I mean, you label ADL as a Nazi party. Then you label Musk as Nazi by association. But if you actually look at ADL they are not Nazis. Here's an interesting things I saw just today: https://x.com/derJamesJackson/status/1887425955494268997
I've done the whole spaced repetition and Anki thing, realized that if it's boring, you won't learn as well, and you won't stick with it in the long term, once you miss a week and suddenly you have 863 reps to get through to catch up. Instead, read stuff you're interested, apply it in your work, learn mostly Just In Time for when you need it. Learn by doing.
Strongly disagree. The worst thing that’s happened to educational theory is the idea that everything needs to be fun and exciting to be worth doing. Some stuff is worth doing because it works, even if it’s a little boring. SRS is one of those things.
Because realistically doing your Anki cards should take 10-30 minutes a day, maximum. If you can’t slog through 30 minutes of learning vocabulary/etc. then the thing you’re learning isn’t very important to you. SRS isn’t about doing flash cards for hours on end, it’s about determining the maximum retention rate while doing the minimum amount of work.
Even then, personally I don’t find Anki that boring at all. But I make custom cards that include audio, images, and other items.
> If you can’t slog through 30 minutes of learning vocabulary/etc. then the thing you’re learning isn’t very important to you
This obviously isn't the case if there's an alternative approach that does work for you.
> Even then, personally I don’t find Anki that boring at all.
Right, and you're having trouble picturing how something you get for free (or take for granted) has a cost for others. Isn't it obvious?
Running 10km a day is a great way to stay in cardiovascular shape. If you like running, it might be a great daily habit for you. But would you say other people must not care about their body just because they're not willing to run 10km, but they are willing to bicycle, swim, or kickbox daily?
I think you are dramatically undervaluing how effective SRS is. 30 minutes of doing flashcards with Anki will accomplish about 500% as many other learning activities in the same time, in my experience.
So the point is that while SRS might be boring at times compared to other types of learning, the benefits are worth it. Effectiveness of learning does not correlate directly with enjoyment.
To phrase this another way: yes, in general you should find learning methods that appeal to you. But spaced repetition is so effective that it’s worth implementing, even if you don’t enjoy it.
Edit: using the metaphor that you added: if say, doing jumping jacks for 10 minutes burned 10x calories than every other exercise does in 60 minutes, and your goal is to lose weight, then yeah, it would seem a little silly to me to avoid doing jumping jacks because you don’t enjoy them. The benefits are worth the hassle.
The science behind spaced repetition isn’t really up for debate, though. I guess if you really hate SRS, don’t do it, but all signs point to it being worth the time.
You know what is more effective than spaced repitition? Reading, listening, and doing to the thing you are interested in. Anki gives you fact that are disjointed from what you need. using the skill forces a study of what is important not facts that are of no use.
i'm not saying spaced repititon is useless - it is a great beginners tool. However you need to move on as soon as possible. once you can use the skill that is better in gereral. Doctors use spaced repition more than most because there are a lot of rare things they will never encounter but mixed in is one rare thing they need to know instantly and they cannot know in advance what that thing is. Almost nothing else is both like that and doesn't allow time to look it up.
note that reading/doing something is itself spaced repititon. It isn't algorithemic but it is still randon spacing of things you need to know and the more you need it the more often it repeats
I built and ran the academic side of a language school for four years, worked on two language learning startups and also learned two languages in different language families from my native language (and from each other) to a pretty fluent level.
In all those years I’ve never met a single really successful learner who made SRS a major part of their studies, though some used it as a small supplementary practice.
The thing is that words aren’t usually discrete pieces of information in the way that names of capitals or things a med student has to memorize often are (unless your goal is just to play scrabble with them, in which case SRS is great). Meanings don’t map one to one across languages, collocations are important , etc, etc. Putting sentence cards into Anki is better than isolation words with translations but even then, you won’t get as much cultural information or even raw quantity of input as you would from extensive reading and this is a topic L2 Acquisition researchers have covered in depth.
I think part of the problem is that SRS sounds really compelling to engineers and it’s generally easy to build into an app, so that’s been the focus of most language learning apps for the past 20 years.
There are some better ones that exist mostly to help learners handle native text and audio, though. LingQ, Language Reactor and language-specific apps that do similar things are great.
> In all those years I’ve never met a single really successful learner who made SRS a major part of their studies
I did, but very differently from how most people do it.
Premade decks with single words sucks, yes. But creating your own flashcards with lots of audio and mined sentences from your learning materials accelerates learning better than anything I tried before.
And yes, it requires a lot of immersion / exposure and other things. But when done properly it works really really well.
