lysozyme

9 ms·

31.
by lysozyme 3y ago
Do yourself a favor and skip right through to the Twitter link to another link to this excellent post by Yoav Goldberg [1] on the actual reason that training new models on ChatGPT output in the manner of supervised learning (in contrast to
32.
by lysozyme 3y ago
If I may suggest a way to train yourself up on this, I’d like to suggest an approach which I learned from the writer Francine Prose in one of her composition courses. The method is simply this: for half a year, every week, go and find the s
33.
by lysozyme 3y ago
I’m reminded by a speech by Frederick Douglass [1] that >This struggle may be a moral one, or it may be a physical one, and it may be both moral and physical, but it must be a struggle. Power concedes nothing without a demand. It never d
34.
by lysozyme 3y ago
My introduction to calculus was “Calculus Made Easy” by Silvanus P. Thompson and I always liked math profs who actively worked to show math for what it is: useful, beautiful but not about the symbols or the jargon. “Any fool can calculate!”
35.
by lysozyme 3y ago
>if LLM-generated content outpaces human-generated content, the useful data proportion will diminish I guess I’d ask why the author thinks that training LLMs on their own output will make them worse. Like, if the problem is that LLM-gene
36.
by lysozyme 3y ago
The translation of the Gagarin’s letter in this article, which explains that it’s “translated by Yves Gauthier … author of a magnificent biography in French on Youri Gagarine” reads better than the translation in TFA, in my opinion
37.
by lysozyme 4y ago
It’s a fun time to be interested in AI for proteins. Every new ML model type is inevitably tried out on proteins. As the functional molecules of life, proteins are uniquely important and fundamental to every process in biology. As the targe
38.
by lysozyme 4y ago
Interesting that they enriched the training data by asking people to point out YouTube videos in specific languages for which they needed data > YT-513-U: We create an additional dataset called YT-513-U to ensure coverage of lower resour
39.
Pandas author: “Ten things I hate about Pandas” (2017)(wesmckinney.com)
1 points by lysozyme 4y ago|0 comments
40.
by lysozyme 4y ago
Blood clotting in mammals is done by platelets, which float around in the blood. Platelets are kind of like (simple) cells in a way. Platelets have a membrane. Inside the membrane are the molecular tools—that’s what I usually call enzymes d
41.
by lysozyme 4y ago
There’s a line of thought that connects the use of leaded gasoline with US crime rates. [1] The idea is that exposure to lead causes people to have problems that lead to violent behavior later in life. The evidence presented includes the co
42.
by lysozyme 4y ago
If someone with access to the GPT-4 vision model wants to ask what’s funny about this picture, I’m sure people around here would be curious to find out
43.
The state of Computer Vision and AI: we are really, really far away (2012)(karpathy.github.io)
3 points by lysozyme 4y ago|2 comments
44.
by lysozyme 4y ago
Especially funny since the author, Andrej Karpathy, wrote at the end of the 2012 article that >we are very, very far and this depresses me. What is the way forward? :( Maybe I should just do a startup and was a founding member of OpenAI
45.
A separate app for classical music makes sense(twitter.com)
2 points by lysozyme 4y ago|0 comments
46.
Hip-hop producers interviewed about the beloved SP1200 sampler(villagevoice.com)
5 points by lysozyme 4y ago|0 comments
47.
Library Music Is Changing the Sampling Game in Hip-Hop(djbooth.net)
1 points by lysozyme 4y ago|0 comments
48.
by lysozyme 4y ago
>sometimes it's better for the public not to know how bad things have gotten behind the scenes Better for whom? It seems to me that institutions that serve the public should do so with transparency. The idea that “the experts” know
49.
by lysozyme 4y ago
I grew up listening to my dad’s vinyl records and remember buying my first CDs in the early 2000s. I also remember borrowing backpacks full of CDs of jazz and classical music from my local library when I was in high school and ripping them
50.
Bioinformatics with Rust(github.com)
2 points by lysozyme 4y ago|0 comments
51.
A Little Bit of Rust(betterprogramming.pub)
21 points by lysozyme 4y ago|0 comments
52.
Create a cross-platform launchbar app with Rust(blog.spyglass.fyi)
2 points by lysozyme 4y ago|0 comments
53.
by lysozyme 4y ago
Doesn’t necessarily make it a bad idea to try again. It seems that photolithography as a process is stretched to its limits. Easy to see the diminishing returns. The electron beam lithography is a different kind of process. It seems like it
54.
by lysozyme 4y ago
To me, the whole point of reading imaginative fiction like Dahl’s work is that it transports you to some other place. A different world with different people in it. The author created that whole world with language, and you’re getting it th
55.
by lysozyme 4y ago
Tangential to the original question, just want to note that having such realizations constantly is one of the primary benefits that accrues if you expend the energy to constantly keep yourself in a learning/teaching mindset. The primar
56.
by lysozyme 4y ago
Former graduate student here. The most important thing to remember is that every paper has a main point, and everything in it is reinforcing that main point. Figure out what the point is. The whole story is in the figures. Start with the ti
57.
by lysozyme 4y ago
Discussed last year too (329 comments) https://news.ycombinator.com/item?id=25854523
58.
Creating dynamic soundscapes in Grand Theft Auto (2005)(soundonsound.com)
1 points by lysozyme 4y ago|0 comments
59.
by lysozyme 4y ago
MMseqs2 stands out as great software in the world of biology. It’s a useful, fast, powerful, flexible command line tool for searching and clustering biological sequences. Unlike a lot of free software in biology, MMseqs2 has a decent user e
60.
MMseqs2 – an example of great software for biology(github.com)
2 points by lysozyme 4y ago|1 comments

More ›