cosmoharrigan

IB margin rates are currently 3.92%, for example: https://www.interactivebrokers.com/en/index.php?f=1595

I was curious to learn more about the project, and would like to share some quick feedback; the first thing I noticed was that the first page of the documentation which I opened has multiple spelling and grammar errors:

“Distrobutors have large part of their net worth is invested in the stock in hand. With ERPNext, you can always keep a birds eye view on your stock availability, replineshment, procurement and sales.“ https://erpnext.org/docs/user/manual/en/stock

Second, I noticed that the CI build is failing on GitHub.

Amazon's 1997 Letter to Shareholders: http://media.corporate-ir.net/media_files/irol/97/97664/repo...

> AGI16 is a futureism conference

I would like to point out that the AGI Conference series [1] is an academic conference, with refereed proceedings published in the Springer Lecture Notes on Computer Science [2].

[1] http://agi-conf.org/

[2] http://link.springer.com/book/10.1007%2F978-3-319-41649-6

For more details, refer to Schmidhuber's own website: http://people.idsia.ch/~juergen/ and his excellent review paper, "Deep Learning in Neural Networks: An Overview": https://arxiv.org/pdf/1404.7828.pdf

See "Logical Depth" as defined by Charles Bennett: http://researcher.ibm.com/researcher/files/us-bennetc/UTMX.p... as well as Chapter 7, "Resource-Bounded Complexity", from "An introduction to Kolmogorov complexity and its applications".

An extremely detailed treatment is presented in "An introduction to Kolmogorov complexity and its applications": http://www-2.dc.uba.ar/materias/azar/bibliografia/LiVitanyi1...

An overview with additional references from Marcus Hutter: http://www.scholarpedia.org/article/Algorithmic_complexity

Kolmogorov Complexity is also presented in Chapter 7 of the classic textbook "Elements of Information Theory": http://poincare.matf.bg.ac.rs/nastavno/viktor/Kolmogorov_Com...

If you happen to find it again, it would be interesting to look over it.

That is interesting. But the dry lease would survive cancellation of the CMI agreement, right?

What document are you referring to?

Units from third party (3P) sellers made up 49% of total units in Q2 2016, and Amazon only reports their commission as revenue for 3P sales rather than the total gross merchandise volume (GMV). A better comparison point is thus GMV, which was estimated at $225B for 2015[1], compared to $482B for Walmart.

[1] http://www.channeladvisor.com/blog/?pn=scot/deep-dive-into-a...

"Amazon’s goal, these people say, is to one day haul and deliver packages for itself as well as other retailers and consumers"

They could choose to offer parcel shipping services to any retailer as part of the Amazon Marketplace Web Services API, just as they currently offer the Multi-Channel Fulfillment API[1], which allows a retailer to fulfill orders from any sales channel using the Fulfillment by Amazon network.

[1] http://docs.developer.amazonservices.com/en_US/fba_outbound/...

In addition to the program to run their own parcel shipping network, referred to as "Consume the City", it also mentions an interesting experiment called "I Have Space", which allows Amazon to utilize excess capacity in warehouses owned by third parties. Note that Amazon has already launched "Seller Fulfilled Prime" for third party sellers which meet certain metrics.

Note that SkFlow is now TF Learn and is part of the TensorFlow repository: https://github.com/tensorflow/tensorflow/tree/master/tensorf... The skflow repository linked to in the poll is being deprecated.

Also see "The deep learning frameworks landscape, August 2016" from @fchollet: https://twitter.com/fchollet/status/765212287531495424

The deep learning frameworks landscape, August 2016: https://twitter.com/fchollet/status/765212287531495424

Are there any sources indicating what percentage of sales or units are from third-party retailers?

The original papers [1][2] for the deep Q-network used reward clipping: "As the scale of scores varies greatly from game to game, we clipped all positive rewards at 1 and all negative rewards at -1, leaving 0 rewards unchanged".

[1] https://www.cs.toronto.edu/~vmnih/docs/dqn.pdf

[2] http://home.uchicago.edu/~arij/journalclub/papers/2015_Mnih_...

For historical background on this part of the algorithm, called "experience replay", see this paper from Long-Ji Lin in 1992:

Self-Improving Reactive Agents Based on Reinforcement Learning, Planning and Teaching

http://link.springer.com/article/10.1023/A:1022628806385

as well as his excellent 1993 PhD thesis:

Reinforcement Learning for Robots Using Neural Networks

http://www.dtic.mil/dtic/tr/fulltext/u2/a261434.pdf

Here are a couple of examples from the Python deep learning ecosystem:

IDSIA, affiliated with Juergen Schmidhuber and many other leading ML researchers, has released Sacred, "a tool to help you configure, organize, log and reproduce experiments." https://github.com/IDSIA/sacred

MILA, affiliated with Yoshua Bengio and the Theano project, offers fuel, "a data pipeline framework for machine learning": https://github.com/mila-udem/fuel

Amazon is pushing very hard to increase Fulfillment by Amazon coverage across the catalog, and is also releasing new products such as Seller-Fulfilled Prime. The customer experience is tightly controlled in both cases.

It's interesting to note that Amazon third-party seller unit share is at 48% of total units shipped, which is a record high. It has grown by 1% per quarter for the last 4 quarters in a row.

Jake Vanderplas previously wrote an excellent blog post about Python performance and scientific computing: https://jakevdp.github.io/blog/2014/05/09/why-python-is-slow...

> meticulously constructed symbolic reasoning world (...) connect such disparate symbols together and maintain that connection strongly for long periods of time

Out of curiosity, could you give an example?