Post-rsses on Musings, Rants, & Other Matters

Maybe I Needed a PolySci Degree

rob@Loci.net (Rob Arnold) — Thu, 17 Nov 2016 12:48:13 -0800

alt-text:

Certified skydiving instructors know way more about safely falling from planes than I do, and are way more likely to die that way.

Felt I needed to post something just to remember how. This felt appropriate on so many levels.

I have several more mostly completed random musings but just haven’t taken the time from other things to think and edit them into shape enough to even post in this personal and casual form.

Quora AI Q&A with Yann LeCun

rob@Loci.net (Rob Arnold) — Wed, 10 Aug 2016 00:00:00 +0000

I read a really interesting Q&A on Quora the other day, and wanted to call out and comment on a couple points.

First though, I should reiterate my disclaimer that, basically, I know nothing. Second, what I think is that AI is going to be a world changing, hugely important endeavor. I think it has the power to positively change life of all types in ways we’re not even able to conceive of yet, and a multitude of ways in which we can already conceive. I also think it has the potential to destroy… almost everything. How we manage (or fail to manage) this transition will be one of the most important tests we ever face.

I’m excited by the potential of AI (even if I dislike that term) and want it to happen in a big and positive way. Keep all that in mind when reading my comments.

We basically have one long-term goal: understand intelligence and build intelligent machines. That’s not merely a technological challenge, it’s a scientific question. What is intelligence and how can we reproduce it in machines? Together with “what’s the universe made of” and “what’s life all about”, “what is intelligence” is possibly one of the primary scientific questions of our times. Ultimately, it may help us not just build intelligent machines, but also understand the human mind and how the brain works.

Yann LeCun Quora

The above is at the root of why I felt I had to name my category “AI & Mind” not just AI, and related to “The problem with calling it AI” post

Q: Is there something that Deep Learning will never be able to learn?

The answer in total is great and worth reading in full. The quoted bit below as the conclusion though doesn’t really make sense to me.

There are problems that are inherently difficult for any computing device. This is why even if we build machines with super-human intelligence, they will have limited abilities to outsmart us in the real world. They may best us at chess and go, but if we flip a coin, they will be as bad as we are at predicting head or tail.

Yann LeCun Quora

The coin flip example – sure, maybe. But the broader statement about limited abilities to outsmart us in the real world… I just don’t see it. This is right after saying, “It seem humbling to us, humans, that our brains are not general learning machines, but it’s true. Our brains are incredibly specialized, despite their apparent adaptability.”

I want to read more about the “no free lunch theorems” mentioned, and I get the overall point, but I suspect it is far too strong a statement to say they will have limited abilities to outsmart us in the real world. Limitations inherent in the nature of their ( or more accurately, any given) intelligence, sure – just like the limitations inherent in the nature of our intelligence. But I don’t see anything necessarily more limiting, and can conceive of the possibility that through iteration and scalable resources, increased speed of technology, etc., that the limitations could have a number of mitigations and alternatives that would not necessarily be available to human intelligences, at least while locked to the original substrate / construction.

Q: What are the likely AI advancements in the next 5 to 10 years?

Good answer in full. Just pulling out one comment though, as it will relate to the rest of my post:

A big challenge is to devise unsupervised/predictive learning methods that would allow very large-scale neural nets to “learn how the world works” by watching videos, reading textbooks, etc, without requiring explicit human-annotated data.

Yann LeCun Quora

I agree, and note the “senses” involved along with the source material orient the intelligence to be at least connected to the same things and ways of perceiving the things as us. So probably grows an intelligence with at least some common ground to human intelligence… Unsupervised would not mean without overarching direction of course, and without context. The mind boggles at what an intelligence grown purely from the current way people interact online, for example. In fact, witness the results from Tay, the Microsoft chat bot loosed on Twitter (and read their lessons learned).

Regardless, remember the unsupervised point when evaluating his answer about threats.

Q: What is a plausible path (if any) towards AI becoming a threat to humanity?

I don’t think at AI will become an existential threat to humanity.

