An extraordinarily arrogant approach to presenting a project.
The TEXT datatype is deprecated in most RDBMSs. A minor and insignificant point but following the 'remove the brown M&Ms' rider principle, it makes me question the integrity of everything else, especially given the insults in the readme and lack of professionalism throughout.
TEXT is absolutely not deprecated in Postgres, which this project is using for its example. It being deprecated by other RDMS seems irrelevant to mention.
The number of times I mis-spelt Stonebraker in my Ph.D thesis... An absolute pioneer. I'm glad he's still around and active - sadly unlike many of his late contemporaries.
But, given how many last names (not only in English, but many other languages too) derive from trade or profession monikers, I can't help suspect that at some time earlier in his family's history it was spelled differently. (Except of course if it was first written down in this form. But then I'd guess that was because spelling wasn't all that stabilised a few centuries ago, so what was meant was still what we'd write as "StonebrEaker".)
"Anyway, it's just a thought." (i.e. idle etymological speculation.)
When I procrastinate, it stems from the thought, 'I don't really want to do this (right now|at all)'.
So one way to jump that hurdle is consider the consequences of not doing it, and how that makes me feel. For example, learning French. I would like to speak French. The consequence of not putting in the hours conjugating verbs means I will not be able to speak French. That makes me sad. I consider that sadness, and conclude I would prefer to spend the next hour reinforcing my knowledge of the passe compose of avoir. That is better than feeling sad.
Some consequences are not obvious, but cumulative. I don't really want to go to the stand-up meeting. What happens if I don't go today? Probably not much. But what happens if I don't go for the rest of the week, or my attendance is patchy? It'll be noticed, and I'll have to explain why I am not on the calls. The thought of the explanation makes me uncomfortable, more uncomfortable than going to the calls. Therefore I go to the calls.
Where this technique is powerful is that it enables me to filter out those activities where there is no obvious consequence of not doing the thing, which means the activities that remain on my daily list are generally pretty important.
> I would prefer to spend the next hour reinforcing my knowledge of the passe compose of avoir. That is better than feeling sad.
But it's not better than catching up on sleep, netflix and/or a great meal with a fantastic conversational partner.
> The thought of the explanation makes me uncomfortable, more uncomfortable than going to the calls. Therefore I go to the calls.
My mind would answer: "I'll take the 10min akward explanation for 5 missed meetings mr Gibbins. No problem."
Why not go to the calls because it is your duty? If nothing else, it makes you dependable and you can be proud of your virtuous follow-through.
Doing things only to prevent the penalty feels like a negative way of looking at things and, for my monkey mind, one that is ultimately doomed to fail. Instead of aiming for the stars, my mind just gets better at dealing with the penalties. It might just be me though.
You're probably high on conscientiousness if that's the case though. In which case procrastination is not something you are allowed to talk about because you literally don't know what it is.
Joking of course, but I have quite a few people who score high on conscientiousness around me. I myself am not such a person. I say they have absolutely no idea what procrastionation even means. Being a day late on a minor thing will be an existential threat to them. It is amazing and I watch them as I watch magnificent wildlife. Full of awe and wonder.
Then you have people like me who have to go throw the 5 stages of grief to overcome procrastination.
1. Denial of the importance or urgency of the task, and denial of my future self also lacking desire and willpower to complete said task in the future.
2. Anger that I cannot magically will myself into not procrastinating, or anger that I even have to do the task in the first place.
3. Bargaining how far back I can push a task back a.k.a. "I'll have plenty of time to do it tomorrow."
4. Depression because I always mismanage my time, overestimate my future abilities, and seem to never learn from the past -- "why do I always do this to myself?"
5. Acceptance that I am at the end of my rope, and I have to do the task now or I will face some kind of consequences far worse than actually doing the task itself.
I agree, doing things to prevent a penalty is probably worse than classic positive reinforcement/conditioning, doing things to get a reward. It's not healthy in the long-term.
However some things are generally pretty awful (such as standups) and don't really have a positive outcome that's easy to focus on and identify as a reward - not in my place of work anyway, YMMV!
