liliumregale

6 ms·

1.
by liliumregale 26d ago
Is it unfortunate, or just a different priority in a different place? Those are also cities with a great public option for transit (unlike, say, Dallas or the Bay Area where different municipalities squabble). One might laud New York drivin
2.
by liliumregale 29d ago
That's a really unfortunate framing. The cat isn't out of the bag; otherwise, there wouldn't be efforts to scale up data centers to support forecasted demand. The inevitability narrative is a self-fulfilling prophecy, and it
3.
by liliumregale 29d ago
Could it be that the goal is not to be competitive, but to be responsible?
4.
by liliumregale 1y ago
I'm sorry, this article reads like AI slop. It has all the hallmarks: grandiose writing ("everything changed"), the classic "It wasn't X, it was Y" (about five times in the first minute of reading the article),
5.
Type-Compliant Adaptation Cascades: Adapting Programmatic LM Workflows to Data(arxiv.org)
2 points by liliumregale 1y ago|1 comments
6.
by liliumregale 2y ago
Regularization as a concept is taught in introductory ML classes. A simple example is called L2 regularization: you include in your loss function the sum of squares of the parameters (times some constant k). This causes the parameter values
7.
by liliumregale 3y ago
We absolutely should treat it just like any other software tool.
8.
by liliumregale 3y ago
Let's distinguish between papers and preprints, please. arXiv has contributed to a blurring of the distinction. The arXiv preprints are useful but should always be taken with a grain of salt. There is nearly no filtering done on things
9.
by liliumregale 3y ago
This repo is a joke, right? I'd be embarrassed peddling this as AGI. We're a long way from AGI existing at all. Even if you disagree, it's agreed upon that we're not there yet. For this repo to call its offering AGI is e
10.
by liliumregale 3y ago
Google absolutely has their own internal models that do exactly this. It wouldn't surprise me if Microsoft indeed does have an internal Copilot that is trained on their data, but even on the smallest risk that they leak their code, the
11.
by liliumregale 3y ago
Well…one that peeks at the test set labels. https://kenschutte.com/gzip-knn-paper2/
12.
by liliumregale 3y ago
Further analysis shows that it doesn’t perform well at all—successes are tied to things like test set leakage. https://kenschutte.com/gzip-knn-paper2/ This paper isn’t any surprisingly effective result. It’s thoroughly
13.
by liliumregale 3y ago
Yes - it's mentioned, but doesn't the framing below make it sound like they're still advocating for this paper? > In essence, it's advisable to take the paper’s reported figures with a grain of salt, particularly as t
14.
by liliumregale 3y ago
The paper has recently been called into question for overestimating their performance relative to BERT: https://news.ycombinator.com/item?id=36758433 . Might be good for the blog's author to take this into account in th
15.
by liliumregale 3y ago
The title wordplay dates back to at least Drew McDermott's 1976 essay "Artificial Intelligence Meets Natural Stupidity" [0]. The intro is phenomenal. --- > As a field, artificial intelligence has always been on the border
16.
by liliumregale 3y ago
By your first paragraph's argument, the semantics are in the Transformer, not the tokenizer. And yes, what they do helps on their two test tasks. I'm not disputing that. It's the fact that there's no scholarship here. Th
17.
by liliumregale 3y ago
I was being generous - stemming is poor man's morphology. Empirically useful (ask the IR folks) but incredibly heuristic.
18.
by liliumregale 3y ago
I'm going to add a contrarian take here: this preprint is not a research paper. While it's nice to see that there is an improvement here on their one task, this is not "semantically" driven tokenization. It's morpho
19.
by liliumregale 3y ago
Yep! Percy Liang in an interview with Chris Potts said he sees BERT and ELMo as foundation models.
20.
by liliumregale 4y ago
This is a God (language)-of-the-gaps argument: we can't figure out this rarely language, but maybe we can figure out an entirely unattested language instead, and also learn the correspondence between it and Linear A. Deep learning can
21.
by liliumregale 5y ago
Is that a pun, because this was done by Pearson?