usmannk

> Kaggle encourages a domain agnostic approach to modeling, in the sense that participants use sophisticated machine learning and statistical methods but typically have no domain expertise in the underlying data.

Yes this is accurate and put very well. This is so much the case that if you have a strong background understanding of the field, the ML part can actually be picked up quite quickly or contributed by someone else. There are a few notable users who are both domain and ML experts and they tend to absolutely clean up in their field. I'm thinking of a couple of med students in particular who are formidable in every medical imaging competition.

As a frequent Kaggler (perhaps too frequent... it's a bit addicting, in a way I'm sure others on HN will understand), I was fairly intrigued to see this one pop up in the competition list a few days ago. Finance shops have tried their hand at Kaggle before, but I think they've normally been out of their domain. e.g. Two Sigma recently did a reinforcement learning game competition.

I'd caution the HN crowd not to expect production-level quant models out of this, like I'm seeing some doing in the comments already. Kagglers are excellent machine learning practitioners and the models that come out of many competitions are top-notch stuff, often making their way into research papers. But this is a short competition on limited data in a non-real-world scenario. The winning models will be very interesting educational exercises and probably wonderful recruiting material for Jane Street, but won't be the underpinnings of a new fund.

That said, I can't wait to see what comes out of this one. It ticks all of my competitive boxes :)

You don't beat their model, just those of everyone else in the competition. A very different game.

My point is the winning model of the competition will not be worth anything beyond the prize. It will only be the best amongst other kagglers, almost none of whom are domain experts in finance. It will not stand a chance in "prod".

You don't have to (and certainly won't) beat all of Jane Street. The goal is to beat everyone else on Kaggle. A still difficult but much more accomplishable task.

If I followed your instructions correctly, that's the response payload not the request.

Interestingly enough it seems to do the shift first in Swift

Hm, just seems like really bad copy. They seem to have meant that any new presentations will be public.

Prezi's response in linked thread:

This copy does not reflect our policy - thank you for bringing this to our attention, and we’re currently fixing the mistake. Any private presentations created during your trial or paid subscription will remain private unless you change the privacy settings. (1/2)

Prezi offers free licenses, and when you cancel your trial or subscription, you’ll have the option to downgrade to a free license. After that, any new presentations created from the free license will be public. (2/2)

They're nothing alike at all. Similar to how a low end laptop GPU differs from a top of the line NVIDIA datacenter offering. Google's cloud TPU offering is the strongest ML training hardware that exists, the edge devices simply support the same API.

Lack of competition and honestly I think they just have high costs. Payment processors like Stripe and PayPal don't like dealing with this stuff because the fraud+chargeback rates are so high. This leaves you with the expensive niche providers.

I agree with the content of the article, but feel a summary is sufficient to get the point across. Beginning with a quote:

> If someone claims to have better information than a mainstream financial market on the value of Google stock, or of copper, they ought to either use this knowledge to make a huge amount of money or stop talking about it. However, it is not true if we are discussing prediction markets. Current prediction markets are so bad in so many different ways that it simply is not surprising for people to know better than them, and it often is not possible for people to make money from knowing better.

Ok so why is this?

- Betting markets are illiquid, there is not enough activity for good price discovery.

- Betting markets are capital inefficient. In real life I can short a security without putting up any cash up front, paying interest as my position ages. To make a x:1 return in the betting market I must collateralize (lock away) the entire "1" at the outset. Another example: Say I think Trump will make it to the end of his term as president, right now on PredictIt I can make $1 for every ~$0.80 bet. In real markets I'd sell "no" shares to collect the $0.20 premium for essentially free (if I'm right). On PredictIt I have to stake $0.80 upfront (by buying "yes" shares) for every $0.20 I want to make.

- Betting markets are EXPENSIVE. PredictIt takes a 10% fee on profits and then a 5% fee on all withdrawals. So not only do you have to be right, you have to hit it on a huge margin!

XLS does in fact have a 16 bit row counter :/

Bugtracker: https://phabricator.wikimedia.org/T264370

I agree. I referenced this (perhaps a bit too implicitly) in a sibling comment. Swift's complexity is there but only when you need it.