So yes, in this case I could go because it's my duty (and try to feel proud of that!) but arguably forcing myself to turn up by focusing on what happens if I don't is also pretty effective, and is basically just like jump-starting a car - as other commenters have noted, merely beginning the undesirable thing is the biggest hurdle.
This breaks down when you stop caring about things. Say I want to learn French but ... not enough to put in the work. Give it up right? Say I want to look like a responsible and professional programmer but I still can't be bothered to show up to meetings. Why does it matter anyway? This yields a cycle of self-loathing and powerlessness.
So I prefer to focus on positive rewards for doing things instead of the consequences of avoiding them. And build from there.
For me when a task is overwhelming enough that I have been procrastinating on it for a while, or have attempted to start it a few times unsuccessfully, the thought of working on it for a whole hour is usually too much. But I can usually identify some trivial first step and say to myself “I’ll spend 15 minutes on that first step”. If I can get into the task, 15 minutes can easily become hours, at least until I find some decision overwhelming again.
But I get your point. I wanted to cite the original comment, though. It's true that the entry point has to be chosen according to the individual and the task at hand... as long as that's not something that has to be chosen so wisely that I should take time to reason about first :)
It’s very possible we mean different things by procrastination. I’m describing things I feel enormous pressure to complete, but also feel completely overwhelmed by. In this case I can spend time on the task, but won’t make any progress. Procrastinating is usually done out of desperation, seeking for some way to feel better to hopefully return to the task and make progress. Everything I do is pretty miserable while the task remains to be done. I’m trying to give myself a way to start. “I’ll just do it tonight” would be nearly unthinkable.
Nah, I meant the 15 mins would be even easier to be delayed again. Towards "tonight" right away, because it's just 15 minutes, just to re-schedule it for tomorrow when "tonight"
has arrived. It's just 15 mins after all, right. "Could do it anytime..."
I'd think the definitions seem similar. The avoidance strategies usually have slightly different hotspots, though.
J’aimerais bien trouver un cours en classe. J’en ai ras-le-bol de la formation sur internet. Malheureusement l’Alliance Française la plus proche est a 2h de chez moi. Je m’étais inscrit à un cours intermédiaire d’une autre organisation locale, mais il a été annulé car il n’y avait pas assez d’étudiants.
Please do not think that French is only in France. We use words here that french people won't understand and they do the same. And the aversion to anglicisms varies a lot depending on the person (ie "footing", a non-existing word in english in the sense used, is used in France to say "jogging").
Actually I can’t speak or write it fully, but I can understand you for the most part. I didn’t reach that B1-2 proficiency because I got hired around that time.
Thank you, this is news to me, I've been using the 'Learn French with Alexa' series on YT, plus a French dictionary and some magazines etc. I'll check it out!
I have looked online a little bit, but apart from some cultural centers around embassies not that I can see. They have an official cultural center, but only 80 worldwide vs 1000+ of Alliance Francaise.
Maybe just language centers or local people offering classes might be your best bet. Maybe an Italian aficionado can chip in on this one.
The weighting you describe is very interesting, I follow the same process semi consciously. I'm dealing with that right now on multiple front (job, family) and I'm curious to see what it yields to push things around as I see fit.
All this makes me wonder about the art of negotiating.. at the existential level. You only have to do things you didn't refuse in a way (figure of speech). Too often I said yes without asking more details, or said yes to things I didn't really like..
This doesn't work for me or probably some other ppl. My brain's attention would be dragged to the consequences and still won't do the work.
What I find useful is the vomit writing technique. I will just drag myself to get started no matter how horrible the work I am doing in the beginning. But once I started doing, I won't feel so bad. This basically solved my problem.
