CompleteSkeptic

6 ms·

1.
(KV) Cache Rules Everything Around Me(completeskeptic.com)
3 points by CompleteSkeptic 7d ago|0 comments
2.
by CompleteSkeptic 2mo ago
constrained decoding tends to make models dumber - this is why it's rarely used
3.
Scaling Laws, Honestly(completeskeptic.com)
5 points by CompleteSkeptic 2mo ago|1 comments
4.
by CompleteSkeptic 3mo ago
I helped train some of the first "magic" models at OpenAI[1] and it was a wild ride. We were a pretty sane + skeptical team and we weren't totally convinced the models were as general as they seemed, but the query that convin
5.
by CompleteSkeptic 4mo ago
TypeSafe AI | Software Engineer | ONSITE (San Francisco) | Full-time TypeSafe AI ( https://typesafe.ai/ ) is a well-funded stealth mode company making a new class of AI models which enable a reliable way to embed composable i
6.
by CompleteSkeptic 5mo ago
Is this the first time OpenAI has published comparisons to other labs? Seems so to me - see GPT-5.4[1] and 5.2[2] announcements. Might be an tacit admission of being behind. [1] https://openai.com/index/introducing-gpt-
7.
by CompleteSkeptic 6y ago
There actually is some work ( https://arxiv.org/abs/2003.13630 ) claiming that FLOPS are a poor measure of real-world performance - with some of the more recent FLOP-efficient models actually running slower than older mo
8.
by CompleteSkeptic 6y ago
This isn't directly relevant to PP-YOLO, but I'm surprised roboflow is still promoting "YOLOv5" - despite that model not having an associated paper and it not being made by the authors of the previous YOLO's.[1] The
9.
Improving Pulse Diversity in the Iterative Setting(medium.com)
1 points by CompleteSkeptic 6y ago|0 comments
10.
by CompleteSkeptic 9y ago
An appropriate quote: "If you can't intelligently argue for both sides of an issue, you don't understand the issue well enough to argue for either." There are many people for whom the declarative paradigm is a huge plus.
11.
Deep Neural Decision Forests Explained(topos-theory.github.io)
10 points by CompleteSkeptic 11y ago|0 comments
12.
by CompleteSkeptic 12y ago
I think he was referring to the fact that the request completes in finite time isn't a tight enough bound.
13.
by CompleteSkeptic 12y ago
State-of-the-art results no longer require unsupervised pretraining with autoencoders or RBMs, but back when unsupervised pretraining was more popular, top researchers were rationalizing that it was consider more biologically plausible than
14.
by CompleteSkeptic 12y ago
You may be interested in DBpedia, which gets its data from wikipedia but presents it in a structured format (for example, the page for Clojure: http://dbpedia.org/describe/?url=http://dbpedia.org/resource
15.
by CompleteSkeptic 12y ago
Yup! We were about to implement something like that, but we realized Reagent's model of having atoms store the state worked pretty well with javelin's FRP, and once you break down the state into atoms, the cursors aren't need
16.
by CompleteSkeptic 12y ago
I wouldn't go so far as to say that Reagent's model is perfect. Definitely pretty good, but we've had some issues that were somewhat non-obvious. If you're worried about it being maintained, there was actually a great di
17.
by CompleteSkeptic 12y ago
If any of my post sounded as if I was complaining about the lack of syntactic sugar, I apologize as that was not at all my intention. The best practices and syntactic ugliness didn't matter to us at all after the learning curve. We wro
18.
Trouble in Paradise with ClojureScript and React.js(diogo149.github.io)
28 points by CompleteSkeptic 12y ago|12 comments