vjerancrnjak

6 ms·

61.
by vjerancrnjak 9mo ago
So, whole english wikipedia in <1 second (~20GB compressed)? Or is it now a lack of proper pipelining where you first load, then uncompress, then chunk, then write? Add a nice strong linear model on top like vowpal wabbit and chunk at 10
62.
by vjerancrnjak 9mo ago
I catch myself sometimes writing the model it is confused and it should just assume what I am writing is true and continue reasoning from there. Sometimes I am actually right but sometimes I am not. Not sure what happens to any future RL an
63.
by vjerancrnjak 9mo ago
Librefm or listenbrainz as well
64.
by vjerancrnjak 10mo ago
They have to start video recording the workouts. Even veterans in this academic field sometimes design workouts that just can’t be done to failure if you’re actually going to failure over 6-12 weeks. Even 1 workout sometimes has so many set
65.
by vjerancrnjak 10mo ago
Asymmetry is normal and you cannot address it (outside of repeatability of movement, aiming for no form degradation during high load). As long as your movement does not degrade horribly, asymmetry is fine. Even before strength training, you
66.
by vjerancrnjak 10mo ago
I cant even register.
67.
by vjerancrnjak 10mo ago
Query expansion works better.
68.
by vjerancrnjak 10mo ago
Same behavior on raspberry pi 5. Might be just lack of arm optimizations.
69.
by vjerancrnjak 10mo ago
I wonder what happened to https://en.wikipedia.org/wiki/Suggestopedia It went from 200 words of long term retention to 3000 words. Spaced repetition is the most inefficient efficient way to learn the vocabulary. I neve
70.
by vjerancrnjak 10mo ago
Made a scrobbler program for cmus when I switched to it. [0] Comparison of listeners really nails the recommendations. Similar minds like to listen to similar things. 0: https://github.com/vjeranc/cmus-status-scrobbler
71.
by vjerancrnjak 11mo ago
But, UK has hummingbird hawk-moth which is equally magnificent. With the right flowers you can have a flutter of hummingbird hawk-moths, creating a similar appearance to the bird.
72.
by vjerancrnjak 11mo ago
Inspired by this, went to look into how much performance I can squeeze and turns out Qualcomm software practices are so bad that I can’t do much but accept old software. It sounds like Qualcomm has to do everything from scratch on their hid
73.
Suggestopedia: Language learning method with Baroque music and relaxation(en.wikipedia.org)
1 points by vjerancrnjak 11mo ago|0 comments
74.
by vjerancrnjak 11mo ago
Interestingly, Knuth swapped dancing links for dancing cells and implemented a bunch of SAT solvers (and counting solutions to polyomino tiling problems) using zero suppressed binary decision diagrams. So algorithms X has newer and more eff
75.
by vjerancrnjak 11mo ago
If you want to participate, paste this into a LLM and ask it questions. Might be a barrier that has nothing to do with language.
76.
by vjerancrnjak 11mo ago
It’s interesting that Hellen Keller describes her experience, before language acquisition as timeless, no perception of time at all. https://news.ycombinator.com/item?id=40466814 I’m curious if this experiment actually test
77.
by vjerancrnjak 11mo ago
ORM gives: 1. migration 2. Validation when inserting 3. Validation when loading 3.1. Serialization 4. Joins 5. Abstracts away db fully if you want, not use db specific features 6. Lazy loading as an encapsulation promoting mechanism None of
78.
by vjerancrnjak 11mo ago
It’s just pipelining. Encryption is free compared to reads or writes to storage.
79.
by vjerancrnjak 11mo ago
No. You can go further and ask me if I imply lazy loading is mandatory. Imagine what happens when lazy loading turns off. You lose encapsulation. How will OOP work now if you have to reason about your whole call stack and know what exactly
80.
by vjerancrnjak 11mo ago
OOP is not simplifying CRUD or DB ops because you want to batch. You don’t want lazy loading. You don’t want to load 1 thing. You don’t want to update 1 thing. You want to actually exploit RETURNING and not have the transaction fail on a si
81.
by vjerancrnjak 11mo ago
Markov Random Fields also do that. Difference is obviously there but nothing prevents you from undirected conditioning of long range dependencies. There’s no need to chain anything. The problem from a math standpoint is that it’s an intract
82.
by vjerancrnjak 11mo ago
If you learn with Baum Welch you can get nonzero ood probabilities. Something like Markov Random Field is much better. Not sure if anyone managed to create latent hierarchies from chars to words to concepts. Learning NNs is far more tinkery
83.
by vjerancrnjak 11mo ago
Some countries have 0 rate if sold after x years. Although many had it in the past, they’ve started eliminating it.
84.
by vjerancrnjak 11mo ago
How are they faster? I don’t think any ELO report actually comes from participating at a live coding contest on previously unseen problems.
85.
by vjerancrnjak 11mo ago
The worst offender is toString which has different types between objects and is everywhere by default.
86.
by vjerancrnjak 11mo ago
I was doing this a lot in US whenever I’d see construction work speed limits and had similar experience. Realized no one cares about these custom signs.
87.
by vjerancrnjak 11mo ago
Let’s solve the problem with ingenuity, not brutality. Brutality is lazy. Dealing with animal overpopulation by introducing carnage. I know nature is indifferent but Homo sapiens should not be.
88.
by vjerancrnjak 11mo ago
Hard to find. I ask for advertising an ipv6 address and they don’t want to do that. Even though they give me an ipv6 prefix.
89.
by vjerancrnjak 11mo ago
Won’t work if behind CGNAT or will be insanely slow. Even ipv6 is not advertised sometimes. I miss the days when I could ssh to my computer with ddns.
90.
by vjerancrnjak 11mo ago
Given that Zig has functions which can return functions, maybe you could capture the top level io and allocator and return a struct with a bunch of functions that now have the top scope io and allocator visible. Don’t know. That’s how peopl

More ›