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.
You have it wrong. Processing LIDAR is way, way more computationally efficient than processing camera footage into a 3D model. LIDAR feeds you direct distance and speed data. Cameras of course do not, meaning you have to try to compute it, something which is very hard and error prone for even the most powerful computers that have ever been created (human brains).
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.
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.
I eventually landed on pipx as fighting with Pyenv and Anaconda - via Miniconda - was an exercise in frustration. There's some mucking about in `$HOME/.local`, but this is mostly self-contained and not a huge chore to keep running.
Coming from the Homebrew/Ruby ecosystem - Hey @mikemcquaid - installing a entirely separate package manager just to deal with a few projects felt like the wrong thing to do.
Occasionally, I have still needed to compile Python myself in order to get things to work, which isn't guaranteed not to blow up w/ `brew`, but this has become far less common of late.
Agreed pipx solves a lot of packaging issues with no downside to speak of. Not just with poetry but also for tools like virtualenv, ruff and black and non-dev command line tools.
We may be unable to provide a concrete definition of what intelligence is, but we can certainly provide definitions for what it isn't. E.g. we don't need a concrete definition of intelligence to say that a rock isn't intelligent. A pencil isn't intelligent. A calculator isn't intelligence
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.
You seem to have a misunderstanding of what the condition "concrete" means.
He's not saying there's no way of judging intelligence, he's just pointing out there's no universal agreement on what intelligence even is.
Edit: To add, this discussion becomes pure semantics. On one side is a strict definition of AGI, on the other side are the most generalized definitions of artificial intelligence. It gets kind of silly because technically, every "if" statement is a type of "AI" by the loosest definitions.
>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."
Why would you find that strange? He's using his own definition of intelligence while acknowledging there's no universal agreement.
Expressing your own point of view while acknowledging that other points of view exist shouldn't be strange. Strange is not being able to see things from different perspectives, those people are abnormal even when they happen to be in the majority.
The only valid criticism I see is that "by any stretch" is hyperbolic, but that's easily forgivable.
> As much as people want to believe, LLMs are not intelligent by any stretch
They said "As much as people want to believe" which means that it shouldn't be counted as intelligent by other people's definition. Even by most liberal interpretation, the comment(which is top rated) doesn't say what you are trying to imply
>Except that he doesn't acknowledge other points of view
False, he explicitly acknowledges other points of view on what defines intelligence:
"There is no concrete definition so there is no concrete way of deciding if something is intelligent."
Acknowledging that your way of thinking isn't the only way of thinking doesn't make someone a hypocrite, it's actually a sign of intelligence (in my opinion).
There's no concrete way to prove my kid's imaginary friend isn't real, but that doesn't mean both sides have an equivalent likelihood and burden of proof.
If that were the only known fact about said imaginary friend on either side, then it would mean exactly that. The reason that there are different expectations of veracity are exactly because there are a bunch of priors being held about imaginary friends, by definition, not being real.
The output of the LLMs would suggest a different metaphor. The supposed divine inspiration of the Bible, perhaps?
Compare "Intelligent Design" vs. the use of genetic algorithms in AI. Simple forms of intelligence can get you a long way and can seem very impressive, especially if they have a lot of subjective experiences, which DNA gets from deep time and which AI gets from transistors outpacing synapses by the ratio to which a pack of wolves outpace continental drift.
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.
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?
How would you secretly hide something like that in FOSS? And why would that be easier? It's seems to me that it's easier to inject into an existing company than to do all the work yourself. This is what they do with most things as I understand.
Yes, but that was a memory leak, giving access to unauthorized random memory. That is not an intentionally created exploit / backdoor which gives the owner easy access to the victim's system.
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.
I use Anki the exact same way. After a lifetime of learning I have accepted that I will never read over anything I write for myself voluntarily - so my two options are:
1. Write an article so good I can publish it and look it over myself later on online. I did this last year with https://andrew-quinn.me/fzf/, for example.
2. Create Anki cards out of the material. Use the builtin Card Browser or even https://datasette.io/ on the underlying SQLite database in a pinch to search for my notes any time I have to.
That looks really nice!
However, for me a flashcard app is one of the apps where I don't want to use commercial offerings if the source code isn't available as well. It involves just too much set up time to have it vanish one day.
I totally agree with the idea of having it all vanish. I’m still building it out, but I’m working towards an ‘eject button’ where you can download all of your notes into a CSV and get out of dodge if you choose!
