Mark Zuckerberg's 2016 personal challenge | Hacker News Reader

Mark Zuckerberg's 2016 personal challenge

facebook.com

53 pointsfrostmatthew52 comments

> My personal challenge for 2016 is to build a simple AI to run my home and help me with my work. You can think of it kind of like Jarvis in Iron Man.

> This should be a fun intellectual challenge to code this for myself. I'm looking forward to sharing what I learn over the course of the year.

I would be seriously impressed if he still had the technical chops to build and code this kind of ambitious project on his own.

It's also interesting that he's opening his life up like this and promising to share his (somewhat) individual work with the world, which most CEOs would never do. It has the potential to open him up to critical scrutiny and/or change his public persona. You would never see Musk releasing any updates on personal (less-ambitious) projects because it would detract from his hard-earned and paid-for superhero/wunderkind persona.

Seems like Zuck wants to be perceived more as a fun, quirky, techie rather than the corporate CEO of Facebook. On the plus side, he'll probably get to spend a lot more time with his daughter if he's hacking around the house all day :)

Really? I wouldn't be very surprised at all. I'm sure the guy has serious programming chops, and the goal is sufficiently vague. He's not literally going to build Jarvis in a year, but there are lots of concrete things you can build to automate your home that are Jarvis-like.

He says "simple", so I would say being able to understand and execute "turn off the lights in the second floor bathroom" probably counts. I'm sure he can do something like that in a year (even though there are programmers who can't).

I think a lot of programmers are under the impression that he is this idiot who knows a little PHP and got lucky building a website. But I'm sure he is probably better than the vast majority of working programmers out there, on account of building the initial versions of Facebook (and I say this as a working programmer). People just don't want to believe that because he seems douchey in many ways.

Prior to Facebook, among other things, Zuckerberg made a music player which learned your tastes using machine learning to produce better matches. Pretty sure that even though he hasn't coded in awhile, his coding abilities just need to be reawoken.

Not to mention that he had built several other things prior to starting Facebook.

There's also a lot of libs/services he can use

And if he needs help in understanding or implementing something, he can always ask for some help from someone @ FB...

This. I think most of the work that goes into building a successful large-scale project is making the design choices that minimize required effort.

Everything after that is a matter of typing correctly and adjusting design choices to unexpected constraints.

Having the knowledge of FB's engineering at his disposal will certainly make things easier.

Facebook acquired https://wit.ai/, which allows you to convert sentences (in speech/text) like "turn off the lights in the second floor bathroom" to structured intents. I'm sure he can hook that up to an arduino or Raspberry Pi with a relay in a couple of days.

The guys of wit.ai are now in the Facebook M team, I've read somewhere last month (can't remember the source though)

I bet against you. In the real world in 2016 even if you are a genius you can't do it alone except in very rare cases. You can impress others with your individual skills but you can't push the boundaries.

If he can achieve his goal will mean that a lot of people could do that and there is nothing special in that project. You can argue that he has more access to data than you and me but several people working at FB have the same access as him. His simple AI will be... yes, simple.

> If he can achieve his goal will mean that a lot of people could do that and there is nothing special in that project.

I think that's kind of the point of his personal challenges. He's trying to set an example of being intellectually curious, not show off that he's a genius. Wasn't one of his previous projects for the year to learn Chinese? A difficult challenge, but not a superhuman one. Plenty of other people are capable of it.

There are things to dislike about Zuckerberg, but I don't think these challenges are one of them. They're pretty admirable imo.

I'm never really sure what to make of Zuck, but it'll certainly be interesting to see what he shares of this project. Even by looking in his replies to the comments on this post I'm seeing him sharing more about himself and his opinions than other high-profile CEOs typically would.

Guess this is what billionaires do with their free time. Bill Gates did this to his house many moons ago.

That is exactly what I thought about.

I love how this post perfectly encapsulates the concept of facebook: Someone with a very comfortable life humble bragging about how he's going to use technology to solve first world problems.

It is a cool idea and an interesting project, and publically declaring it will probably help to motivate some people to take on projects of their own. But I do get the feeling that he could be aiming a little higher. With great power comes great responsibility and all that. The alluded to Iron Man was a billionaire who used technology to fight gods and monsters. Zuckerberg is...tinkering with the internet of things? Can't tell someone else how to live their life, but I would have expected him to make public announcements about the Big Problems and leave projects like this for cocktail party conversation.

I've had a sci-fi-villain-hunch about this guy for a long time and this just takes the cake. This is straight out of Ex-Machina, except for the hot robots but who knows, maybe next year...

There's a lot of neat libraries out there and papers to read. There's a lot of hero worship, and a lot of disparaging comments as well between here and the facebook post but a year is a serious amount of time, and with a little dedication I'm sure he can expect to see some decent results out of his pet project.