I just need to start doing and do a shit job, then improve it to be less shitty.
maybe another way around it is, 'do I really need it' -> 'and what that need is? how do I create a need for it?'. for language, that could be - applying that language. finding something that you do want to apply it to, that really gets you going and gives you joy. like, talking to someone, watching someone, listening to them talk, watching a stream, trying to talk to someone on that stream, reading, watching some piece of entertainment, in that language. and creating that 'need' for a 'want', and a 'want' itself - 'i want to watch those things and interact and chat with people - so i need language knowledge for that.' and maybe that could move it a bit closer. finding those bits that you might 'want' more readily - and then have those things move your goals into more of a 'need and want' zone, where you'd be both feeling a need of something more tangibly, and feel the joy in those things and possibilities of them more acutely and have that draw you in to do something.
and maybe it's also eh, it's counterproductive, but maybe it's fine to just have those curiosities - and have them just be that. even if it's kinda 'non-committal'
- maybe that's just kinda the dynamic for it, and that's alright. maybe some knowledge of language is just fine, and it's gonna be enough for watching or chatting, and it could be just those little bits of 'knowledge gain' where you look something up (like a word, definition, etc.) as you come across it, and amass some knowledge that way. this - does not help with 'how to do the thing', but hey - maybe there's also isn't really a need to beat yourself down over it either, over some 'thing about it you don't really want to do' when maybe it's just fine without it. (cause maybe that thing could be a buzzkiller that sucks the joy out of it, only further deterring you from it)
> Where this technique is powerful is that it enables me to filter out those activities where there is no obvious consequence of not doing the thing, which means the activities that remain on my daily list are generally pretty important.
This filtering is important. Doing X with a block of time means not doing Y, or Z, or any of A through W. I can do anything with this next minute, but that means that I am not doing all the other things with that minute.
So there are a pile of things that I "should" do that I am not doing. Which ones should I get to (even if that takes finding a way to defeat my avoidance), and which ones do I not need to get to, ever? In fact, which ones should I avoid, because they're going to take the time that should go to other things? I need some way to think about those questions.
I've always felt the downvote option was a colossal mistake and just allowed controversial or unpopular points of view to disappear, strengthening the filter effect for everything else.
I mean even 4chan doesn't have downvote functionality, if people disagree with you they'll just call you a f* or n* which, while nasty, doesn't actually change visibility of your post, plus only the admins can ban you or delete a post, not jumped-up mods, and frankly you'd have to be pretty extreme on there to earn that.
But 4chan is also a complete cesspit. Like any place that adopts "free speech absolutist" attitudes. Reddit's heyday is long passed, but at least it is still very useful for grassroots information amalgamation. If you've ever had to google some issue, typing reddit into the query will land you some posts with discussion in them. Good luck finding that on 4chan.
It is a cesspit but for broadening perspectives it is much much better than reddit, even with less structure of discussions and worse discoverability.
"free speech absolutism" wasn't ever a huge problem, on the contrary, the suggested alternatives always seem to end up worse.
You have to suffer through idiotic opinions if you expose yourself to new ideas. There is never any alternative to that and there is no magic moderator that can make this decision in your stead. It isn't any more complicated than that really.
> You have to suffer through idiotic opinions if you expose yourself to new ideas. There is never any alternative to that and there is no magic moderator that can make this decision in your stead. It isn't any more complicated than that really.
The main issue with that is that it assumes anyone on the internet is arguing anything in good faith. Boosting checkmarks has turned twitter to nearly unusable to actually unusable, because the signal to noise ratio on the replies from checkmarks tend to be somewhere on the factually provable incorrect to outright slurs spectrum. Expecting any better on 4chan of all places will be a a fruitless endeavor more akin to self-harm. If you don't want to wade through shit, maybe don't have your discussions in a sewer.
HN, while not perfect, will often be a place that cultivates some rare insight.
There is no trivial way to detect motivation, on the contrary leveraging the accusation of someone arguing in bad faith is usually seen as ending an intellectual exchange.
To be realistic, this can be that case, especially on the net, and there are ways you can make an educated guess. But it stays a guess and in the end I want to do the guessing myself. It cannot be outsourced aside from the most trivial cases like spam, advertising, etc.
That aside people with a foul mouth might use words differently while others people might receive it as a slur. We now have words that have much more might than they had before with the better approach to such issues.
This seems very naive to me. There are enough people yelling "OK Groomer" at me on certain platforms that I know the signal to noise ratio is bad enough that engaging is at best a waste of time and at worst getting more soundbites than I could possibly debunk, so I lose the optics part of the debate by default.
