AI Overview:

00:00 – 00:26 In this candid episode of the Phrasing podcast, the creator steps away from standard “language log” format to share hard-won insights on phrasing philosophy after two years of building and using the app himself. The core message: long expressions (20+ words) outperform short, I+1 sentences for genuine language acquisition.

Product Evolution & the Turning Point (00:26 – 02:38)

• Original Phrasing used classic spaced-repetition I+1 logic: one new piece of information per card.

• Early testing showed it was “easier to get started” but delivered poor long-term retention.

• A single 25-word Chinese sentence (learned while knowing zero Chinese) flipped the approach forever.

The Chinese Sentence Breakthrough (02:38 – 03:51) At 02:41 the speaker recites a full long Chinese expression he still remembers months later—tones imperfect, but the sentence intact. This subconscious absorption proved short, targeted cards become routine and forgettable, while long ones force deeper processing.

Why Short Expressions Cheat Your Brain (03:51 – 05:15)

• Miller’s Law (7 ± 2 items in working memory) explains the problem.

• 4–10 word cards quickly become predictable “fill-in-the-blank” drills.

• Users report knowing entire short sentences without understanding individual words—classic rote memorization, not language learning.

Benefits of 20–30 Word Expressions (05:15 – 11:42)

• Forces active listening, deduction, and contextual reasoning.

• No “cheating” possible—brain cannot instantly complete the sentence.

• Creates the famous “click moment” (10:42) where a sentence suddenly makes perfect sense after months of reviews.

• Expressions remain useful for years: pronunciation, intonation, speed, and advanced nuance.

• One high-quality extraction = months of learning + years of review.

Phrasing’s Unique Technical Advantages (12:16 – 13:57)

• Word-by-word teaching with phrase prioritization and language-graph interlinking.

• Contextualized learning (no direct English word shown—users must map meaning themselves).

• Full-sentence testing on any word, any phrase, any position.

Pricing Model That Rewards Quality (08:30 – 09:08) Charged per expression (not per word), so longer cards are actually cheaper per unit of learning value—strong incentive to create rich, long-form content.

Final Advice (15:22 – End) The host openly admits his first version followed conventional Anki-style advice and was suboptimal. After real-world testing he reversed course. His plea: don’t fear long expressions. Dive in, trust the process for 1–2 weeks until spaced repetition warms up, and experience the rewarding “unlock” moments that short cards never deliver.

SEO keywords covered: Phrasing app, long expressions language learning, spaced repetition long sentences, contextual language acquisition, Anki alternative, effective language learning method, deep retention fluency.

AI Transcript:

[00:00] So today I’m going to talk about something a little bit different. This isn’t going to be a language log, but more just a chat about phrasing and some of the philosophies and the learning method. Because I’ve been working on phrasing for over two years now, and in that time the product has changed a lot.

[00:26] I made some decisions along the way that I wouldn’t have expected myself to make ahead of time, and I think I’ve made plenty of controversial decisions. So I’d like to dig into them and explain the rationale behind them. Today I especially want to talk about the use of long expressions.

[00:55] In phrasing, you create expressions anywhere from a couple of words to a couple of sentences, and we teach them to you one word at a time, two words at a time, three words at a time, up to full phrases. We slowly build up. The way phrasing is set up, the financial incentives, and the way I encourage it, it is more beneficial to create longer, more complicated expressions.

[01:32] When I was first building the application the idea was you’d upload content and it would find I+1 material—classic spaced repetition that shows you just one novel piece of information beyond your current level. That was the very first version, and I used it myself.

[02:17] I quickly found it didn’t have very good long-term value. It was easier to get started, but easier doesn’t mean better for ultimate learning. What really changed my mind was when I added one very long Chinese sentence. I don’t speak a word of Chinese, so I thought it would be suboptimal.

[02:51] I tried to make everything shorter and more beginner-friendly, but I left that long Chinese expression in my library anyway. Over the next couple of months I eventually learned the entire thing. Keep in mind I still don’t speak Chinese, but I can recite it.

[03:13] It’s a very long sentence. I wouldn’t have expected that information to stick, yet it did—very subconsciously, very naturally. My brain just held onto it. Meanwhile, all the shorter four- or five-word expressions I created became routine almost immediately.

