HN user
While you may know the guy better, notice this Druck comment comes at a time when the long term buy-and-hold thinking is at its weakest point. A slightly respectable guru saying this sort of things is enough to make some people sell.
This is the best comment in all. HN folks don't understand the purpose of this article. You think Druckenmiller is so bearish? for the next 30 yrs? Not a chance. He must have a huge short position on the S&P500, and he likes you to sell your stocks so he could cover at a lower level.
I heard this so many times. These goddamn CEOs are so myopic; all they care is next quarter's earnings. After the fall of every megacap, tech or otherwise (recently GE), someone always blames them for short sightedness. Is that really? Well, many CEOs of many huge corporations just plan for the next quarter, and they seem to be doing just fine.
Bravo!
To rephrase the famous quote attributed to G. Box,
`All languages are wrong, some are useful.`
And R is one of them.
Bravo! This is exactly right.
I went to this really nice talk by Ghahramani on Bayesian Nonparametrics at 2011 NIPS. Worth seeing polya urn model in action.
I used to cringe whenever I see a prompt to upgrade driver. But it's getting better. Version 460.39 hasn't caused any pain so far.
When a product originates from China, it must be full of spyware. Let's don't even question whether that's true or even reasonable. Then imagine one day, China's scientists invented a new semiconductor, a new material to make chip, or a new technological product that's order of magnitude better than everything out there.
My question: are you going to avoid it like the plague; don't buy anything made of it, which may be in a lot of the things you touch. Btw, that's not very far-fetched, judging from the number of breakthrough research coming out of that country.
Title is totally misleading. I use Rstudio, which is an IDE, and there's none of those "terrible" features in Copilet. Does it make my code terrible?
I wonder if the site is down because of the interest from folks on HN?!
I'm stunned why this comment is downvoted. It's exactly what's happening in US relationship with China.
The view expressed in this blog and many of the comments on HN is a prime example of survivorship bias[1].
In 1921, you did not know DJI would have done so well over the next 100 yrs.
A very real counter example is the Japanese stock market. The Nikkei peaked at 39,000 in 1989. Thirty yrs later, it's only 28,000. Many blue chip stocks on Tokyo Stock Exch have never recovered their previous high.
there are at least three books on this period,
Lefevre, E. (2004). Reminiscences of a stock operator (Vol. 175). John Wiley & Sons.
Kramer, C. (2000). " Devil Take the Hindmost: A History of Financial Speculation" by Edward Chancellor (Book Review). Finance and Development, 37(1), 53.
Mackay, C. (2012). Extraordinary popular delusions and the madness of crowds. Simon and Schuster.
WSJ has a link to its digital archive, which only goes back to May 1996. And unfortunately it doesn't have API so getting articles out of site for NLP is a pain. Let me know if you have code for that. I'm working on something similar
I just bought two yubikeys; a month later, I returned both. Here is a (major) problem. On a ubuntu box, I installed `libpam-u2f` and set it up for one user account. Turns out it breaks all other user accounts on this ubuntu box, meaning no other user could log in without the key. I contacted their support. No solution.
I just realized Vitalik never finished college. That's even more impressive for what he's done.
See my link to his ICML talk above.
See my link to his 2013 ICML talk above. There's a very nice photo of L'Avion III de Clement Ader, a plane modeled as a bird.
I found the source. He and Ranzato gave a talk at ICML 2013
https://cilvr.nyu.edu/lib/exe/fetch.php?media=deeplearning:2...
Slide #9
Let's be inspired by nature, but not too much
It's nice imitate Nature,
But we also need to understand
For airplanes, we developed aerodynamics and compressible fluid dynamics.
Question : what is the equivalent of aerodynamics for understanding intelligence
The better/faster part is my embellishment :)
See my link to his ICML 2013 presentation above.
While Hinton's view need to be noted, I heard a quote attributed to Yann LeCun, something like,
"If you want to learn flying by modeling the biology of birds, you're doing it wrong. Just look at today's airplanes. They have no resemblance to birds at all. Yet they're million times better and faster than any bird."
....and they are Chinese!
What you said about daily data is precisely what makes stock mkt so interesting and challenging : nonstationarity.
"outliers and events outside of the data, news" : these are precisely the stuff your models need to learn, and the fact that you consider them noise tells me most folks have no clue how to predict these "noise".
Ther are plenty of free datasets out there. You can get upward of 10 yrs of daily OHLC stock data on yahoo finance. The amazing thing is yf has S&P 500 index since 1927. Free!
Quandl has many free, or low cost stock market/commodity datasets.
I'm not sure what you mean by a "simulator". One of the greatest challenge applying RL to stock mkt is precisely that the market itself is not a MDP.
"most ML people are not really interested in trading"
You couldn't be more wrong on this. Stock market trading has the lowest barrier to entry of any endeavor. All you need is $1000 and Robinhood account, which you can open one on your phone in 5 min or less.
I've been following HN for a while, every time someone comes up with a trading algo or posts a link to algo, there's were hundreds of upvotes, lots of comments.
While it's so easy to dismiss someone's work as flawed (sure, backtest is illusional but do you have anything better?), which I think it may be, I always read it and try to understand what they're up to. Sure, academic folks may have no clue about market microstructure and other complexities, but if they could solve, or make some way toward solving the difficult problems in stochastic processes, they're already worth my effort.
Good and convincing example. thanks for this robotic perspective.
references?
But then how do you simulate, or imagine all the possible ways of falling and all the possible places this could happen? You have one sample, that's all. Ultimately, you have to use domain knowledge, but domain knowledge comes from observed data. High fidelity comes from having a lot of data. This takes you back to ground one.