Have you ever gotten to a C1 level in a language you learned as an adult that isn't close to your native language? If so, what kind of effort did it take you?
Or if you only learned closely related languages, what aspect of it was "really successful"? Accent? Writing ability? Something else?
It is a lot effort, I find Spaced Repetition actually very demanding – but effective, that's the point.
If you take Paul Nation's math (the famous https://files.eric.ed.gov/fulltext/EJ1044345.pdf) and instead of doing that you mine the hardest sentences you find to an SRS (basically, separating grain from chaff) you save yourself a lot of time, because you will never find a piece of text long enough where you know exactly 98% of the words.
In my case, I use flashcards primarily for reading and listening comprehension, but lately I am using it a bit more for pronunciation, and it can also work well for spelling.
Some repetition is good and children in fact often repeatedly read their favorite book or watch their favorite show. However, a key thing to keep in mind with Paul Nation's study is that "repetitions" of a word represented encountering it in different contexts.
This is important for three reasons. First is polysemy, which he mentioned. This is words with multiple meanings like fair (reasonable/just / light or unblemished / a type of public event with entertainment and vendors).
The second and in my opinion even larger issue is gaining and understanding of how the words are used and their scope. E.g., there's no word in Chinese that quite matches "nose" in English. The closest is 鼻子, but that can refer to an elephant trunk or a pig's snout, which "nose" can't. In some languages, there's a word that can be used for humans and pigs, but not elephants. In others, the same word also encompasses bird's beaks. The only way you'll learn this is from encountering the word in a lot of different contexts, not from drilling it repeatedly in the same context. Furthermore, there are a lot of words that tend to be used with or near each other but near others (collocations). In English, it's normal to say you're doing "pretty good" or "absolutely fantastic", but saying you're doing "absolutely good" would be very strange.
Finally, there are a lot of shared cultural stories each language community has. In English, these would be from Christianity, from classical Greco-Roman figures like Aesop, from German storytellers like the Brothers Grimm, other European storytellers like Hans Christian Anderson, etc. In Chinese (or Korean, Japanese, Vietnamese or other nearby countries) people are going to have a lot of shared stories from ancient China—philosophers from the Spring and Autumn era, historical stories and dramatizations of the interesting times in history such as the Three Kingdoms era, and many, many folk tales. If you don't understand at least the core of the cultural cannon, you'll regularly be confused by things in TV shows like soap operas or variety shows, even if you understand every word in transcript. Reading fiction will be even harder.
The more distant the language is from the one you speak natively, the more the second and especially third points will impact you unless you take in an enormous amount of input.
> You know what is more effective than spaced repitition? Reading, listening, and doing to the thing you are interested in.
No, they are not more effective, they are less effective. The fact that they are less effective is the only reason that people do spaced repetition. It takes less time to learn a fact if you optimize for the way memory works. Reading, listening, and doing the thing you are interested in builds memory just like spaced repetition does, just in a haphazard, random way.
> note that reading/doing something is itself spaced repititon.
But worse. There are plenty of things that simply can't be learned by spaced repetition because nobody has come up with a good way to do them, and/or they are physical. But some facts simply have to be memorized, and when doing the thing you will have to recall them. You can either look them up every time you have to use them, and a) forget completely if there's too large a gap between incidents of having to look a thing up, or b) spend all day studying something that you'll completely forget in a week. Spaced repetition is about using a timing trick to make those facts reflexive and long-lasting.
When reading/listening and doing you encounter thousands of facts/words/whatever in the same amount of time spaced repetition will give you tens. If you are a complete beginner spaced repetition does well, but that is misleading - it does well in the easy to study situation of someone who knows nothing and doesn't examine the much harder to study place where someone knows a little something where you spend most of your time.
If you know nothing, spaced repetition can get you over that initial hump well. However I maintain that you quickly reach a point where you are beyond that and then space repetition is less useful than other courses of study in general. Use well it can still be a supplement to other study methods, but those others should be the priority.
Anki is just a software tool for presenting pieces of data at an optimal rate for memory. It’s perfect possible to use it for reading, listening, or even prompting activities.
> If an approach is so boring that you don't do it, then what does it matter how effective it might be?
Yeah true, but an obvious argument is that this is where discipline comes in. If you are one of the people Anki works for, then you have to find the level of discipline required to stick with it.
> once you miss a week and suddenly you have 863 reps to get through to catch up.
You only have to "catch up" ASAP if you want to minimize the risk of forgetting any card at all, and then having to memorize it again. Otherwise, just working through the backlog at your preferred pace is the best strategy; you'll still be recalling most of the cards and the system will push repetitions for those that you do recall further out in time.
