mxwsn

10 ms·

121.
by mxwsn 5y ago
I finished my PhD earlier this year and am taking a break until I start a research job next year. Right now I'm reading and enjoying the freedom to dive however deep into learning and thinking about research topics I feel are fundament
122.
by mxwsn 5y ago
That's so cool! It's inaccurate for Manhattan though ;)
123.
by mxwsn 5y ago
So, do you like ML?
124.
by mxwsn 5y ago
Systems thinking is great and all but why should causality be discarded? In the example about population size, a causal graph is built that connects walking distance to population size that certainly allowed for reasonable counterfactual re
125.
by mxwsn 5y ago
Good take but I'm not convinced. I would suspect (see epistemic status) that while correctness is easy to reason about and maintain by a single accountant using spreadsheets, since there is a clean mapping from a precise excel api to f
126.
by mxwsn 5y ago
(2011)
127.
by mxwsn 5y ago
Do y'all have interest in fighting games, specifically SSBM?
128.
by mxwsn 5y ago
That has got to be one of the worst possible use cases one could imagine. In page 33 of the appendix, the authors note that nearly 40% of RSA encryption keys created by Codex are clearly insecure.
129.
by mxwsn 5y ago
Tools are relatively low level compared to any single use case or field because they should universally support all uses cases or fields. The more narrow your field or use case is, the fewer resources there are to create a higher level lang
130.
by mxwsn 5y ago
Tangentially related, but I found this article on positions that humans understand better than engines fascinating: https://www.chess.com/article/view/10-positions-chess-engine... Especially Hikaru's game whe
131.
by mxwsn 5y ago
I am not a fan of the axiom of choice, and open-minded about truth values beyond true and false. So I'm really curious about the progress of constructive mathematics, regarding proving known results with non-ZNF axioms, and maybe estab
132.
by mxwsn 5y ago
The art collective behind this project, MSCHF, has done some pretty interesting stuff: https://en.m.wikipedia.org/wiki/MSCHF
133.
by mxwsn 5y ago
If you are goal-oriented and don't already work out regularly or go to the gym, I would highly recommend it. The physical challenge really requires focus and can clear your head and your body is hardwired to find exercising pleasurable
134.
by mxwsn 5y ago
Notion has been free for personal use with unlimited blocks for a year or two now.
135.
by mxwsn 5y ago
I've been eagerly anticipating this, and am very pleasantly surprised by the price. Excited to have pre-ordered in batch #1 :)
136.
by mxwsn 5y ago
The latest NFT craze is primarily for one-of-a-kind items. I'd rather see an org like Noora Health sell unlimited NFTs. What's exciting is the capability for decentralized apps to interact with NFTs. For example, other ethereum-b
137.
by mxwsn 5y ago
Using illusory afterimages to give a perception of non-natural color is a cool idea, but I'm not sure that it'll be a part of the future as far as Pixar is concerned. Pixar must care immensely about precisely controlling the viewi
138.
by mxwsn 5y ago
w/o paywall: https://outline.com/AMVLqP
139.
Mozart and Board Games(tablescopegames.com)
1 points by mxwsn 6y ago|0 comments
140.
by mxwsn 6y ago
Large model capacity enough to perfectly memorize/interpolate data may not be a bad thing. A phenomenon known as "deep double descent" says that increasing modeling capacity relative to the dataset size can reduce generalizat
141.
by mxwsn 6y ago
I agree with your logic but not your premise. Two alien species could effectively communicate if they happened to agree on a shared set of fundamental axioms. The axiom of choice is somewhat contentious here on earth, since it underlies the
142.
by mxwsn 6y ago
32 kb addressable memory, and 9 MHz cpu https://pico-8.fandom.com/wiki/Memory https://pico-8.fandom.com/wiki/CPU
143.
by mxwsn 6y ago
As a current MIT PhD student, this ideas are incredibly important to hear. Establishing an appropriate widespread sense of what normal is crucial for changing culture. More people need to talk about changing labs as well. A surprising fract
144.
by mxwsn 6y ago
Zuckerberg owns Facebook in a way that no other major company (to my knowledge) is owned. He is a monarch that can override any and all board politics. While power sloshes around in other companies by tug of war, Zuckerberg's word is f
145.
by mxwsn 6y ago
A tremendous loss. I would highly recommend Debt: The First 5000 Years to anyone here.
146.
by mxwsn 6y ago
https://www.slate.com/human-interest/2013/12/mandarin-chines...
147.
by mxwsn 6y ago
'From my experience playing a few games on it against decently strong opponents has been that at high levels it becomes "Chess, but it's much much easier to checkmate, and 2~3 times a game some time travel shenanigans happen.
148.
by mxwsn 6y ago
Less than 48 hours after that video was published, OpenRCT2 patched their pathfinding algorithm to not prefer some directions more than others: https://www.youtube.com/watch?v=b-5aX2oLOgU . The GitHub commit message says it
149.
by mxwsn 6y ago
Location: Cambridge, MA, US Remote: Yes Willing to relocate: Yes Technologies: Python, Pytorch, Dash plotly, scikit-learn, pandas, Heroku, Photoshop, Illustrator Résumé/CV: https://www.maxwshen.com Email: maxwshen [] gmail.
150.
by mxwsn 6y ago
My Ph.D. work has used machine learning to model CRISPR editing outcomes, see https://www.maxwshen.com/

More ›