud as partly a production service, and partly an education/hobby resource a while back - but I'm not sure if that's what's now:
DonHopkins on Nov 14, 2016 | parent | context | favorite | on: The NHS's 1.2M employees are trapped in a 'reply-a...
Back in the days of ARPANET mailing lists, there used to be an "educational" mailing list called "please-remove-me", that was for people who asked an entire mailing list to remove them, instead of removing themselves, or sending email to the administrative "-request" address.
So when somebody asked an entire mailing list to remove them, somebody else would add them to the "please-remove-me" mailing list, and they would start getting hundreds of "please remove me" requests from other people, so they could discuss the topic of being removed from mailing lists with people with similar interests, without bothering people on mailing lists whose topics weren't about being removed from mailing lists.
It worked so well that it was a victim of its own success: Eventually the "please-remove-me" mailing list was so popular that it got too big and had to be shut down...
...Then there was Jordan Hubbard's infamous "rwall incident" in 1987:
I think I can count the number of kernel changes I've submitted on one hand, but I work on core virtualization that involves a lot of pretending to be hardware and (these days) a lot of poking directly at hardware registers.
I would say James Mickens sums things up nicely in "The Night Watch[0]." For example, you mention debugging with logs and metrics -- this snippet came to mind:
“Yeah, that sounds bad. Have you checked the log files for
errors?” I said, “Indeed, I would do that if I hadn’t broken every
component that a logging system needs to log data. I have a
network file system, and I have broken the network, and I have
broken the file system, and my machines crash when I make
eye contact with them. I HAVE NO TOOLS BECAUSE I’VE
DESTROYED MY TOOLS WITH MY TOOLS. My only logging
option is to hire monks to transcribe the subjective experience
of watching my machines die as I weep tears of blood.”
Mind you, I absolutely _love_ working on low-level stuff, and I wouldn't trade the time I get to spend actually doing that for anything. That said, the complexity of modern operating systems, CPU architectures, interconnects, and peripherals creates opportunities for frustration and confusion that honor no bounds of reasonability or decency.
Building your own custom launcher utilities can be fun (and fully tailored to your needs). I’ve had good results using completion frameworks like ivy, accessible from any app (not just Emacs).
I disagree. I came to Emacs from a combination of VSCode/Pycharms and Jupyter notebook. I think the latter are still held in highest regard in the IDE space. To this day there is nothing in another editor or IDE that I miss. I find that Emacs allows me to invest as much as I want into my tools: editor, notebook, document generator, and shell. If you are looking for a maximal out-of-the-box product, then sure Emacs will come up short. However if you are looking for a product that you can develop and use for a lifetime then personally I don't think any one piece of software comes close
NOTE: Emacs ecosystem has had support for multiple cursors for over a decade. Emacs (like Python) should not be valued in terms of built-in features, but in terms of its community and ecosystem
I sincerely feel that the discipline of building user interfaces is long lost and perhaps never to come back. From car dashboards to loud typography, the whole field is regressing. Minimalism, too much white space, infantilization of users, design trends, loss of predictability, etc. are all symptoms of this.