And as an aside - HN is not guilt free on this topic. There's a general tendency for comments to go right past the topic even outside this. For instance, every remote work thread turns into a generalized remote vs onsite debate, without the specifics of the article being weighed in. Even worse, HN also has a tendency to devolve into calling authors of posts deranged (or deranged, but with nicer choice of words) more for their identity and politics rather than the words they wrote, as was the case with Alyssa Rosenzweig's article being posted a few weeks ago.
It simply isn't viable to assume good faith for everyone. I often (unwiesely) get into fights with transphobes that I shouldn't, and I had maybe 2 insightful well-reflected discussions over the past 10 years. Sometimes they'll complement me on not fitting their expectation of an "SJW" (which is, presumably, a screaming child), but then go on to call me a slur anyway. And when I say slur, I am very positive that this is the right classification. Unambiguously.
I recommend the entire Alt-Right playbook series, in fact. A lot of these concepts are not limited to the alt-right and are often used incidentally by everyone, but it is important to know when you have no chance at rhetoric overcoming... anything. I guess sometimes the sword is mightier than the pen, especially in a post-fact world.
there's a few semi-shady things that 4chan does better than anywhere else though.
like finding pdfs for rpgs. not to say it's not a pit, cess or otherwise, but there's a good few nuggets swimming around in there.
I agree, downvoting is a bad idea across the board for silencing people that are not in the majority group-think. You don't even have to have a reason to down vote someone.
What I find interesting is that on HN, downvoting doesn't seem to have a negative impact, not as much as Reddit anyway. And I wonder if that's because folks who use HN really care about HN - I do - and make a good-faith, concerted effort to contribute in a positive way, with moderators who will enforce rules, but will not remove posts they disagree with (Reddit), or mandate arbitrary and inconsistent standards (Reddit).
Maybe it's the ol' HN hug-of-death but I'm getting a mostly blank screen on the #ALL page with a single post from 'Krazy', contents a single < symbol, and nothing else. Browsing tabs doesn't change the main screen, seems to be no content here.
If it helps, I'm on Edge v.114.0.1823.43, Windows 10.
Great concept though, I like the idea of subscription money going to popular content creators but worry this simply encourages lazy posting of popular meme content, basically a monetised karma farm. The capitalists that we are, many will find a way to optimise post engagement against the algorithm and many will be pursuing hard cash rather than social interaction.
As someone in this space I can attest that AI teaching in most (UK) universities is generally poor on detail, abstract and behind industry by at least 3-5 years.
Not to mention that there is zero appetite from undergrads or postgrads to get into the nitty-gritty of it. To learn CNNs at the deep-dive level you need calculus, at least differentiation and integration. Calculus or even pre-calculus doesn't form part of the degree programme for most compsci BScs any more, because it is 'too hard'.
The way most students 'learn' AI is to use a method out of a Python library with near-zero understanding of how it works, and regurgitate it for an assessment.
Professorial research staff in most UK universities are light-years from AI within industry, and there's no clear path to that gap tightening, especially while universities are being run like second-rate consulting houses (don't get me started on THAT).
lmao even in my university (the serbian uni), we have at least calculus+linear algebra before any nn course. also to "learn" what's a cnn you just need gradients not integrals (unless you use some kind of non-lipschitz function as activation?), plus the idea of what a convolution is... but even Mobius knew it back in the 800s.
anyway i think your statement that industry is light years away from unis is just misleading. i think the two are trying to answer different questions:
1. how can i achieve a "somewhat" decent chatbot that gets me rich albeit not even knowing what it does [industry in case you wondered]
2. try to understand, quantify and measure how well a model works, is it stable? does it converge if we have small datasets? and so on so forth.
just my two cents, to conclude i think a good analogy to the current climate is the 700-800s with electromagnetism: plenty of people discovered "empirical" laws but didn't understand really the phenomenon.
You might be able to understand what a convolutional network "is" without calculus, but you'll be woefully unequipped to ask even obvious questions like "what if we put Fourier transforms around the convolutional layers" (a cursory search suggests it provides the expected speedup but is for some reason not a standard thing to do?). As someone outside of the industry, I'd also imagine any effort to explain what NNs are actually "learning" (or I suppose dually, how to design network architectures) is going to have a lot of fruitful overlap with signal processing theory, which is heavy on calculus, linear algebra, probability, etc.
