Profile cover photo
Profile photo
Kenton Varda
The Cap'n Proto Guy (formerly The Protobuf Guy)
The Cap'n Proto Guy (formerly The Protobuf Guy)

Kenton's posts

Post has attachment
Not bad... not bad.

(with +Jade Q Wang​)
2 Photos - View album

Ugh, why does Google Maps still report incorrect Caltrain schedules on holidays? This has messed up my travel so many times.

Ironically I only realized that today was a holiday when I noticed Google Maps was showing me a Saturday schedule for BART. But it combined that with a weekday Caltrain -- which is bogus and would have made me late for my flight.

Dear lazyplus,

With an Android phone, is it possible to use a strong password for encryption, but require it only at boot time, and use pattern unlock at the lock screen? I would like to increase the strength of my device's encryption (especially while travelling out of the country), but typing a long password every time I want to unlock it is pretty rough. My current plan is to switch to a long password just while going through customs...

Googling suggests the answer to this is "no, your regular unlock mechanism must be your full encryption key", but that seems so obviously broken...

Played *The Turing Test* yesterday at +Jake Weisz's recommendation. Pretty good game! Same feel as Portal and The Talos Principle. Only 5-ish hours long.

Like The Talos Principle, in parallel with the puzzles there is a bunch of philosophical dialog. In The Talos Principle I found this dialog annoyingly cliche. In The Turing Test, though, the dialog turned out to be really interesting, particularly as the plot progressed. I'm not going to go into any detail because I don't want to spoil it -- you should play it.

Tangentially: The game makes reference to John Searle's infamous "Chinese Room Experiment", which always makes me angry. I'll talk about that since it isn't a spoiler, obviously.

The Chinese Room Experiment is a thought experiment that goes like this (in my words, not the original):

Imagine you don't know Chinese, but you are given a set of instructions which, if followed, allow you to take a query written on paper in Chinese and produce a response written in Chinese, without ever knowing what the statement actually says. The instructions are low-level thing like: "If the color of the paper at coordinates 5, 9 is black, write the number 2 into your scratch paper at cell 5." or "If your scratch paper cell 52549 has the value 1, color your output paper black at position 6, 8.". Now imagine you are placed in a room with these instructions and all the scratch paper and time you need, and you are passed notes under the door written in Chinese, to which you produce responses written in Chinese, indistinguishable (other than time taken) from how an actual Chinese-understanding human might respond.

Searle: This simulates what a computer does. But nothing involved in producing the response actually "knows" Chinese. Hence a computer cannot truly "know" Chinese and, by similar argument, cannot be conscious.

Me (and many others): The system as a whole (as executed by you plus the rest of the room) knows Chinese even if you don't.

Searle: OK, so imagine you memorize the instruction book, and now you do the whole thing in your head. You still don't know Chinese!

Me: Memorizing the instruction book and running it in your mind does not change anything. It's still that process, which you are hosting -- but which isn't you -- that knows Chinese. Yes, two consciousnesses can occupy the same physical space while still being separate, because consciousness is not physical, it is a process. What you have here is like a virtual machine. If I run a Windows VM on a Linux box, they are occupying the same physical space, but the Windows VM for all intents and purposes is a separate machine, and the Linux host still doesn't understand how to run Windows programs even though Windows programs are running within it. It's the same thing here. Honestly this is really not that complicated and I'm surprised that a professional philosopher with an interest in computer science would have trouble grasping this.

Searle: I have catalogued hundreds of counter-arguments to my argument but none of them are correct. I can't be bothered to actually counter them individually. I'm world-famous philosopher John Searle and I say they are wrong. So clearly I am right.

Me: WTF?

Just played The Last Guardian.

This is not a good game.

I had high hopes since it's from Team Ico, who made Shadow of the Colossus, which is one of my favorite games of all time.

Unfortunately, The Last Guardian did not take after SotC. It mostly felt like Ico, except much more annoying.

Apparently they spent some 10 years on this game. But apparently those 10 years mostly went into making your giant-feathery-dog companion behave "like a real animal". And indeed, the companion's behavior is incredibly convincingly animal-like. It would be incredibly impressive except for one problem: the game is not fun.

The animal behavior does not make the game fun. In fact, it does the opposite: it makes the game incredibly annoying. Because if your animal doesn't do what you want it to do -- for instance, jump to a nearby ledge -- there's two possible reasons:

1) It can't jump to that ledge. There's something different you need to do next.
2) It's being moody, and if you keep trying for half an hour eventually it will indeed jump to the ledge, which indeed is the thing you need to do next.

As a result of this awful mechanic, a large percentage of the game is spent trying to coax your companion into making jumps it can't make, or spent searching in vain for what to do next after falsely concluding that you can't jump to that ledge.

The other half the game is spent missing jumps and falling into pits because the game controls are basically QWOP. Apparently, team Ico spent zero of those ten years fixing their egregiously play control issues -- indeed, many parts of the game seem explicitly designed to highlight the crappy control. (The controls were bad in SotC, too, but it felt "right" when you were trying to climb up the side of a giant golem who is struggling to throw you off. SotC was supposed to be stressful. But for a "find your way through the inanimate obstacle course" game, it's totally wrong.)

I was hoping for a puzzle-platformer, but The Last Guardian does not have "puzzles". None of the areas required any kind of mental capacity to figure out. When it was unclear what to do next, 100% of the time it was because there was a thing that I could do which I didn't know about, and so the challenge was to wander around until I accidentally tripped over the answer. These are not puzzles. In a good puzzle-platformer, you have a set of tools and you have a problem, and you can figure out in your head exactly how to use the tools to solve the problem and know you have the right answer before you actually do it. This is not that.

I could have lived with a dexterity-platformer instead, but that requires controls that aren't shit.

The Last Guardian is a QWOP-platformer.

I guess the story was good, though.

Applied for 2-week visitor visa to China.

Received 10-year multi-entry visa.


Post has shared content
Sandstorm for Work is now free of charge and Blackrock -- the scale-out tech behind Oasis -- is now open source.

The project is transitioning to being a classic community-driven open source project, and -- for now, at least -- shedding our business aims, as they haven't succeeded. We are, however, committed to keep working on it, and we'd love your help -- check out the new contributing guide outlining high-value projects.

Post has attachment
If you were wondering why the tap water in the bay area smells like spent fireworks lately...

I've often found that talking to Trump supporters is like talking to someone living in a parallel universe, where truth is not based on facts or logic but comes from the gut, any sound rational argument is simply dismissed out of hand, and mature "western values" which seemed central to the American identity a few years ago are now simply ignored or even mocked.

Today I'm feeling the same way talking to liberals about "punching nazis". Apparently, a large contingent of people on "my side" not only think that straight up punching someone for holding extreme racist views is OK, but that if I think maybe this isn't OK, I'm guilty of "normalizing" Nazism as an acceptable political viewpoint.

I give up. We're all screwed. I'm going to attempt to ignore politics and instead focus on coding and playing video games.

Holy crap, Clang can do code completion inside a macro body, when the context is dependent on a macro parameter...
Wait while more posts are being loaded