> Rust aims to be extremely reliable and efficient.

Fair point, Swift will never be as predictable or efficient as Rust (not a negative perse, just different goals). But I disagree on the complexity re: Go. What about Swift gives you the impression of extreme complexity? Go is definitely comparatively simpler but I do not think it is so much so as to put it in a different league.

To me Swift is one of the most interesting languages right now. It's killer for iOS development of course, but Swift for Tensorflow[0] is exciting and could become a refreshing alternative to python for machine learning. Here Swift System + Swift for Linux could make it a compelling alternative to Go or Rust in the future. Looking at Swift System, one interesting thing to note is that they are wrapping the library functions rather than calling into the kernel directly. The source for that is here: https://github.com/apple/swift-system/blob/main/Sources/Syst...

[0] Yes, I'm aware of the staffing issues. No need to beat that dead horse again.

Yes, it's a typed, safe wrapper around a subset of Darwin or Glibc (depending on platform) functions. See: https://github.com/apple/swift-system/blob/main/Sources/Syst...

Yeah, that's more or less the only way to go if you want to use SwiftUI. Personally at that point I just prefer going full UIKit instead.

> I feel we won't see "complete" SwiftUI apps for another 2 years or so, maybe thats the intention?

Unsure about intention but it is definitely the expectation :)

The author’s takeaway here is “If your target platform is iOS 14, you’re now good to go with hobby projects or individual screens in SwiftUI.”

I think that’s generally pretty reasonable but in my experience you won’t even be able to reach the end of your hobby app without hitting some wall where a feature you want is simple in UIKit and not supported at all yet in SwiftUI. SwiftUI is great, the paradigm is refreshing but unfortunately it just isn’t ready yet.

For local networks I use the builtin python http server `python3 -m http.server` (or `python -m SimpleHTTPServer` in 2.x). By default it binds to all so anyone on your network can access it.

> If Microsoft say, added something to your host file, that blocked all ad trackers except from the Microsoft network, everyone would be up in arms.

You'll love this: https://www.bleepingcomputer.com/news/microsoft/windows-10-h...

There are multiple posts in the #1 spot on any given day.

I've been doing research on this exact category myself (after having previously had some modest success on the image editing side.) You're severely underestimating the number of competitors. Try searching the relevant app stores instead of Wikipedia (why would you ever even check here?).

Isn’t exposing ports in your docker-compose services:redis:ports, how you’d do that? (Docker hobbyist here)

It certainly suggests this.

Full bullet: If you use the SRD to find, test, validate, verify, or confirm a vulnerability, you must promptly report it to Apple and, if the bug is in third-party code, to the appropriate third party. If you didn’t use the SRD for any aspect of your work with a vulnerability, Apple strongly encourages (and rewards, through the Apple Security Bounty) that you report the vulnerability, but you are not required to do so.

Does anyone have experience with Hidary's Quantum Computing: An Applied Approach? How is it?

I take it you haven't spent much time on wallstreetbets.

Their whole point on the stock market is this:

> The big issue with all of these is that it’s very difficult to participate in the stock market anonymously. The SEC has all sorts of monitoring in place to catch more common forms of insider trading and fraud and you can guarantee that they would conduct a long, thorough investigation into a hypothetical hack-based market fraud. Unlike Bitcoin transactions, wire transfers and stock purchases can be reversed after the fact, and the exposure and risk go way up when you’re actually working with US dollars.

But this is divorced from reality. In reality you'd be one of thousands of people holding HTZ calls or TSLA puts in your Robinhood account. You could make a huge payday and be indistinguishable from the crowd. With the crypto scam, one person will eventually have to turn this BTC into fiat. With market manipulation, you've made thousands of retail investors indistinguishable from yourself rich. Which one sounds like a better idea?

I recommend ignoring this misinformed WSJ article and reading this: https://www.kalzumeus.com/2019/6/26/how-brokerages-make-mone...

After that come back and see how the WSJ article is just a poorly written hit piece on RH.

Looks like a mistake, should only be in New Features.