Blog
The MemDecks blog
Posts about how vocabulary learning actually works and how MemDecks implements it — including the parts where MemDecks made a trade-off and lost something. Everything here is written by the person who builds the product, and technical claims about MemDecks describe what is live, with a date.
All posts
- How MemDecks decides when to show a card againNo scheduler, no intervals, no due queue. The card HP mechanism in full, with the numbers, and what it trades away.
- What learning a non-Latin alphabet actually takesThe letters take days. Reading takes months. What the middle stage needs, grounded in Armenian and Greek.
- Why vocabulary sticks better inside a phrase than on a flashcardWhat a two-sided card leaves out, what a sentence supplies for free, and why Phrase Builder lessons are five words long.
- Games as retrieval practice: what the format buys you and what it does notFour things games genuinely add, four things they cannot do, and the failure mode nobody mentions.
- Choosing your first 200 words in a new languageWhy frequency lists mislead beginners, and a split that works: 130 structural words plus 70 of your own.
- Why a translation app is not a weekend projectThe demo takes an afternoon. What it skips: word identity that survives leaving the database, one pivot language instead of 240 pairs, four providers per card, and progress that outlives the content.
- Why language apps get boring, and what games actually do differentlyPlayers memorise hundreds of things without trying. What transfers to vocabulary, what cannot, and why streaks are the wrong fix.
- Vocabulary first, grammar later: what the first thousand words changeWhy rules land better on words you already have, the three things you cannot postpone, and the failure mode the method produces.
Every post is written by Sergey Belousov and carries its own publication date.
What this blog is for
Two things. Explaining the reasoning behind MemDecks in enough detail that you can disagree with it — which means stating the costs of each decision, not only the benefits. And writing about vocabulary learning generally, where the useful answer is often "it depends" and most published advice pretends otherwise.
What it is not for: announcements, feature releases, growth numbers, or roadmaps. Product documentation lives in the docs, and it is kept current rather than announced.