191–200 of 304 posts
Wait, do programmers actually think that programming is fiddly? The fiddly part is always the concerns of the real-world domain; what's special to programming is the ability to use abstractions.
Something I've noticed now that we're all communicating online in written form: a lack of people saying "I don't know" It seems that everybody knows everything. Whatever the issue, a little bit of Googling and suddenly you know all there is to know. Even more wondrous, no matter what position you take on anything, some more search engine Kung Fu and you can find a hundred people willing to support you with arguments,…
There's a language that has citations built into statements of fact or belief; you modify statements based on how you came to know them. If someone asks, "Where is Alice?", your answer of "Alice is fishing" would be phrased differently depending on whether he told you he was going fishing, or whether Bob told you he saw her fishing, or whether you saw her fishing. There's nothing like that online, so any given advice…
This is such a good explanation of why I have a hard time working with folks who always approach technical problems with the attitude of "why don't we just do this?" It's the just that gets me. They are too action-oriented to think about the details or ask someone who's done it before, so they start down the path of (to use the author's metaphor) buying cheap lumber and calculating the angles for cutting the stair bo…
"they look terrible" Well then don't do a shitty job? I mean if the budget is there, then yeah, just build me some stairs.
This is such a good explanation of why I have a hard time working with folks who always approach technical problems with the attitude of "why don't we just do this?" It's the just that gets me. They are too action-oriented to think about the details or ask someone who's done it before, so they start down the path of (to use the author's metaphor) buying cheap lumber and calculating the angles for cutting the stair bo…
The flip side of this are the people who can’t accomplish anything novel because they’re unable to simply try. At some point you have to “just” do it.
You can do research and seek out advice, and indeed you should. But for most things, you will not develop a perfect understanding or solution without jumping in and finding the practical gaps in your knowledge. You generally will have gaps and you can spend infinite time trying to fill them. Or you can learn enough to start and understand that you will likely make mistakes and learn along the way.
I’ve never done anything substantial that didn’t involve mistakes. Every major new system I’ve built had mistakes. Every home project I’ve done has imperfections that annoy me. Not once have I completed something substantial and novel and then said I’d do it exactly the same if I were doing it again. That only happens for things I’ve done repeatedly so I learn the optimal way to get it done. Build one deck and you’ll screw up at least three things. Build a dozen and you’ll get it down to a consistent product.
This is of course why trades have apprenticeship programs. You can screw up stairs a dozen times and eventually figure out all the details you missed, or you can work with someone to learn to do it correctly more efficiently. The person you’re learning from can probably fix your errors, too, so the end product is better than you’d have produced.
Something I've noticed now that we're all communicating online in written form: a lack of people saying "I don't know" It seems that everybody knows everything. Whatever the issue, a little bit of Googling and suddenly you know all there is to know. Even more wondrous, no matter what position you take on anything, some more search engine Kung Fu and you can find a hundred people willing to support you with arguments,…
Here's what I think the most chilling takeaway is. People who have no appreciation for details do not know anything . They are dangerous people who will conclude incorrect things with certainty on a regular basis.
So intuitively we have a tendency to make models of everything around us. And now we are encouraged to smash those abstractions to pieces, so we can find better ones. Does that sum it up?
I personally took this to an extreme... I learned to program at quite a young age, and dual-majored in EE and CS specifically so that I could understand the systems that ran my software. I'm happy writing Python, but if I need to dig into the CPython run-time to figure out a bug I'm ok with that. And if I need to run tcpdump to figure out why a TCP stream is stalling, I'm good with that. The abstractions are beautiful and work 99.9% of the time, but if you don't know how the abstractions are built, you're going to have a hard time figuring out what's wrong when they break down.
Earlier quoted context omitted.
If anything it should make you feel the opposite. The brain is an extremely effective optimizer that solves these problems in a way that is opaque to us. We find it very difficult to solve them. An AI is an "unlocked" brain. It can see all the things that it does and then see how it does them. If it learns to ride a bike -and if it's a humanlike AI, it will be able to- it'll be able to see exactly what it does and bu…
> An AI is an "unlocked" brain. It can see all the things that it does and then see how it does them. Except this AI doesn't exist. You're imagining there is such a technology, and this technology will just be able to see all the detail in the world it needs to learn, and presto it does everything better than opaque humans. If you think deep learning is such a technology, then ask yourself to what extent ANNs underst…
Obviously I'm imagining it. Strong AI does not yet exist. It's also obvious that it could exist, because humans do it. I'm only making two logical inferences here:
1: Future superhuman AI will have at least the capabilities of the human brain, because we know those abilities are possible, because we do them.
2. A future superhuman AI will have the ability to examine itself in memory and identify things about itself in a way that far exceeds human introspection: we can barely examine out own emotions, much less the actual neuronal contents of our heads.
> And it's also why I think he paperclip maximizer and gray goo scenarios are silly. Maybe it's theoretically possible to create something that would eat the world, but in order to do so, it would have to overcome every obstacle the world throws at it. [...] If you think deep learning is such a technology, then ask yourself to what extent ANNs understand themselves and you'll see they don't at all.
Well first off, they're quite good at it[1], but more importantly that's weak AI rather than strong AI. Arguing that weak AI is unlikely to be superhuman is plausible, but strong AI is definitely self-improving.
What I think you're saying is that you're skeptical of us being able to create strong AI out of current techniques, which is also reasonable. NN are not gonna evolve into skynet any time soon. But believing them to be categorically impossible requires the human brain to be special in some way- either beyond human comprehension or comprising a supernatural component.
[1]: https://en.wikipedia.org/wiki/Generative_adversarial_network