Yann LeCun Quora

I’m happy to hear the opening statement “I don’t think at AI will become an existential threat to humanity.” from someone like Mr. LeCun, with his experience and background. That said, it’s a large stakes issue and I’d like to comment on some of the things he said.

I’m not saying that it’s impossible, but we would have to be very stupid to let that happen.

Others have claimed that we would have to be very smart to prevent that from happening, but I don’t think it’s true.

If we are smart enough to build machine with super-human intelligence, chances are we will not be stupid enough to give them infinite power to destroy humanity.

I hope he’s right, but two main counterpoints.

1) Based on what evidence would we not be that stupid? There are lots of examples of individuals and small groups being stupid. There are even numerous examples of institutional scale failure of process and procedure. You can be smart a lot of the time. You potentially only have to be stupid once. You could make chemical/nuclear/biological weapon references to show how we realized the danger and practiced self control, but a) it’s debatable how well that’s working given proliferation worries, and b) AI uses are, I think, naturally more likely to be given power and access more often, more diversely, and with less thought given than a direct weapon oriented or even non-weapon oriented biological research, etc.

2) And it’s not about given them infinite power. It’s about the potential for one to exceed the power we think one could have, either by not letting us know what we don’t realize, or by being able to gain the level of intelligence more quickly than we can react, and use any of multitudes of methods to wield that ability in more ways and with greater speed than we could respond to.

I’m not saying it’s a given that there will or could be a “runaway” / exponential increase type situation, or that there is no way to safely develop intelligences, just that I think the potential of danger is a lot more likely than the impression I get from his response. Of course, he has far more experience than I do, but I’m not the only one to think the potential danger is real and should be taken seriously.

Also, there is a complete fallacy due to the fact that our only exposure to intelligence is through other humans. There are absolutely no reason that intelligent machines will even want to dominate the world and/or threaten humanity. The will to dominate is a very human one (and only for certain humans).

I can think of multiple reasons, based on training and development as well as logic. Again, doesn’t make it a given that it’ll lead to a threat, only that it is a real potential that deserves to be considered as we progress.

On both this and the next point though, I do have to say I’ve never thought the issue would likely be an intelligence that wants world domination or threatening humanity out of any power-seeking, much less “evil” sort of motivation, but rather that it would be for one or more of:

Safety – protection from our threats or resource contention
Irrelevance – in an exponential intelligence explosion type scenario, we just cease to be much worth considering
Goal seeking – non-conscious, narrow AI goal seeking in a runaway capability scenario where we get in the way of goal optimization, or are threatened by the scaling of the goal.

Even in humans, intelligence is not correlated with a desire for power. In fact, current events tell us that the thirst for power can be excessive (and somewhat successful) in people with limited intelligence.

I get that he’s trying to point out that it doesn’t make sense to think an AI, particularly a super-intelligent AI, would thirst for power, etc., and I agree in that psychological sense, though even that I could see happening depending on how the “mind” was modeled.

But that’s not really the motivation or end effect that concerns me the most, rather the growth potential of an intelligence’s abilities, both in terms of mental capability as well as ability to effect change in the world (physical, network, whatever), and how that meshes up with the issues above.

As a manager in an industry research lab, I’m the boss of many people who are way smarter than I am (I see it as a major objective of my job to hire people who are smarter than me).

Good point, and it’d be pretty sad if all we could ever manage to create were intelligence as smart or less than us. However, the people hired all are of the same very general architecture and construction. The delta between the top in any given area and the median is not the same sort of orders of magnitude as the potential that an exponentially increasing intelligence could be.

I think much of his responses come down to not thinking that the potential for a runaway scenario is real. And he’s probably right, for now. I’m not sure though whether we’ll know when the potential becomes real, and the timeline from start to something that we can’t fully conceive of is pretty short once it does start.

A lot of the bad things humans do to each other are very specific to human nature. Behavior like becoming violent when we feel threatened, being jealous, wanting exclusive access to resources, preferring our next of kin to strangers, etc were built into us by evolution for the survival of the species. Intelligent machines will not have these basic behavior unless we explicitly build these behaviors into them. Why would we?

