count_duckula

joined 2 years ago

Assimil is a great way to throw yourself into the language. Each lesson is in the form of a conversation with audio and the pdf has the text along with the translation.

Listen to the lesson without reading the text first. This gets you used to the sound of the language. Then read the text, then text with audio, and finally read the translation along with whatever notes on grammar (don't focus too much on the grammar aspects when you are first starting out), neither on spellings. Later on you'll be asked to go back to earlier lessons and reproduce the text. The first phase is to internalise the language. You can read the recommended Assimil way of learning and adapt the steps to something that suits you.

Assimil works well along with Language Transfer for me. Assimil is more immersive while Language Transfer is more explanatory.

I find that music is also a great way for me to learn new words. Once I listen enough times to a song I like, I start humming along, maybe repeat a word or two. The important thing is to not stress yourself out trying to sing along to everything. Maybe there is a catchy chorus or bridge section that is memorable. That is good enough to form associations with words. In this, I find pop songs are a better genre because they are catchy.

Something else I do is have a notebook where the only rule I have for myself is: no using my native language. I try to explain new words to myself using a sketch or whatever basic words I have already learnt. Don't worry if you can't draw well, neither can I. But I can draw something that looks like a spoon or a hill. Then I label them, and bam I've already learnt two new words. To build on that, I can draw a stick figure on the hill - this has taught me the verb climbing. You get the general idea. Just don't stress yourself out trying to journal every new word you come across. Be creative and you'll have fun.

DivestOS has ceased maintaining Mull if I remember correctly. I use Ironfox on Android now.

I'd pay for Youtube if Google would guarantee to not track me. I donate to open source projects that I use, rotating every month whom I choose to donate to. I even donated to Manjaro recently even though I don't use it any more, but it was something I had used in the past and I was poor and couldn't donate them then. But I refuse to feel any guilt for watching Youtube for free.

I decided to use my phone to read after my Kindle broke. I use KOReader. I would like to not support proprietary products as much as I can, just my personal philosophy since the last 7-8 years. So far I can't complain about my setup, and there is one less device to lug around. I do miss an e-ink display sometimes though.

I do this too. I do not like anything on my Desktop, but I download files to there, which forces me to deal with them. Interestingly enough, my Downloads directory is a barren wasteland.

I switched after development ended on the package manager I was using on neovim. I didn't at that moment want to simplify my vimconfig, so I looked into helix.

Helix highlights the action you take, so if for example, you are deleting 5 lines, you select the lines first then hit delete. Sometimes the vim actions end up taking fewer keystrokes though. And I still prefer some ways vim does things. And I don't always agree with the kakoune inspiration of helix (I haven't used kakoune, just going by what the docs say) - for example, movement always selects text which I then have to unhighlight.

But the biggest reason I stuck to helix was sane LSP defaults out of the box with minimal config. I was tired of having to fix LSP related bugs in my vim config after package updates.

TLDR: saner defaults for helix + lazy to fix my bloated vimconfig.

[–] count_duckula@discuss.tchncs.de 2 points 9 months ago (1 children)

I've been using KOReader on my phone now, ever since my Kindle one day decided to be unrecognisable on my computer. Couldn't find a solution to fix it so it became a glorified paperweight.

The screen real estate is slightly degraded, but fuck if I give Amazon any more of my money. Besides, I get to store epubs as epubs instead of converting to that god awful mobi format.

I switched to Pipepipe from Newpipe because I wasn't sure Newpipe was being maintained. Pipepipe has SponsorBlock.

[–] count_duckula@discuss.tchncs.de 3 points 1 year ago* (last edited 1 year ago)

The reason they are blackboxes is because they are function approximators with billions of parameters. Theory has not caught up with practical results. This is why you tune hyperparameters (learning rate, number of layers, number of neurons ina layer, etc.) and have multiple iterations of training to get an approximation of the distribution of the inputs. Training is also sensitive to the order of inputs to the network. A network trained on the same training set but in a different order might converge to an entirely different function. This is why you train on the same inputs in random order over multiple episodes to hopefully average out such variations. They are blackboxes simply because you can't yet prove theoretically the function it has approximated or converged to given the input.

You want upvote, you get upvote. Now, would you like updog?