Similar to that hacker who built a "self driving car" in his garage, which, in the end, although not being a polished finished project did self drive a decent bit, I'm sure we can expect similar if not better results from Zuckerberg here.

Anyone who's claiming differently may not have fooled around with some of the libraries out there (I built something for named entity recognition with Stanford's library in a few days once I found it), and might also be overlooking the fact that we're talking about a person with a serious resume of impressive accomplishments. I'm not sure how his Chinese is and to be fair I was a little less than impressed with the two books a month goal (I do that on the train it seems like anyways, but I don't have a huge company to be CEO of or a family to spend time with), but we are certainly talking about someone who qualified for entry into Harvard (based on merit I presume), and has the drive to sit down and create websites (one of which eventually ballooned into facebook, which he then steered through a public ipo). And let's be fair, getting started is generally the hardest part.

I'd have to be hardpressed to discredit the guy, think ill of his challenge, or feel he couldn't accomplish much regardless of my personal feelings about his attitude and to that I say "Good on you Zuck! Here's to hoping I hear more about this interesting project you've got going."

In the end, lord knows I always enjoy reading about cool projects.

> There's a lot of neat libraries out there and papers to read.

Can you please link those? Because I will move into a house in the middle of the year and I do already have a concept for life form detection based on sonar and 16X4 pixel IR thermal array (MLX90620 / MLX90621).

Hey, so I've mostly done stuff with natural language processing and Stanford's NLP library which I don't think would help you here. But Google's TensorFlow could maybe help you? I'd also look at some of the Python libraries. Just search around and I'm sure you'll find what you need.

Hmm. That's what I'm doing. Ok, so the pressure's on...

Why all the nay-sayers?

He's got a lofty goal. Better yet, it's lofty goal where you can get 70-80% of the way quickly, and if passion serves, you can slog through the next 5-10% throughout the year.

But, why all the negativity?

I, for one, am jealous of the resources (time + money) he has at his disposal to make an actual go at this. But, that jealousy doesn't extend to me hoping he fails.

I feel like I've been seeing more celebrity type stuff on here lately. I guess I would've been more impressed if the comments were about an actual project instead of a personality. If this project is interesting enough to be on the front page, why not talk about that?

I somehow doubt thats the most efficient use of his time but thats his right I guess.

Anyone can have a go at this. Why not pick something that is both challenging and out of reach of "lesser" mortals. Kinda like Bill Gates & his crusade to eradicate Malaria.

I wouldn't be shocked if he goes that route, but he's still the CEO of an engineering-driven company. I think it would inspire confidence that the person at the top cares about technical stuff and can empathize with other employees who write code for a living.

Also I imagine he probably misses just hacking on stuff for fun.

> Also I imagine he probably misses just hacking on stuff for fun.

Exactly... Is that difficult to imagine that he just enjoys the process of learning/building as many of us do? I've also been looking to start a new pet project with simple AI just for the sake of grasping a better understanding of NNs...

> Why not pick something that is both challenging and out of reach of "lesser" mortals. Kinda like Bill Gates & his crusade to eradicate Malaria.

Did you miss Zuckerberg pledging $45 billion to his foundation last month?

>Did you miss Zuckerberg pledging $45 billion to his foundation last month?

I did forget about that yes. Its not quite what I meant though. Its not about the money. There is a bit of a loss of "star" power at work here here. Take the hyperloop example - it sounded a bit outrageous but because it was Musk that proposed it people had a serious look at the numbers. Had I proposed a plan of equal technical merit it wouldn't have gotten the same attention.

Thats why I think people like that should leverage that.

He has plenty of time to do real altruistic stuff. Remember it's only recently that Gates started his foundation; at Zuck's age he was still in full on embrace-extend-extinguish mode.

>He has plenty of time to do real altruistic stuff

Perhaps I wasn't quite clear. I didn't specifically mean altruistic stuff but rather things that can leverage his celebrity status & implicit ability to execute that comes with it. A little bit like if Musk were to announce that he is building a thorium reactor - people will take that seriously even if it is a bit far out.

Building a home automation system on the other hand - there was a guy on reddit a while back that built a pretty good one in his garage.

As I said - he is free to do whatever he wants but it strikes me as a missed opportunity to do something cool.

Well, I guess if he succeeds in doing all of this in a year, it tells us that AI is actually a much simpler discipline than, say, OS development, or compiler design.

It could also suggest there are more tutorials and/or easy to use tools available, making it easier for novices to make a quick impact in the field.

A lot of the groundwork has already been done. It's not as though he'll be piecing it all together from nothing.

I wonder if he means he is going to lead efforts to build an AI such as Jarvis with some help from his friends at FAIR or whether he's actually going to sit down and train his own models (for recognizing his family's faces for example). Quite motivating actually to see someone who is leading a billion dollar company to find the time to study machine learning on the side and apply it.

