rsrsrs86

9 ms·

91.
by rsrsrs86 3y ago
Best you can get is some bounds for whatever effect you want to predict and those bounds will open wide very fast. As for lots of data, doing proper econometric models requires really well gathered data. Data is published quarterly or yearl
92.
by rsrsrs86 3y ago
Economics is a science and what an economist calls economics is not what the market things an economist knows. In a science you are interested in testing hypothesis. Econometric/statistical models will generally serve to test general h
93.
by rsrsrs86 3y ago
One really cool thing to do is consider the agents to have limited computational capacity. There are a few NP economic problems, and you can’t expect agents to really be solving enormous computational problems in their heads.
94.
by rsrsrs86 3y ago
Undergrad Econ courses still present micro and macro economics as two disciplines, but that’s theory from the 70s at best. Graduate level economics packs a lot of economic models where for example the interest rate depends on the intertempo
95.
by rsrsrs86 3y ago
This. At least a general equilibrium model.
96.
by rsrsrs86 3y ago
One way to make this realistic is to add game theory to the simulation. Otherwise, the most important assumption in economics will be left out: rationality in decision making.
97.
by rsrsrs86 3y ago
This is cute but very different from what an economist would consider interesting.
98.
by rsrsrs86 3y ago
You will find active learning interesting. It starts by attributing a value to each point in your domain that it learns to match the expected gain in some performance metric. This metric can be learned so it’s okay if it’s really hard to sp
99.
by rsrsrs86 3y ago
Whichever has the most information wins. When the information has structure you can heavily exploit it for generating synthetic data. For this I point you to Apple Sim. It’s a repository of 3D models for interiors. You can generate many lay
100.
by rsrsrs86 3y ago
You can find the answer by trying the following: generate random data according to a model, fit a linear regression (or any other distribution), sample from the distribution, add it as to the training set.
101.
by rsrsrs86 3y ago
Asking as a newbie in database programming. Are vector databases built on top of rd trees?
102.
by rsrsrs86 3y ago
It failed to render on iPhone 13.
103.
by rsrsrs86 3y ago
Author writes about types being applied for ergonomy, which is interesting. Types add static information that can be used by IDEs. But there are other reasons for using type systems: correctness and performance (unboxing, optimization)
104.
by rsrsrs86 3y ago
I think it won’t ever hurt to know the topics in SICP. The interpreters and other components presented in the book will help the user navigate so many different problems, and the use of Scheme will instruct the reader in functional programm
105.
by rsrsrs86 3y ago
This is homophonic propaganda.
106.
by rsrsrs86 3y ago
You would be surprised to see what the older generations think about what we consider to be formal and professional, and how much of our lexicon started out as more informal vocabulary imported by upper classes. Are you sure your opposition
107.
by rsrsrs86 3y ago
It doesn’t need to be AGI to be dangerous. As argument, just consider how the next elections will be. GPT is at least a cannon for fake news. A fake News machine gun
108.
by rsrsrs86 3y ago
I heard about LISP in 2014. Back then I worked in finance and it looked so similar to spreadsheets.
109.
by rsrsrs86 3y ago
You will be sometimes wrong and sometimes right. So don’t punish yourself when you are not sure. Always be transparent over how confident you are in your analysis and what will be the consequences if you are wrong. Try to always focus on th
110.
by rsrsrs86 3y ago
This is awful and disgusting indeed. It is telling that the author defaulted to girlfriend. It shows utter alienation from what is ethically expected in or time. This is a case in point of a morally uneducated person messing with technology
111.
by rsrsrs86 3y ago
I think it is appropriate to thoroughly consider the ethical implications of whatever you are doing, with the support of one of the many researchers interested in ethics
112.
by rsrsrs86 3y ago
Did you coin the hyper porn term? I think it will catch on.
113.
by rsrsrs86 3y ago
This is very bad. “her” is not a particularly uplifting movie.
114.
The Chinese Room and GPT(ai.stackexchange.com)
2 points by rsrsrs86 3y ago|0 comments
115.
by rsrsrs86 3y ago
A truck has more bandwidth than network adapters technically
116.
by rsrsrs86 3y ago
The feature list is impressive
117.
by rsrsrs86 3y ago
I second that. It feels like an abacus after a while
118.
by rsrsrs86 3y ago
A jazz generator, pre-machine learning. It had instruments and each instrument had Markov Chains to control pitch and note duration. I then used a bunch of helpers and functions to write markov chains that sounded cool. I used it to play ba
119.
by rsrsrs86 3y ago
No proof inside?
120.
by rsrsrs86 3y ago
“Social AI chat” How can TechCrunch adopt buzzwords so fast?

More ›