splonk

19 ms·

151.
by splonk 6y ago
From having spoken to people who worked on object clustering for Froogle in the past, I think the base issue is that it's surprisingly hard to extract all the metadata accurately from descriptions. Even if you provide an file format a
152.
by splonk 6y ago
Gmail was internal. Maps was mostly built by a couple acquisitions, but it wasn't anything like Youtube, where the acquired companies were big established players in the space.
153.
by splonk 6y ago
Both of them also have GDS integrations. I don't know exactly how much but I'd assume it's the bulk of their traffic. As a general rule the large majority of anything you don't book directly with the airline/hotel
154.
by splonk 6y ago
The Havana stream is pretty cool if you're into classic cars. Quite a few just on the streets, since the embargo basically stopped US car imports after 1960.
155.
by splonk 6y ago
I agree with you in principle, but having worked on both sides of this, I think there's very little chance they get blocked at their current traffic levels. I do think that if they ever get traction they'll have a lot of problems
156.
by splonk 6y ago
Just try to make sure you've found the real website and not one of Grubhub's autogenerated fake sites. They all look pretty similar so you'll be able to identify them after you've seen a few, but it's easy to mista
157.
by splonk 6y ago
I'm aware of at least one community that shares copyrighted videos on Google Drive, and I'm sure there's more. There's at least two ways videos get taken down. 1. If there's too much burst traffic (like when a new
158.
by splonk 6y ago
> Furthermore, when I installed and ran a VNC server, I didn't detect any difference in site behavior - so why is it looking for it? Not an eBay employee, but used to work in fraud detection. Two very obvious related guesses from m
159.
by splonk 6y ago
At the end of the day, there's a lot more legitimate transactions than fraudulent ones. It's very difficult to restrict a particular type of purchase en masse without having a huge false positive rate. That's exactly why co
160.
by splonk 6y ago
Lots of ways to monetize a working stolen card, although as GP says, the people stealing the credit card information are generally selling those cards to other people who'll actually try to turn them into cash. Gift cards and any sort
161.
by splonk 6y ago
I used to work on fraud detection on a product with transactions totaling billions of dollars a year, and for a period of time we could have stopped something like 90% of our fraud attempts (with like a 99% accuracy rate) by simply blacklis
162.
by splonk 7y ago
Yeah, I'm not going to say United is some shining beacon of a company, but all I'm seeing here is that somebody representing a company that "book[s] a lot of travel" doesn't seem to understand how fare classes work.
163.
by splonk 7y ago
Google ran one internally for some time maybe 10 years ago. (Maybe still does, for all I know.) IIRC it was for points, maybe transferable for massage credits or something, so no huge economic incentive. I have no idea if it was in any way
164.
by splonk 7y ago
Companies with more than 50 employees are required to offer some sort of commuter benefit out of a predefined list. In practice I've seen the pretax deduction the most. https://mtc.ca.gov/whats-happening/news&#x2F
165.
by splonk 7y ago
> Hipmunk's parent company got a piece of my $60k in travel last year (most of that was work, maybe $2500 was personal), because that’s where your percentages can matter. I think for corporate travel management, the fee structure is
166.
by splonk 7y ago
It does to certain business travel platforms (search for "swabiz"), and I believe they're expanding that exposure through their integration with Amadeus. Whether they plan it make it available on consumer sites eventually, I
167.
by splonk 7y ago
Assuming skiplagged is still doing hidden city ticketing, it's important to recognize the risks involved here (primarily that overuse of it may cause airlines to reject your bookings and/or shut down your frequent flier accounts).
168.
by splonk 7y ago
Concur's a business travel platform. These differ from consumer platforms in a few ways: 1. Private negotiated rates - medium-large companies have bulk rate discounts individually negotiated with various chains that can be integrated
169.
by splonk 7y ago
1-2% is more of an upper bound. The numbers I've seen have been more like 10's of cents per conversion. And for whatever reasons, flight availability providers are not in the business of providing high throughput for clients, so
170.
by splonk 7y ago
This is also not uncommon. I've seen similar at Oaks, Bay 101, Wynn, Casino Arizona, Commerce, and probably others as well. I tend to believe that it shouldn't be allowed, if only for the pace of the game, but it's not uniq
171.
by splonk 7y ago
Usually only two decks are used at a table, one for the current hand, and one in the shuffle machine. That's only 104 unique tags. Decks tend to be changed out only every few hours at most, because it delays the game (new decks are v
172.
by splonk 7y ago
Almost every poker room allows phones at the table. Turns out poker's a really boring game when you're at a table for hours and hours. Many of them technically don't allow use of them during a hand, but those rules tend to
173.
by splonk 7y ago
>> In San Francisco, if you sit down at a restaurant, take the written prices on the menu and multiply them by 1.5 to get what you'll actually be paying. >Is this anything other than taxes and gratuity (which, technically, you
174.
by splonk 7y ago
Chowhound basically got destroyed a few years ago when it got acquired and redesigned into an unusable mess, and now traffic has cratered. Some of the regulars split off into new sites - HungryOnion (mostly SF Bay) and FoodTalkCentral (mos
175.
by splonk 7y ago
TA is pretty close to last on my list for restaurant search. The top of every ranking in any even mildly touristy city is going to be spammed to hell and back. It's basically a list of "where do tourists go who can be preferenti
176.
by splonk 7y ago
Cheers, thanks. One of the reasons I asked about 1H/5AI vs 5H/1AI is that historically the new bots for a given form of poker have played a bit wider than the accepted wisdom of the time, so I was curious if there were relatively
177.
by splonk 7y ago
The other issue is that increasing the number of branches also decreases the number of hands that go into each bucket, to the point where it might not be effective any more without being able to randomize the branch choice for specific thre
178.
by splonk 7y ago
I haven't finished digging through the paper and the supplement yet, but I'm curious about how many hands were multiway to the flop (and whether the percentages differ significantly between 1H/5AI and 5H/1AI). I'd
179.
by splonk 7y ago
They're credible enough. I'd like the sample sizes to be bigger as well but they're enough to verify that even if the bot got lucky over the sample size, it's close enough that it doesn't really matter. Add a bit
180.
by splonk 7y ago
It's basically been true along every step of the the poker bot evolution (HU limit, HU NL, and 6-max NL) that the bots donk a lot more than the humans. 10 years ago you could find pros arguing that donking in any situation is always w

More ›