The penultimate paragraph ends with "But it's a different kind of rewarding to build things yourself, so this year my personal challenge is to do that." and the final paragraph begins "This should be a fun intellectual challenge to code this for myself."

I saw an article somewhere, I don't remember where, about the $10 Echo/Alexa. He could customize that and put one in every room/dungeon in his house(s).

I seriously doubt Zuck is that smart or capable. Intelligence wise Zuck is no Bill Gates. Go look at what Facebook looked like when it was just Zuck working on it. Zuck has not been honing his skills at networking or AI or deep learning for 10 years, he's been leading a company. He's not going to write any of this. Zuck wants to come off like he's the brains behind all of it... and sure he has business sense and billions of dollars to put that business sense to the world. Technically he's very limited.

tl;dr Zuck is not Iron Man.

That's pretty harsh considering you don't really know him at all, and he was at least smart enough to build a company that was good enough to "get lucky" and balloon to the billion dollar one it is now.

It's hardly judging. Considering the amount of effort it takes to master a single language or even a set of algos or putting those algos into a running program. Unless you have been a programmer the last 10 years, you're a manager. A manager should not try to take credit for what his programmers do. That's really the really stupid thing about Stark, he did it all himself. No one does it all him self. Everyone stands on the shoulder of giants.

But I think saying "Mark Zuckerberg had help building facebook" is different from saying "Mark Zuckerberg can't possibly implement the new AI algos on his own", and the latter seems to imply he's some sort of an idiot, which he's not.

Doesn't sound like he is planning to implement new AI algorithms on his own:

"Thanks buddy! I'm still deciding between using the FB environment and AWS. The FB environment gets me access to all of the great stuff the Facebook AI research team has worked on, so I'll probably do that." (https://www.facebook.com/zuck/posts/10102577175875681?commen...)

To be fair, it did look "better" than MySpace. I use quotes because I preferred MySpace's customizability but I understand why uniformity won out.

Take a look at early versions of Windows.

Not talking about how it looked. Talking about what was required in terms of technical skill to produce that. Gates was a pretty good programmer in his teens. Zuck was a LAMP developer, I very much doubt he matured from a LAMP developer to a AI guru in 10 years while running a billion dollar company.

You did say:

> Go look at what Facebook looked like when it was just Zuck working on it

He learned Mandarin Chinese in one year. I don't think there is any question of intellectual capacity.

http://www.theguardian.com/technology/video/2014/oct/23/mark...

http://qz.com/532834/mark-zuckerbergs-20-minute-speech-in-cl...

Does not seem as if he learned it. He's learning it. See this is the sort of shit that I am really against. He speaks a few words and everyone is like see amazing. See he created Facebook all by himself, amazing!

Let's do some more hero worship.

Being able to give a 20 minute speech in a passable - if poorly pronounced - second language seems like a level at which one could say they "knew" that other language.

Clearly no-one ever has learnt a language completely. I think it proves he's no layabout at the very least.

Keep inventing and love innovation. That's the reason why he can be successful. Some say you need to become a billionaire so that you can have free time to invent. However, he could become a billionaire because he used all his time on inventing. This is the fundamental difference between ordinary people and billionaires.

Sounds like a fun challenge, but some of the comments here are a bit too hero-worship-y for me... He's a mere person.

There always seems to be that kind of vibe on any of his posts, they seem to be mostly from people in Asian countries. I guess everyone needs a role model, and when people see the chance to interact with theirs they really reach out for it.

I think what's cool is that Zuck has actually replied to a good chunk of top-level comments on his post. Most bigwigs on social-media don't really engage with their "fans".

To be fair, replying on the platform he owns is a fairly significant value-add process.

A lot of what he mentioned can be done without writing a line of code. He is basically talking about a large scale home automation system with a voice interface to it

127.0.0.1 facebook.com

> My personal challenge for 2016 is to build a simple AI to run my home and help me with my work. You can think of it kind of like Jarvis in Iron Man.

I wouldn't throw those names around too much. The Marvel movie empire might not be worth as much as facebook, but it may be close. Those trademarked names are valuable assets. There are firms out there ready to pounce on any unauthorized use. Not the acceptable wording "like Jarvis in Iron Man." That's ok as it uses the terms descriptively. His legal department would have rejected "I'm building Jarvis".

https://trademarks.justia.com/862/94/jarvis-86294162.html

"Computer application software that may be downloaded via global computer networks and electronic communication networks for use in connection with mobile computers, mobile phones, and tablet computers, namely, software for use as a voice controlled personal digital assistant." .. not fictional software. The Jarvis mark is protected in the real world just like Siri.

In the international descript Jarvis is also a "fire-extinguishing apparatus." A coincidence, but it made me laugh.

Reply on news.ycombinator.com