lb4r

Because a lot of English is spoken by non-native speakers.

You aren't wrong, but my assumption was (perhaps incorrectly?) that they would add Lidar sensors in addition to their cameras, not replace them.

I'm not saying what the better option would be (because I don't know), but many people approach the problem from a very myopic point of view.

Adopting Lidar would of course provide Tesla with higher-quality input for their self-driving model. But the quality of the input isn't the whole equation; you need to process it as well. In other words, adopting Lidar would incur costs not only on the hardware side, but also on the software side, which of course would result in more expensive cars. More expensive cars means less cars sold, and less cars sold means less data, which in turns means less input.

Does this result in a worse model? Again, I don't know, but I do know that the issue is more complicated (and not only because of the reasons I mentioned here) than many people seem to think.

I think of it like this: Many people would "happily" spend 5 years in prison for a more than probable chance to get filthy rich. That's a superset with, I imagine, significantly greater cardinality than the set of those willing to spend 25.

Obviously there is a sweet spot. For example, if you're okay with 60 years, than you're probably okay with 80. I'd imagine 20, give or take 5 years or so, is near that sweet spot, but that's just my gut feeling. Obviously statistics is key here, if there is any.

See also: "Why not tell people to "simply" use pyenv, poetry or anaconda"

https://www.bitecode.dev/p/why-not-tell-people-to-simply-use

Just curious: what are the downsides of poetry installed with pipx? The article mentions having to install poetry in another venv, but that's hardly an issue with pipx (you just add an 'x' after 'pip'), and installing pipx is as straight-forward as it can be.

Just out of curiosity, why is the below example more git-friendly than the above? Smaller deltas? (and if so: why?)

Sure you can. Just don't shut down counter-arguments from, say, believers in certain form of animism, by implying that their side is invalid because there is no universally accepted definition of intelligence without giving your own assertion that same treatment.

Except that he doesn't acknowledge other points of view:

As much as people want to believe, LLMs are not intelligent by any stretch.

Acknowledging other points of view would have sounded more like "People are free to believe what they want, but LLMs don't strike me as intelligent."

I don't disagree for the items you listed, but for something that exhibits what in many aspects can AT LEAST be mistaken for 'intelligence' (among signs of the complete opposite, of course), I would just say that there is no way for anyone to know.

he's just pointing out there's no universal agreement on what intelligence even is.

Which is why I find it strange that he takes it upon himself to proclaim in a definitive manner that LLMs are not intelligent, and not "by any stretch."

But why are his toys floating around in the air and moving on their own? Why is there a guitar playing itself? Burden of proof just shifted.

LLMs are not intelligent by any stretch.

There is no concrete definition so there is no concrete way of deciding if something is intelligent.

You say there is no concrete way of deciding if something is intelligent, yet you yourself have decided that LLMs are not intelligent.

Wouldn't it be easier (since they probably have very skilled programmers working for them) and way, way more effective to just set up a team and create a quality open source project with one or two extremely stealthy backdoors?

Or just pay or threaten a struggling company or dev to insert them?

My workflow will probably seem a bit clunky since I try to make the notes as good as possible. That can mean spending anywhere from 10 seconds to 3 minutes writing a note if I intend to memorize it. And yes, this is all through the "Add" screen.

Obviously, if you're in class, you can't really spend that long on a single note. So in class I would usually just take notes (in Anki) without thinking about it too much, then rewrite them when I got home.

When browsing the web, I'll usually take screenshots (not bookmarks! though there are benefits to those too) when I see something interesting if I don't feel like putting it into Anki immediately. Then once or twice a week I'll go through my screenshots folder and either make proper flashcards or one-sided notes (which sometimes just means dragging the screenshot to the add window and pressing "add") of the stuff in there.

When watching a tutorial, I'll pause often and take my time writing notes if I feel like it's important. Otherwise I'll just take screenshots and decide later what I'll do with it, like above.

When it comes to tags, I'm hesitant to recommend the following because it's probably not "objectively" the best practice, but I organize my notes mainly by decks, and I don't bother at all with tagging them. I just found that it was too much of a hassle having to tag every card with multiple tags.

The key is really about finding joy in making the cards. I agree that the "Add" window isn't the most joyful experience, but I can live with it. Oh and I'll sometimes play around with stable diffusion to generate images to help me remember some of the stuff. Of course this will make creating cards take even longer, but it certainly adds to the fun!

The pitfall to watch out for with Anki is that it can become very time-consuming and feel like a burden when you have too many reviews. This might be something everyone has to figure out for themselves by trial and error, but if you're just using it to take notes without trying to memorize them, it's not something to worry about.

While I don't know of any Youtube videos dealing with this very specific thing, I think The Anking[1] has some of the higher-quality videos when it comes to Anki in general.

Either way, I'd love to hear about other people's workflows.

[1]https://www.youtube.com/@AnKingMed/videos

For note-taking specifically, I've tried everything from plain old pen and paper to more modern solutions like Evernote and emacs (if you can call that modern), but nothing I've come across really beats Anki.