Perhaps the Anki devs should add some kind of special workflow in the UX that shows up when the user has a huge backlog to go through due to missed reviews, and tells them this. Because you're right that just seeing a bare number (863 cards to go through!) can be quite scary.
Spaced repetition is the cherry on the cake: learn by practicing, but create your own cards along the way, and when reviewing just focus on making a recall effort, not nailing everything.
For language, 10-20 mins / day is great, and if you are putting enough effort, 6-8 reps is enough to learn a card with 2 new concepts at ~15secs / rep, which really makes a difference long term.
I made a tool[1] with that approach in mind and tailor-made for language acquisition patterns. The key part is that creating cards (audio included) must be super-easy and fast, but they need to come from you nevertheless.
This looks very interesting. The language dialect that I am learning now has no online sources, but many words are similar to other dialects that are online. How does the app find the native pronunciations of words? What can we do with the voice notes, other than just listening to them? Transcription?
I think the mistake here is simply to many new cards a day. If you miss a week and have 863 reps, you definitely are going way too fast. You shouldn’t have to spend more than 5-20 minutes daily going over your review, and if you miss a week you shouldn’t need more than two or three half hour sessions to catch up.
I think there are diminishing returns as well by having to many daily new cards, as there is only so much you can commit into memory every day. So you will probably end up having terrible recall as well (which further adds to your reps).
I’m using Anki to learn Japanese. I actually go a step further in taking it slow as I’m still on my N5 vocab (first 1000 words) after 9 months of learning. I think it will be another 9 months before I can read stuff I‘m interested in in Japanese (which will probably be Go books).
I used it for language learning and CS among other things. The thing is, life gets busy. Suddenly you have a kid or two. Spaced Repetition falls into the same category of things that people in their 20s do because they have too much time, like using intricate note taking systems, journaling, training for triathlons and being really into artisanal coffee.
I haven't found that at all. I'm well past my twenties but find Anki is one of the things I can fit in, mainly because even with kids and responsibilities you can often find small periods (say 15 mins) of time through the day. It's not enough time to sit down and start into something really complex, especially as the time is sometimes interrupted, but it is enough time to try a few questions.
I guess it just boils down to what you want to do with your spare time. You find time for HN. Me, I can find time for 20-60 min a day studying Japanese (of which 5-20 is in Anki). I also find time for 20-60 min a day playing go so ¯\_(ツ)_/¯
Fine, time AND energy. HN is entertainment. SR feels like work. When I have an hour at night on average to yourself, I don't want to spend it studying. I've tried taking breaks at work to do it, but you need those mini breaks at work to be a bit unproductive as well. I can't be "ON" 100% of the time, I need rest.
An interesting thought experiment what would happen if we treated people as corporations. I guess we'd have to raise income taxes quite a bit to make ends meet, as most people spend everything they make.
> what would happen if we treated people as corporations
You'd rename line items and move on. Companies can't deduct any expense. The chief ones are things they've already paid tax on, e.g. wages subject to payroll and purchases subject to sales taxes.
Yes, of course. But it's also inevitable and overall good. If we stagnate then we are headed towards certain doom via climate change, nuclear war, an asteroid, a vulcano eruption leading to cooling and crop failure or a pandemic. If we are to survive and thrive long term we need to become true masters of our environment, and that means we need to be smarter, stronger and more productive.
I think this is my main disconnect with the pessimists, I don't see "stop AI progress" as a valid option anyway.
Yes, collective suicide means we no longer have any Climate Change issue. Calling this "a solution" stretches the definition of the word tho.
> If we are to survive and thrive long term we need to become true masters of our environment, and that means we need to be smarter, stronger and more productive.
Yes. I fundamentally agree with that vision. We want this for us. We/Us humans.
If we build AGI we won’t thrive. We won’t be smarter. We won’t be stronger. We won’t be more productive. We won’t be masters of our environment. The AI will be all of this. We’ll just be relegated to passive, helpless spectators.
I think you've got a very strong view on what it will definitely look like, but I don't see any exploratory thinking on how we move to that point.
The world is chaotic. We have no way to predict that this is what it will end up like.
One possibility is that AGI remains a tool, without a mind or direction other than what we give it. We may manage to constrain it in that manner, and it is used to do "work" in the sense of solving mathematical problems, engineering problems, etc. We could end up with a reduced scarcity (if not post scarcity) scenario, where people do other things for meaning (come on, use your imagination here, plenty to imagine).
I could even imagine neural implants and regenerative technology. Meaning suddenly we're not so constrained by our physical limitations. What possibilities does that unlock?
I sometimes feel that the doomers lack imagination.