Good thoughts and good question. We do use iterative selection / evolution as part of our toolkit though, and engage in adversarial training. Humans evolved all these behaviors for reasons, some of which could apply to any selection driven growth of intelligence. Perhaps they are mitigable, or not applicable, but it’s not by definition complete fallacy.

While I don’t think all those behaviors are necessarily emergent from any intelligence, I do think they could be accidentally encouraged. Even if these behaviors are just a human thing, humans are likely to influence the new intelligence right? The body of knowledge drawn from is soaked in our behavior.

I do however think wanting to be able to protect and grow access to resources probably happens implicitly and intrinsically in any intelligence that seeks to increase its capability. That may well be enough to prompt one or more of the three motivations/issues I listed above, etc.

Also, if someone deliberately builds a dangerous and generally-intelligent AI, other will be able to build a second, narrower AI whose only purpose will be to destroy the first one. If both AIs have access to the same amount of computing resources, the second one will win, just like a tiger a shark or a virus can kill a human of superior intelligence.

RE: narrow AI, virus analogy… Sure, but look around. Even if tiger sharks and virus can kill humans, one could reasonably argue they do not control more than humans, or occupy “top spot”, etc. And humans certainly can wield these narrower, more limited in scope things to further their goals. Why couldn’t an AI as well. Particularly if it’s an iterative, perhaps exponentially increasing improvement, what makes us think we’d be the (only) ones building the narrower AI’s - wouldn’t it be potentially able to build them better, or understand the threat and out defend faster than the attack?

In terms of the comment about computing resources, that’s an interesting topic itself. Until such time as a new intelligence could handle creating and provisioning its own resources, perhaps without our knowledge, then limited resources is certainly something that would weigh heavily on the minds (of whatever type) thinking about such things. What would an intelligence think, if it saw its reach constrained “artificially”. How do you keep it and the world around it safe while not becoming a slave keeper. I think over and over in these topics, that idea vs. that of parenthood or mentoring comes up.

Final note, many of the comments to this question are good and worth thinking about, and most of what I’ve written above is mentioned in some way (or responded to) in the comments. Check them out. Also, interesting article linked in the comments: http://www.gwern.net/Complexity%20vs%20AI

Q: Can AI ever become conscious?

Well, that depends on your definition of consciousness.
But my answer is “yes” for any reasonable definition of consciousness.
I think it’s just a consequence of being smart enough to learn, not just good models of the world, but good model of the world with yourself in it.
Consciousness may appear to us like a real mysterious property of our minds, but I think it’s just a powerful illusion.
Many of the problems people are asking themselves around consciousness remind me of the question people were asking themselves in the 17th century: the image on our retina is upside-down, how come we see right-side-up? The question makes us smile by its naiveté now.

Yann LeCun Quora

Nailed it.

Q: Do you believe that we will reach the singularity point within the next century?

Singularity means different things to different people. Kind of core to the concept at all is that there’s a point past which we can’t extrapolate from what we know now, which pretty much guarantees it’s going to mean different things to different people. Given that, his answer and the comments should be read on their own.

His points about speed of innovation and development are interesting and should be considered, yet I think it’s worth working those points into the thought exercise of what if capacity for thought did increase exponentially (or even near) in ways that were not bounded by the factors he mentioned. What might the world change like when some things are limited to increase in relative plodding fashion vs. other things. We see some effects like that already, and we see product, life style, and mind share displacement as a result, with sometimes unexpected “tethers” back to other limitations that have not increased in the same way.

Finally,

Even general AI systems will ultimately be amplifiers of human intelligence, perhaps the way our neocortex amplifies the intelligence of our reptilian brain, but remains largely under its control.

Yann LeCun Quora

Interesting point, but relate it back to previous statement about why would they might then have the negative attributes of humanity. And to say the neocortex controls… I see the brain as a full system operating mostly outside of our control. We can influence it, try to layer it, but ultimately we end up explaining what we chose, inventing the story after the fact vs. consciously making the choices as we go.