[03:51] With short sentences my brain just completes the phonemes. There’s this idea in memory science that you can only hold about seven items in your head. My theory is that limit starts as seven letters, then seven syllables, then seven words, then seven phrases. If it’s less than seven words, your brain saves the whole chunk together and you’re not actually learning the individual words you’re filling in.

[04:45] Now that I’ve been using the app for over six months, almost all my expressions shorter than ten words are completely predictable. If even one word is missing I already know the rest. I can fill them in without knowing what the words actually mean—which is not ideal.

[05:15] Compare that to the longer expressions. With a twenty-five-word sentence I don’t immediately know which word is missing, especially at lower levels. My brain can’t cheat. I have to actually think, listen, read the context, and figure it out. That effort is incredibly valuable for learning.

[05:57] You can’t blast through hundreds of cards per hour like some Anki users do. When you use phrasing as intended—with longer expressions—it takes time to parse, recall, and truly save the information. That time and work is what makes the learning stick.

[06:53] Because of that one Chinese sentence I started shifting all my expressions to longer sentences. I almost never create anything under twenty words anymore, and it has been remarkably helpful. Even in languages I already speak, these long expressions still sit in that seven-phrase short-term memory bucket, but I haven’t been able to cheat my way through them yet.

[07:36] It also leans into the technical benefits of phrasing: teaching by full phrases, interlinking cards, connecting everything to our language graphs so you have this interconnected web of language. Your expressions actually get shelf life—they can last you years of learning.

[08:04] You extract an expression once and you’ll be learning from it for years. That’s exactly what you want. I don’t want a card I memorize and then trash; I want something I can tease information out of slowly over time while my brain fills in all the grooves.

[08:30] We charge per expression, not per word or character. A twenty-word expression actually costs more to create, but because you’re only charged once, it’s dramatically more cost-effective and gives you far better value. That’s the incentive I want to give users—because it provides such a superior learning experience.

[09:25] It’s counter to common Anki advice, where you put things you already understand and want to remember. Phrasing is different: you put things you would like to understand one day, and we teach you slowly by unpacking them over time. Don’t worry about understanding the whole thing from day one.

[10:10] You’ll struggle at first, not getting a single word. Then you get some rote memorization, a little cheating from context, but you keep going. Then one day the sentence comes up and suddenly the entire expression makes sense. There’s this almost literal “click” where everything slots into place.

[11:21] That feeling—when a large, complicated sentence suddenly makes perfect sense after months—is super rewarding. It saves the information far better than any short sentence where you can fill in every word without thinking.

[11:54] This is why I encourage longer sentences on phrasing. It’s how the app is built to work. You won’t see this approach on other applications because they don’t have the same teaching mechanisms, prioritization, contextual learning, and phrase-level testing that make it possible.

[12:59] We used to show the English translation of the word you needed to fill in. I did it all the time because that’s how other apps work. But that creates non-contextualized learning—your brain takes the lazy route and builds one-to-one relationships instead of truly figuring it out. So we removed it.

[13:12] Now you see the full translation, we highlight the corresponding phrase, and you have to map which English words match which target-language words and figure out what’s missing. That extra step—four or five mental steps—solidifies the learning. It’s called contextualized learning and it’s massively helpful.

[13:43] I built a way to learn a language where you aren’t cheating, where it is more difficult, weirder, different, and unique. But if you stick with it, it is insanely rewarding and insanely effective. It’s great to study the same expression for months and still have more to learn.

[14:30] I have thirty-word Arabic sentences I’ve been reviewing for four months. I still get words wrong every day and I haven’t fully understood them yet—and that’s fantastic. One expression can give me a year’s worth of learning and ten years of review.

[15:16] So that’s why we have longer sentences and why you should try them. This is also a plea to anyone who sees the app and thinks short expressions and massive volume are better. I understand that point of view—my first version followed all those principles. But through experience I found it was suboptimal and highly gameable.

[15:57] I made an ungamable system that is just language learning. Give the longer expressions a shot. Don’t be scared of them. They feel overwhelming at first, but you don’t need your hand held that much. Just dive in, trust the process. It takes about a week or two for the spaced-repetition system to warm up and start making great recommendations.

[16:53] Those expressions will start to make sense. In six months you’ll look back at a library full of long, complex sentences you can practice saying, work on your accent and intonation, and speak quickly. It’s worth it. That’s what I got.