coldcard

Attention over time

7-day rolling share of analyzed posts/comments mentioning this entity · 90 daily points ending 2026-10-08

Final complete day: 2026-10-08. Attention was 0.3% — this entity appeared in 27 of 9,401 analyzed posts/comments in the seven days ending that day (33 mentions).

Each point is a seven-day rolling window of complete UTC days; a point is comparable only when at least 100 analyzed posts or comments fall inside it. A day with no analyzed records counts as zero; missing coverage is marked and never shown as zero. The series ends at the latest complete analyzed day — it never extends into the current incomplete day. How attention is measured.

Recent mentions

1. 2026-10-08 21:48 UTCComment

   Using the 99-die method generated enough entropy to work just fine. Coldcard's problem was the generation of a seed-phrase *on device* without using the dice to introduce entropy. I would know: didn't use my Coldcard for a couple years, and wasn't paying attention to the news, but it was my key to a dozen grand or so, and it remained safe while thousands of people lost theirs.

2. 2026-10-08 21:45 UTCComment

   i think i could, but i don't have the time or desire to. for me it's quite obvious, that if you have 1024 pieces in a bag and you shake for a while, it will be truly random. if you don't agree, ofc feel free to use any other way. just be careful with algorithms written by few people that stand to gain millions by slipping a line in there that reduces entropy, as with coldcard

3. 2026-10-08 21:43 UTCComment

   if i put these pills in a bag, shake the bag for a few mins, i truly doubt that this will not be entirely random. where as with coldcard or other algorithms, there are people writing them that stand to gain millions by slipping in a clever line that reduces entropy somewhere...

4. 2026-10-08 21:28 UTCComment

   > simple sha256 of the string of numbers Ironically, this is exactly what ColdCard does, if you use the dice-only mode. Is it possible to mess that up? Of course, but it seems harder to do than with some other methods. As any HW wallet, it does need a correct SHA-256 implementation for multiple purposes, so a bug with it seems less likely. Also, here SHA-256 acts as what we call a debiasing function or randomness extractor. It's quite effective, and requires a minimal amount of input compared to some others. But being a cryptographic hash function, it's complicated, and would take forever to calculate by hand if you wanted to try that. A very simple method is the Von Neumann algorithm: For each bit, flip a coin twice, and if the flips are opposite, use either the first one or the second (as long as you make the same choice throughout), otherwise try again. The obvious downside is the inefficiency, since you need twice the flips as bits. Your method does seem a lot easier and more direct (but, like others, still having one little problem that you mentioned; see below). Just to clarify, since your language is ambiguous: It's only adding 1024 that is to be done according to the coin flip. That gives an extra bit, so the converted value that goes from 0 to 1023 is extended up to 2047. One problem, though, is it doesn't account for a biased die or coin, which the other ones mentioned do. (This does matter. I recently used cheap dice to generate entropy, and the bias was quite obvious when looking at the distribution. Fortunately, SHA-256 deals with this quite nicely, and only a few extra rolls are needed to make up for the loss of entropy.) > btw, the last 4 bits of the last word are a checksum so you can’t choose that one completely randomly. (The actual number of bits depends on the phrase length. Specifically, if you have *w* words, then the last *w*/3 bits are checksum, where *w* is one of 12, 15, 18, 21, or 24. But anyway...) Yeah, that's the one snag with all of this. Using coin flips and dice rolls, and converting in a simpler way, does avoid using a complicated hash function. But unfortunately, you still need SHA-256 to calculate the checksum. With 12 words, it's only 4 bits, so you can choose randomly and get a valid phrase one time in 16. However, the calculation still needs to be done to check it, which, practically speaking, means you need a computer of some kind at the time you create the phrase. Since you need to know the phrase is valid before putting money in, it's not something you want to leave till later.

5. 2026-10-08 18:08 UTCComment

   Nobody understands that what happened with Coldcard and other similar cases was just a bunch of amateurs selling solutions that work, but were flawed from day one. It’s clear that these devices were never actually audited. A specialist would have spotted all the weaknesses in less than an hour. It’s either intentional or just careless engineering.