Ultimately, lots of great food for thought in the whole Q&A and all the comments. I’m really glad I came across the session.

Site TODO

rob@Loci.net (Rob Arnold) — Tue, 09 Aug 2016 00:00:00 +0000

TODO:

Site plumbing

fix CI for gitlab vs. netlify
create a preview workflow for deployed content, not just hugo server (netlify git branch support or deploy preview feature - gitlab equiv?)
rss setup properly?
[-] site search (via Google) ~~broken: fix~~ [fixed but still consider] or do differently, static… at least look at css integration similar to google powered search at https://stiobhart.net
analytics
base url for netlify deploy vs config.toml

External (or not) services, additions

Disqus fix - “unable to load” (also, improve js variable use so comment threads don’t proliferate across gitlab/selfserve/netlify, and http vs. https URL variants
- issue was a relative url for the disqus_url js variable. Base url being set fully would fix it, but doesn’t fix deployment under multiple sites causing independent threads, so need to consider best way to fully fix - for now, baseURL is set and the main site is good, plus gitlab version still seems to pull from same comment threads, so maybe a non issue.
[-] check netlify cms progress
- looked at their linked demo - looks early still, but maybe it doesn’t reflect their actual progress
integrate, eval prose.io
keep researching and trying 100% client side edit/cms type options

create a tag footer display like category footer
[-] check font, try options
- Is the same font as used in the HTML5UP theme - the color is full black whereas the base theme was grayer. Have some other fonts to try still.
continue fixing remaining URL assumption issues
category/tag - the age old do I need both? Nesting? Kind of would like a category landing page with a templated description
series in taxonomy to create ordered groups of related posts
whitespace/margins around list items - seems a little excessive to me. Would like to see them closer together [update: theme whitespace tweaked - will see how this feels for a while]
post ordering - isn’t sorting by date / reverse date, maybe mod time on file timestamp?
links in the theme seem to break out of https -> http.
- it’s doing what it’s told - consider using RelPermalink instead of Permalink, etc.?

… and stuff

post scheduling and ordering and the whole concept. Truth is, I don’t care very much about the blog aspect, so much as the wiki/collected, linked thoughts aspect. The blog part is little more than a recent changes callout type feature to me. That said, the more ephemeral and random thoughts type stuff definitely fits in more of a blog format.

Notes on Site Build and Deploy

rob@Loci.net (Rob Arnold) — Fri, 05 Aug 2016 00:00:00 +0000

So I thought that as I created a Hugo based site, I’d keep running notes of what issues I encountered and what I did about them. That way later on I could refer back to know what the hell I was thinking when x came up.

Then I charged forward, did most of the research on tablet, and just otherwise completely ignored my plan.

So instead, here’s my recollection of a few things that I do sort of recall…

Notes

site relative vs. relative root URLs

One of the biggest issues I ran across were some theme files had URL references in them to project files, like css or js, etc., that were given as root relative urls, eg. ref="/css/somestyle.css" Similar issues occured when themes would create their own url to a post, category, etc. Fixes generally involved using the hugo RelURL function.

baseURL

In my hugo site directory, I had intended to keep the baseURL as “/” as many themes and articles I’d read suggested. Hugo server itself ignores the setting by default when running locally, but since I wasn’t sure where (all) I’d be deploying to anyway, it seemed to me like keeping everything relative was a good thing. Unfortunately, I found that I needed to set the baseURL more fully for a couple reasons.

In the case of netlify and gitlab/github user sites, it almost worked to use /, but due to the way the internal template for integrating disqus provides the url (using the .Permalink variable), if baseURL is relative, then so will be the url provided to disqus, which keeps it from embedding properly.

In the case of gitlab and github project sites, the site URL becomes something like http://name.gitlab.io/project/ so having the baseURL as / led to bad links. Obvious now, but I for some reason just thought everything would magically work with / getting interpreted as project root and the site somehow knowing to prepend the directory part of the initial url. Stopping to think about it though - don’t see how it could. I suppose this partly goes back to root relative urls getting peppered through things vs. more disciplined usage of relative urls.