Although its main selling point is as a program for flashcards with spaced repetition, it comes with pretty much all the features of a good note-taking app, like tags, easy to organize, synchronization across devices (you can set up your own server), good interface for searching through your notes (which are stored in an Sqlite db if that matters), and yes, LaTeX. Not only that, it's also highly extendable with third-party plugins, so if there are features that you miss chances are there's a plugin for it. In other words, you can use it perfectly fine just taking notes. However, where it really shines is in all of this in combination the spaced repetition algorithm, which is now on steroids with FSRS[1][2]. The downside is that for this to be effective for the things you want to memorize, you'll have to write your notes to be suitable for a flashcard, but if you do it consistently you'll soon notice that you can store most of your notes in your head. Needless to say, any student would benefit greatly from this. Now, if that's too much work, you can still just use the scheduling to have it remind you of your notes. Either way, even as someone who sometimes goes out of his way to shoehorn everything into Emacs, I can't see a reason not to use anki for note-taking.

Edit: I should probably mention that I primarily use Anki on my desktop. The smartphone app might not be as good, though I wouldn't know because I haven't given it a good try yet.

[1]https://github.com/open-spaced-repetition/fsrs4anki/blob/mai...

[2]https://www.youtube.com/watch?v=OqRLqVRyIzc

Speaking for myself, and except for just being curious, it's mostly for similar reasons as to why you'd want to read, for example, CLRS, even though you'll probably never implement an algorithm like that in a real production environment yourself. It's not so much about learning how, but rather why, because it'll help you answer your why's in the future (not that the how can't also be important, of course).

How easy will it be for any company like MS to correlate that data with all that other data they extracted without user consent?

Doable to some extent, but would they really learn much that we haven't already told them, given our propensity to Google for symptoms and diseases?

Personally, my worry here, if I had some embarrassing medical history that I wouldn't want people to know about, would be some malicious party gaining access to de-anonymized data and using it to blackmail, or just simply making it public.

Edit: Come to think of it, insurance companies could probably have a field day with a data set like this.

https://stackoverflow.com/a/66309040

This was the answer I originally linked to, because I found it to be an exceptional explanation of the inner workings of git. For some reason my submission now just links to the original question.

What I don't understand is this: Out of the thousands and thousands of participants in the study, some of them are sure to have unknown deficits. You'd think then that in studies like these, even assuming multivitamins are useless to the vast, vast majority of people, you'd see some slight reduced risk for things like heart disease and mental decline. Yet, the studies claim there is no benefit.

Unless people get sick from taking the vitamins and almost exactly offset the number of people that get better from them, I feel like something is missing. They clearly also state that multivitamins ARE in fact beneficial for some groups. If those groups were excluded from the studies (which I don't think should be that hard for them to do), shouldn't we then see an increased risk associated with taking multivitamins for people not in those groups?

I'm probably missing something?

I suggest people pick up the book "Immune: A Journey Into the Mysterious System That Keeps You Alive 978-0593241318" as it does a much better job at explaining these systems then my short paragraph.

I read that book just a few weeks ago and I don't recall reading anything about the stuff you mentioned.

The book is amazing though.

The idea is that there should be some amount of 'friction' involved. I used to have a script that made it easier, but realized after a while that it defeated much of the purpose of blocking the sites in the first place.

Unfortunately, nowadays I instead get addicted a lot to various websites, like YouTube and ... Hacker News.

I realized this about myself recently as well. I found a pretty good solution that works for me and may or may not work for you.

1. Disable all Youtube search history (no more recommendations!)

2. Use your hosts file to block sites you spend too much time on. To visit them you have to go through the effort of opening the hosts file and edit it (which for me requires typing in a long password because sudo). This was enough to break my habit of mindlessly visiting them. Here's an excerpt from mine:

127.0.0.1 youtube.com

127.0.0.1 www.youtube.com

#127.0.0.1 news.ycombinator.com

127.0.0.1 chess.com

127.0.0.1 www.chess.com

127.0.0.1 reddit.com

127.0.0.1 www.reddit.com

127.0.0.1 facebook.com

127.0.0.1 www.facebook.com

3. For your smartphone, uninstall all browsers and just use Firefox Focus. Turns out having to manually enter URLs and searches without auto-complete is pretty shitty user experience (which is good in this case!)

Technically, you just send a request to OpenAI and they are the ones who feed it into GPT4. Although I'd argue this is irrelevant to your question, the law works in mysterious way so perhaps it carries some importance.

Did you read what I wrote? When did I ever claim that Tesla or SpaceX were founded/not founded by Musk?

Around a decade ago it was considered pretty dumb to spend all your time and money building rockets and electric cars. I sometimes wonder how high the bar is set for 'intelligence.'

Interesting. The wording makes it sound like it would apply to controllers as well. So for a system like the Switch, they'd have to make sure that not only the battery of the console is replaceable, but also those of both Joy-Cons.

Any natural or legal person that places on the market products incorporating portable batteries shall ensure that those batteries are readily removable and replaceable by the end-user at any time during the lifetime of the product. That obligation shall only apply to entire batteries and not to individual cells or other parts included in such batteries.

So would this apply to something like the Nintendo Switch, and/or Steam Deck?

I'll look into it, thanks for the suggestion! The risk de-synchronization poses can probably be mitigated with automated local backups.

Thanks for the suggestion! By the looks of it, it seems to be a complete solution similar to Evernote, whereas I just need to figure out the 'syncing' part of things (I use Emacs for the rest).

I've contemplated this, but in general I'm not a big fan of hosting personal files on the cloud if it can be avoided. Maybe if I use rclone and encrypt it...?