This won't really help. The point of an SRS is the scheduling information. Even then, you may not have a compatible system to make use of the scheduling info. The switching cost is immense, and I would personally need an ironclad guarantee that the tool cannot be taken from me, such as by having the source code. Never going to be using a non-self hosted SaaS for this, sorry. There is a lot to dislike about Anki but the good thing is it will be only taken from my cold dead hands.
I'll give it a test run :) I probably will not be a convert, but I love seeing people spin up new SRSes. It's a surprisingly vast design space, and Anki is pretty maximalist within it.
Pohtia also means to ponder in Finnish, not just Norwegian, fun fact!
Very surprising… never thought this this way. How can you guy even thought of this. Using anki for language learning and latex, vim … how can this even occur to your mind … really wonder.
I actually did use Anki for Vim, lol. I started by putting the tips in Practical Vim in, then a few easy Vim Golf questions to remember how to do some advanced maneuvers.
I use this approach also and it has served me well now for many years. What I really like is that you can break out of the classic 'flashcards' mindset and don't even necessarily need to have it 'test' you for anything all that difficult. Often times, I want the note to mostly capture my thinking and be available for later reference, so I cloze out a simple word that I'll always recall. Then it's not hard to go through during my reviews, and even if I don't remember exactly what it said without prompting, I'll know I can find the note with a quick search.
I have never heard of Anki used for live notes in class. I find that surprising because Anki cards are said to be good if you use them to capture quanta of understanding. While you can make cards on whatever facts or understanding statements made explicitly in class, I find it hard to believe that you would be able to create those novel long range understanding cards during class.
And you can't create those cards for existing cards. You need more comprehensive notes (like your class notes) to create those cards.
I use both Vim (with wiki.vim, a plugin I've made based on Vimwiki) and Anki. I find the combination is very powerful!
I keep a journal for work related stuff and a wiki/personal knowledge base for everything (both work and personal). I use Anki to memorize things I believe may be useful to have readily available and I use my wiki as an extended "long-term memory", sort of.
I'm interested in learning this workflow... are there any YT examples that you know of? I just can't see how it would work in my head. Are you using the "Add" screen to constantly add cards using hotkeys?
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.
I’m speculating there could be a more ergonomic workflow combining the vim-latex flow from op and this comment(and the reply). I know you can import anki decks like a .txt(?)(or similar), so I’m imaginig a ‘flow where you live write notes in the ~.txt file and later just import it to anki. It might be a hassle to set it up all but seems way better than using the Add interface, and also I prefer latex over screenshots (but I get that ss is just more convenient for some situations)
I love Anki, but my main gripe is the UI is just so bad. That said I truly believe I would not have graduated from college without it, and I'm forever grateful.
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.
The civilized world has universal health coverage, so the insurance thing isn’t really a problem except in one particular country. But the malicious release is definitely a problem.
It never is a problem until it is. Insurance companies in Germany are also not benevolent. If there is a case where they do not have to pay they won't. If they can use data to save money they will do it, even if it worsens your conditions.
Not all public insurances perform these elections but some, e.g. Barmer GEK, Techniker Krankenkasse, DAK-Gesundheit, Kaufmännische Krankenkasse – KKH, Handelskrankenkasse und BKK RWE and I think also AOK.
It doesn't feel adversarial as a customer. Never did for me, never heard it from others. They aren't even regular companies - they have a special legal status. Private insurance, which is also available, is different.
The "civilized world" does not, however, have universal life insurance nor disability insurance that would cover the full standard of living. Those are, at least in Germany, privatized, and companies could benefit a lot from having access to some extra data.
I'm not sure that getting caught using such data would be sufficiently bad news to be not worth it.
No, because medial data can also be used for things like life insurance, insurance against loss of earnings caused by sickness, and probably a lot of other things.
I've been in Europe two decades and aside from health insurance being key to more than a few countries, private health insurance is relatively common in tandem with public health services.
Quite typically there is basic service via a welfare system and additional or premium service with insurances. I pay into the public health system at a rate proportional to my income (which is more than what I would pay if I had a worse-paying job for the same service), and I have an additional dental insurance which insures my teeth.
No, I mean full-blown private health insurance. In France, Belgium and Netherlands it's very common (and I think mandatory in NL).
Outside of those countries, it's still somewhat common to have private health insurance (not just dental), such as Mapfre or Regina Maria. Heck, even in Germany something like 1 out of 10 people have private health insurance (again, not just dental).
