HN user
+1 For real world business problems that I most frequently encounter doing consulting, it's hard to beat Random Forests and/or Gradient Boosting. Truth be told, most business problems I encounter turn out to be largely helped by good old linear models.
I'd settle for figuring out how to send an email to a previous point in time. I'd like to email myself in 2009. Something about buying a bunch of ASIC's and mining something for a long-term hold.
With a little luck, in about two months we'll finally get to watch Shazam again.
Mark is also the that recently tweeted a picture indicating that he as working through an "into to python for machine learning" book. He may have been successful in tech 25 years ago, but I don't place a premium on his opinions of where things are going. Advising someone who does not intend to go into the academic field to major in any of those is pretty poor career advice. Either you want to work in academia or you just love the subjects and will figure out money through other means... only plausible reasons to pursue those majors. Of course you don't have to worry about automation impact on degrees that have little to no correlation to job-placement anyway.
This guy is just trying to tout his own product. That aside, Kaggle can be quite useful for learning... at least it used to be. Kaggle got me into machine learning 2-3 years ago and now it is about 50% of my job in consulting. Problem is, Kaggle is running too many deep learning, image based comps and they are lucky to get a couple hundred competitors each. The wider user base engages much more heavily in transaction based competitions, and these are much closer to the real world problems you are likely to encounter as a professional.
I think this is because at it's core, the math behind deep learning is pretty simple and there's not much of it. Linear algebra, some simple activation functions and gradient descent. Implementing a net from scratch in python is pretty concise. The simplicity-to-effectiveness ratio is what I find so interesting about deep learning.
Short answer is that their employment contracts probably don't allow them to sue. Almost all employment contracts contain binding arbitration agreements now.
My "Fake News" comment aside, I think what y'all are doing could be transformational for many reasons. Imagine a scenario where a person loses a loved one, and similar technology is able to allow them to "have conversations" with the deceased as a form of healing and closure. Not to mention, this could add a personal touch to assistant bots that will make them a pleasure to use.
The top two are representative of what it sounds like when doing true text to speech. The middle five are just resynthesis of a clip saying the exact same thing.
Gotcha, now I understand.
Based upon what little is posted there, I thought they were taking the original recording, then training the model on that recording against the text of the recording... reproducing the recording. I would think next step is to sample enough audio and text to be able to produce new outputs entirely. It should in theory even be able to learn when/where/how to use inflection.
Interesting. They are not TTS like we are accustomed to, they are replicating a specific persons voice with TTS. Listen to the ground-truth recordings at the bottom and then the synthesized versions above. "Fake News" is about to get a lot more compelling when you can make anyone say anything as long as you have some previous recordings of their voice.
As already mentioned, they do have Windows VM's but there are some caveats that indicate it's not fully baked yet. 1.) They require that each VM MUST have a public IP address so that Windows can talk to an activation server every 30 days. 2.) You cannot yet bring your own license.
czep, even though I don't agree with much of what you wrote (in the article), I do appreciate you taking the time to so thoroughly expound on the ideas. Paul Graham most likely fell victim to projection of a Libertarian ideal when he viewed "best programmers" from his perspective.
I'm also frustrated to see a powerful man perpetuate self-serving ideologies without acknowledging the influences of power and luck.
I agree with this above statement maybe or maybe not for the reasons you do.
The "Power Game", or even the lack of willingness or know-how to play it is the reason a lot of programmers think themselves superior to the sales guys who peddle the product of their labor.
The "Power Game" is also the same reason the good sales guys feel that despite being so technically smart, programmers can be damned idiotic fools.
As an introverted "programmer-type" myself, life would be way easier for me if there was no Power Game. But the Power Game is as human as eating, drinking and pissing.
Hell, I get irritated daily that I even have to eat, drink or piss. It feels like a waste of time when I am in the zone with something. Same thing with the Power Game.
Luck... hugely important. Heck, we've all played RPG's and know to fill up that skill bucket ASAP.
Maybe what would really be helpful to programmers is something that can stir inspiration like Paul Graham, but that covers something akin to "The 48 Laws of Power" for the introverted modern day employees.
And I forgot to add above: I am sorry for the loss of your friend. When a friend dies in a freak accident, there is at least the comfort that you probably could have done nothing. When someone commits suicide your mind turns endlessly looking for sign and signals you missed and you always wonder. I hope you find peace.
You specifically asked how you could use your technology skills to help people.
First though, there are two kinds (at least) of depression. One is the normal kind where you feel blue when things go wrong in life. This is normal and it should fade. There is another kind related to an incorrectly functioning brain. For this medication is needed to help correct that problem.
When a malfunctioning brain combines with things not going your way in life... you get a high probability of a downward spiral and suicide. It's a terrible self-feeding cycle.
Back to technology: Having a sense of meaning and purpose is quite possibly the most important component to a healthy human experience. When you are completely down in the dumps, just going down to a local soup kitchen and offering to volunteer for a day can recharge you and spark understanding of the human condition.
What you can do with technology is endless. For instance, I have never encountered an app or site where I enter a zip code and see everywhere I can go and help other people that day, right then and there. If such an app or site could list close by "AA" style depression sufferers meetings I could drop into anonymously on my way to they metaphorical "soup kitchen"; that would be even better.
You won't ever understand depression through tech, but understanding it through the human side will allow to you to envision how to use your tech skills to help.
I see a future where a developer is paid the salary of a retail cashier and it scares me.
The only way that will happen is if Point-of-Sale systems become so complex that a developer has to run them, or development becomes as easy as punching buttons on a Point-of-Sale system.
"Why am I getting reprimanded?" "Because you ran a unit test..." "What is wrong with that?" "It's Brian's job to press that key, not yours. You violated Union rules"
Yup, I see 8-10 hours of real use since 10.12.3
"Trump's huge unpopularity right now shows..." After watching what happened with the election, I don't think it's safe to say Trump is unpopular. Sure the mainstream media and loud social media voices say so, but that looming "silent majority" are still silent. Assessing his popularity among those willing to speak only tells us how popular he is among that group.
Ask yourself this: Is there a market for database developers that don't have a PH.D in Relational Algebra? "AI" will be democratized very quickly just like data (SQL, Excel, etc.) has and solving business problems with technology will be more important than ivory tower tinkering. That being said, if you're interested in Deep Learning... do yourself the favor of truly understanding what is happening with backpropogation and eventually learn to code networks by hand, from memory; simply to aid your own intuition.