levesque

52 ms·

121.
by levesque 8y ago
I mean, the ideas are there. The scope of the project is probably too big to reproduce for now, but eventually it will be accessible to your average spammer / scammer. We will get there. We won't be able to get these tools loc
122.
by levesque 8y ago
The presentation is very interesting. That's always what amazes my with those GAN outputs. These people do not actually exist. Obviously, there are some funky examples. Nothing wrong with mine at first, although his buddy should probab
123.
by levesque 8y ago
Seems very nice, but it also appears like they have few stock images per keyword/topic and they would be overused after a relatively short time. Am I not getting something? Are they supposed to take out an image once the logo is sold?
124.
by levesque 8y ago
I only described my personal experience, I'm not making assumptions of evilness on Google's side, just pointing out that it feels a little restrictive. As somebody else pointed out in this thread, the information is still stored o
125.
by levesque 8y ago
> "Seriously considering buying some illegal drugs to try to turn them back into cold medicine." Now that sounds like a project.
126.
by levesque 8y ago
It's pretty useful when you use your phone a lot for navigation / GPS. Saves you a bunch of time, but true it does have the cost of letting Google know where you live (although there are so many other ways for them to get that inf
127.
by levesque 8y ago
Yeah, I agree it was a bit strong, I edited to "downgraded".
128.
by levesque 8y ago
As of last month it still wasn't the case for me. In the end, I turned back search history on, I decided to pay the 'price' for the features.
129.
by levesque 8y ago
iOS is becoming more attractive by the day.
130.
by levesque 8y ago
I tried using Google Maps without search history turned on. It's basically a downgraded version. They tied saving locations with the search history, so it won't let you search for friends' contacts, set a home address, etc. I
131.
by levesque 8y ago
Sometimes it helps to read the article ;)
132.
by levesque 8y ago
Bias is one of the new hot topics in ML. People act surprised that a face recognition software trained on a database of white people is unable to properly classify pictures of people from all around the world. The algorithm is only learning
133.
by levesque 8y ago
Perhaps a keyboard shortcut to replace mouse-button-down, that definitely sounds do-able.
134.
by levesque 8y ago
Also, consider that several of these services still exist in one form or another.
135.
by levesque 8y ago
There's probably a project to make here. Take those pixel2pixel models that are able to generate images from drawings, generalize them to 3D, say a pixel2vertex. Would be pretty neat. The key would be to find a way to constrain the out
136.
by levesque 8y ago
Pretty underwhelming. From the very limited experiment demonstrated, the LSTM seems to be playing catch up rather than actually predicting trends.
137.
by levesque 8y ago
Happy to see more work based on prediction. I've been of the opinion that predictive rewards should largely bypass the hand-tuned rewards we have been using for reinforcement learning so far, or at least speed up learning by providing
138.
by levesque 8y ago
Fair enough, probably a bad example.
139.
by levesque 8y ago
Even though probably everyone in your town eats sushi, a variant that is most likely wildly different from what the Japanese call sushi, people don't go around calling it a Canadian or American dish. Same goes here, I'd say. At it
140.
by levesque 8y ago
College degree, working a 9$ / hr job. Not saying this is impossible, but probably not the case of most college degree holders.
141.
by levesque 8y ago
Yeah SigOpt clearly has taken the space left by Whetlab. They must not be too pleased about this announcement, I mean generally the worst thing that can happen to your startup is a tech giant launching in your space. Then again the problem
142.
by levesque 8y ago
Yeah bicycle routes don't seem to work. It's sad I was really curious to see how far an hour route could take me!
143.
by levesque 8y ago
I'd rather just watch content with proper subtitles in the language I'm trying to learn. Say I'm learning German, I could either watch some original movie in German with English subtitles, or watch an English movie with Germa
144.
by levesque 8y ago
That's not gonna happen, unless they dislike being sponsored/funded.
145.
by levesque 8y ago
You save time by getting it right now, so that's worth something.
146.
by levesque 8y ago
From my experience with UPS, they probably burned the shirts.
147.
by levesque 8y ago
The data generation is indeed one key aspect of it. To train a reinforcement learning model such as this one, you do need an insane amount of data (they wrote somewhere that the model played the equivalent of 180 years of Dota per day). Ove
148.
by levesque 8y ago
Agressive tipping is also of the utmost importance
149.
by levesque 8y ago
I like UIs that feel premium. Needless to say, Xubuntu and i3 do not make the cut. Left with Gnome 3 (ugh) or KDE (options galore). All other DEs feel clunky (graphical glitches, sub-par design) and are also usually buggy (which is normal f
150.
by levesque 8y ago
Wow, makes the whole thing much more pleasant to use. Pretty bad that I only discover this now, after a year and a half.

More ›