> One possibility is that AGI remains a tool, without a mind or direction other than what we give it
"Solve for the equilibrium" is the name of the game. "AGI remain a tool" is nice, until you notice that some people will use it lightly (solve some day-to-day problems I encounter when running my business, but I still do most of the work and decisions), some people will use it less lightly (I give you complete access to my mail box and my bank account — now run my business for me), and the first one will get absolutely crushed by the second in any market that is remotely close to free.
If you can ensure that "AGIs remains a tool", you just get to the same result, "AGI controls the future", with an additional step of nominally-in-charge humans rubber-stamping decisions made by AGIs.
> I sometimes feel that the doomers lack imagination.
It’s really the opposite.
I have asked a lot of people, "doomers" or not, "what is the positive vision of a future where there are AGIs around that dominate human in all cognitive tasks ?". All the answers feel superficially good and correct, until you actually use your imagination to poke and probe at it. It always fall apart. The only exception is those accepting : "humans not in control of the future is good actually, I don’t mind having infinite entertainment and zero responsibility". I have nothing to reply to that except : I personally abhor that picture.
Did you really think I have not enough imagination to generate by myself the hypothesis "what if AGIs stay tools ?". Of course I have. It’s just that I have used my imagination to dig deeper at that image. And the result is still not the pretty happy story it initially looks like.
Once you generate a dozen of such hypothesis, dig, and find the same outcome, you start to get the feeling that it is not exactly an accident.
It is not a lack of imagination. The problem is fundamentally extremely hard, possibly impossibly hard.
> The world is chaotic. We have no way to predict that this is what it will end up like.
I strongly disagree with that. Chaotic behavior on the micro-scale is not incompatible with predictability on the macro-scale. Almost all science is like that actually. Even tho you can’t predict when and in which individuals a particular mutation will arise, and the exact life path of those individuals, you can still have laws relating relative fitness and fixation rates, and the prediction that eventually, a beneficial mutation will fixate in a population.
Everything you know works like that. Electromagnetism does not require knowledge of individual photon/electron interactions to describe Maxwell equations. Thermodynamics do not require detail knowledge of molecular dynamics to relate quantities like pressure and temperature.
Today we summarize the native/settlers conflicts by "Eventually the Settlers displaced the Natives from their lands". And while it leaves some nuances out, it is not a bad summary. If you ask a human living this period, he will say what you say "world is chaotic, who can predict what the outcome will be ?". Actually the overall outcome was pretty obvious given the significant technological gap. Details may be surprising ; the big picture isn’t.
Edit : I focus too much on AGI on this message, not enough on humans. The problem is at the end mostly a human problem. Accepting diversity — true diversity, "I abhor your vision for a future, but I agree you should have a place to realize it with other like-minded people, as long as you don’t interfere with other people realizing their vision of the future" is the cure. If humans can accept genuine diversity, I believe there are pretty straightforward solutions that are robust to at least my level of poking and probing. I do not expect humans to be able to accept genuine diversity tho, and the more I reach out to other humans, the more I despair on this particular axis. It’s very easy to make a human say "Yes Diversity is Good" in the abstract. Then you start to paint some concrete pictures, and the mask fall off: "actually the diversity I like means accepting some people slightly, superficially different from me".
AI does fuck all about climate change or nuclear war or any other existential threat.
I'm not all that bothered about stopping AI progress, but if capital is overwhelmingly directed towards replacing blue and white collar workers with machines while climate change and every other existential threat is starved of capital then as a species we are doomed.
It's not like this is being done for the benefit of humanity either. Ask those same profit-obsessed oligarchs (or their media outlets) who are effusive about AI's ability to create a better future what will happen if the retirement age is raised in response or even kept the same. They 180 on their techno-jobs-killing-optimism so fast you will get whiplash: https://www.forbes.com/sites/dandoonan/2024/01/30/demographi...
Teachers in most countries are, at best, mid-wits with no practical or real world experience. I know teachers who barely passed math in high school who are now match teachers. It's like a basketball teacher who went to "Basketball Teaching School", who's never played basketball in his life, teaching kids how to play basketball.
The fact that you believe the ultra rich conspire to control and abuse the uneducated shows that you are part of that group of average people parents want their kids to stay away from.
True in terms of text, but not if you include video, audio, touch etc. Sure, one could argue that there is much less information content in video than their raw bytes, but even so, we spend many years building a world model as we play with tools, exist in the world and go to school. I don't deny humans are more efficient learners but people tend to forget this. Also, children are taught things in ascending order of difficulty, while with LLMs we just throw random pieces of text at it. There is sure to be a lot of progress in curriculum learning for AI models.