> a cursory search suggests it provides the expected speedup
What do you mean? Processing a CNN layer takes an amount of time that does not depend on the input data, only the input/output sizes. Fourier transform is just a change of basis. Why should anything speed up?
Because convolution is an O(N^2) operation, but a Fourier transform (and its inverse) can be done in O(NlogN) and turns convolution into O(N) multiplication. So if you do FFT, multiply, Inverse FFT, you get convolution in O(NlogN). I would guess that you don't even need to do the inverse FFT and can just learn in frequency space instead, but maybe there's some reason why that doesn't work out.
Computing convolutions using FFTs is efficient for large kernels (or filters). Most convolutions in popular ML models have small kernels, a regime where it is typically more efficient to reformulate the convolution as a matrix multiplication.
I think your complexity argument is correct for N=pixels=kernel size. But typically, pixels>>kernel size.
Disclosure: I work at Arm optimising open source ML frameworks. Opinions are my own.
I see, the wording confused me. You don't really "put Fourier transforms around the convolutional layers" because you have to completely replace the convolutions.
This seems to be done in some cases. I guess it isn't done more widely because the "standard" convolution kernels are very small and the performance would actually be worse?
>just my two cents, to conclude i think a good analogy to the current climate is the 700-800s with electromagnetism: plenty of people discovered "empirical" laws but didn't understand really the phenomenon.
Sounds dead on. Do these large """language""" models actually even implement any concepts from linguistics? Or is the entire "language" part of the model merely derived from the fact that it's inherently part of the training data?
I don't fault Chomsky at all for being fed up with the hype here.
The entire field is also glossing over the fact that other languages which aren't English exist.
>> anyway i think your statement that industry is light years away from unis is just misleading. i think the two are trying to answer different questions: 1. how can i achieve a "somewhat" decent chatbot that gets me rich albeit not even knowing what it does [industry in case you wondered] 2. try to understand, quantify and measure how well a model works, is it stable? does it converge if we have small datasets? and so on so forth.
GP here is, IMO, confusing what the corporations want (1), with what corporate R&D people want (2). As long as the corps see good ROI on throwing infinite money at their AI R&D departments, then those corporate researchers are better positioned and better equipped to do actual, solid science, than academia ever can be. This has happened many times before, including in this industry. Research is best done by well-funded teams of smart people left to do whatever they fancy. When those conditions arise, progress happens, and it doesn't matter whether it's the government or industry that creates them.
(Conversely, the best hope for academia to become relevant again is that corporations lose interest in this research, and defund their departments. This could happen if e.g. transformers end up being a dead end, or compute suddenly becomes very expensive.)
> Do these large """language""" models actually even implement any concepts from linguistics? Or is the entire "language" part of the model merely derived from the fact that it's inherently part of the training data?
The latter. And guess what, they're not trying to solve the issue of linguistics. They started as tools to generate human-sounding text, but in the process of just throwing more data and compute at them, they not only got better, but started to acquire something resembling concept-level understanding.
It turns out that surprisingly many aspects of thinking seem to reduce well to proximity search in a vector space, if that space is high-dimensional enough. This result is both surprising and impactful well beyond the field of AI. It's arguably the first potential path we identified that the evolution could take to gradually random-walk itself from amoeabas to human brains.
i'm not saying LLM are modeling linguistics in any way lol. i only meant that there's some kind of phenomena related to scaling+attention that produces good enough result for most "human language stuff", which is kind of unexpected (i mean everyine knows that if you build a large enough model you can teach it any function, but cmon it is architecture+scaling that made it possibile not scaling alone). Moreover, the architectures used, from attention layer, even LSTM for that matter are not completely understood, are being used because "they works" just as in the old days of electromagnetism the empiral laws "just worked" for their usage.
btw, in other languages i guess it is decent although it depends on which language, at least gpt4.
> Calculus or even pre-calculus doesn't form part of the degree programme for most compsci BScs any more, because it is 'too hard'.
In the US I’ve never seen a BS in computer science that didn’t require calculus. I can’t speak for the UK, but it would surprise me that what you say is true.
