sapphireblue

6 ms·

1.
by sapphireblue 9y ago
>But neither the NN nor the researcher will be able to give you a rule like: "F=(G m_1 m_2)/(r^2)" to explain the underlying reasons for the object-trajectory dynamics. It is not impossible. There is a combination of a spe
2.
by sapphireblue 9y ago
>Everyone talks about extending life and eventually becoming immortal, yet no one asks what happens when we are immortal. I see the opposite picture when these discussions pop up: everybody tries to find a downside in increased lifespan,
3.
by sapphireblue 10y ago
This is very important. It seems that the automation of labor is underexplored due to political and economical inertia (I know that industrial robots do somewhat improve, but they still are very costly, very proprietary and require experts
4.
by sapphireblue 10y ago
There was a couple of interesting threads on HN about ultrasound machines recently: https://news.ycombinator.com/item?id=13230741 https://news.ycombinator.com/item?id=13241295 I stumbled upon this startup
5.
Smartphone-based Ultrasound device Clarius passes FDA tests(clarius.me)
4 points by sapphireblue 10y ago|1 comments
6.
by sapphireblue 10y ago
I agree with the first sentence, but I'd like to note that there are practical (though weak) approximations of AIXI that preserve some of its properties, and while not turing-complete, prove to be more performant when compared to other
7.
by sapphireblue 10y ago
Actually there are at least two decades-old branches of computer science/mathematics that have formulated precise definitions of AI, and proved many theoretical results that gave way to lots of practical applications. These branches of
8.
by sapphireblue 10y ago
Meta- reinforcement learning could prove to be such breakthrough, see [1],[2]. Also next generation ASIC accelerators (Google's TPU, Nervana) can give 10x increase in NN performance over a GPU manufactured on the same process, with ano
9.
by sapphireblue 10y ago
Looks like the "UNREAL" ( https://arxiv.org/abs/1611.05397 ), "Learning to reinforcement learn" ( https://arxiv.org/abs/1611.05763 ) and "RL^2" ( https://arxiv.
10.
by sapphireblue 10y ago
Is he a CEO though? Wikipedia and other press articles say that Hassabis is the CEO: https://en.wikipedia.org/wiki/DeepMind and Suleyman is Chief Product Officer, the head of applied AI at DeepMind.
11.
by sapphireblue 10y ago
You have a lot of solid points, but note that Linux is currently being used by SpaceX in an even more safety-critical aerospace setting. Also note that interpreters have their place in safety-critical aerospace as well: some satellites run
12.
by sapphireblue 10y ago
Note that Facebook also has a formidable AI research group called FAIR and they are pursuing goals close to DeepMind's, while openly publishing their results and tools. There is a lot of social media unicorns that don't contribute
13.
by sapphireblue 10y ago
A general purpose reinforcement learning (RL) agent is a machine that can be taught to perform any task from a very wide range of tasks via sparse rewards given by a human or software trainer. The agent can, like any software, be snapshotte
14.
by sapphireblue 10y ago
DeepMind looks like a hilariously wrong project to criticize because it is a true moonshot, something very different from the majority of other SV projects. If hiring hundreds of PhDs to create a general purpose learning agent, all while pu
15.
by sapphireblue 10y ago
The unpleasant truth is that this dystopia isn't coming out of nowhere, it is being built with our own hands. Our everyday desires, decisions, actions and inaction decide what will be built (more apps) and what won't (nextgen cure
16.
by sapphireblue 10y ago
Turing test is relatively easy to pass if you creatively design your program to roleplay an irrational character. See this program https://en.wikipedia.org/wiki/Eugene_Goostman Obviously such programs, while being work
17.
by sapphireblue 10y ago
I very much agree with this point of view. It seems to me that dry metaphysical debates (i.e. questioning the nature of these conversation models) are obscuring interesting practical questions that could be asked and tested empirically. Giv
18.
by sapphireblue 10y ago
I liked Anathem, but it is more of a philosophical work. Personally I think The Diamond Age is Stephenson's hard sci-fi masterpiece.
19.
by sapphireblue 10y ago
Hannu Rajaniemi, Ted Chiang and Neal Stephenson come to mind as hard sci-fi authors worthy of comparison to Greg Egan.
20.
by sapphireblue 10y ago
Well said. Political and intellectual elites tend to attach negative connotations to the word "utopia". Perhaps they feel that the very concept of utopia attracts undesirable attention to the bleakness of our current status quo as
21.
by sapphireblue 10y ago
I disagree. There exist examples of system displaying very non-trivial, open-ended behavior, and yet they are still not agi. A good example of this is https://arxiv.org/abs/1603.01417 . When trained it is able to answe
22.
by sapphireblue 10y ago
Even if we relax hardware requirements, e.g. assume a Xeon+Titan X hardware, it is still a question if a viable conversational agent that uses deep learning to generate conversations and adapt to surroundings can be developed around it. May
23.
by sapphireblue 10y ago
Am I the only one who thinks that dystopian scifi got boring a decade or two ago, while utopian scifi is an almost entirely neglected genre? I prefer a techno-optimistic point of view shown here http://foundersfund.com/anato
24.
by sapphireblue 10y ago
>Promising to deliver an AI that people could see as a friend is absolutely insane though. I don't see people being friends with something that couldn't complete the Turing Test This is an interesting case. Turns out, given a c
25.
by sapphireblue 10y ago
... but without new Hardware there will be no new future. Just more of the same. We really need to embrace hardware more.
26.
by sapphireblue 10y ago
Maybe it is possible to design a robot-friendly rug or a rug-friendly robot locomotion mechanics!
27.
by sapphireblue 10y ago
I think the problems you described could be solved if we made tiny adjustments to our lifestyles and environments to make it much easier for automated systems to navigate and interact with it. For example make a choice of not having rugs, h
28.
by sapphireblue 10y ago
>The skills for building a robot are quite different from the skills for building "classical" software. I'd object that on the contrary, arduino and http://www.espruino.com/ require mostly the same skills
29.
by sapphireblue 10y ago
I also wonder how we manage to stay busy given that most home tasks (dishwashing, laundry, cleaning (roombas) and even cooking - see multicooker devices) are already automated, with "doing laundry" meaning merely loading and unloa
30.
by sapphireblue 10y ago
Building a shack isn't that hard, see https://www.youtube.com/watch?v=P73REgj-3UE Try building it in the forest though - it is most likely illegal due to some rentier (or government) already owning it. I think the gran

More ›