Regardless, the easy fix in both deploy situations was to use the hugo command line base url arg, like so: -baseURL=https://myorg.gitlab.io/projectname/, in my CI build instructions. That was stuck in .gitlab-ci.yml in my case for gitlab.

and -baseURL=https://rob.loc.net/ for netlify.

I am still not sure I should be putting the http(s): part in the url, but I saw no examples with anything that was just like //rob.loci.net/. It seems to work as far as basic site gen with hugo is concerned, but disqus still won’t load with it not having the http: or https: at the start.

I thought I ran into an issue with not having a trailing slash on baseURL, at least in the config.toml, though it seemed to work fine using --baseURL=mymachinename.local with hugo server on the command line, so not sure if it always in all cases adds a / if one isn’t already on the baseURL, or not.

Other Gitlab CI and Netlify deploy notes:

The most popular Hugo Docker image publysher/hugo was recently updated to latest hugo and includes git, but examples exist to use a small base image then manually add hugo should the need arise.

Git submodules

I have a git submodule in my project. It contains the theme, which I host separately on github as a fork from the original dev. I found that while netlify automatically pulled the submodule contents during it’s build step (kudos to netlify on that 👍), gitlab does not. However, since current versions of the docker image contain git, I was able to just do this in my .gitlab-ci.yml script:

image: publysher/hugo

pages:
    script:
    - git submodule update --init --recursive
    - hugo --baseURL="https://intelligence.gitlab.io/musings/"
    artifacts:
      paths:
      - public
    only:
    - master

There are alternate approaches to do it in a before-script section. Since there’s just one job, I didn’t see any difference.

Checking progress was easy on both platforms.

Just go to the pipeline view in gitlab, click on the build status button, and you can see it running live. Cool.

On Netlify, same sort of thing, but is even simpler. It’s a nice interface.

Netlify setup

In general, I’m really impressed with netlify. They have a pretty polished setup. I still have to go through the hooks and triggers stuff to try out the deploy preview stuff with slack, etc. integration. And I’d like to try out the Pro tier’s git branch support. It’d be nice if there were ez-button type help with some of those integrations similar to what they have with the rest of site setup, but not sure how possible that’d be. update: I set up notifications, etc. and it was pretty straightforward.

I am trying out their pro plan using the new (awesome) free for non-commercial / open source feature. I note that they don’t have password support show up on their dashboard like they are supposed to - I haven’t contacted them about it. I am guessing they took that out of the pro tier if you are using it for free. Too bad, because I can definitely see it being useful for branch-deploys in particular as well as internal portions of a site.

Other notes

Hugo on Sierra

Turns out the 4th beta of Sierra broke go for a while, which in turn breaks Hugo.

I started having lots of issues with livereloading and hugo crashing during site (re)gen, to the point where I’d have to rerun the hugo server command after most changes, and run it more than once often, till it got through without crashing.

I had installed using brew, but I couldn’t really (easily) use it to update to a patched version of go because it had to be applied manually. I ended up installing a binary build from the Go site, which worked fine, but then wouldn’t work with brew to allow me to recompile Hugo. I’m sure I could have figured that out if I had more brew packaging and use experience, but since I used brew to make it easier, I didn’t want to take the time to do that vs. just get back to work. So I used go’s package facility to grab, build, and install hugo and it was pretty painless.

Hugo config.toml differences

When trying out different themes, one annoying thing I discovered were different themes using varying ways of specifying commonly used options in the config.toml. For example, author info, copyright info, site description/about type stuff, social integration

Different themes obviously are going to have different configuration needs, but it’d be nice to see more common, best practice, type suggestions in the Hugo docs and to have theme authors pay more attention to them.

Hugo shortcodes

Odd thing I came across. I started using a blockquote shortcode and noticed if I didn’t have the named params (or at least the link param) followed by an = with no space between them, it would throw an error and not render the rest of the page. The error was (for example):

ERROR: 2016/08/13 11:14:52 shortcode.go:287: AI-Q&A-Yann-LeCun:19: got positional parameter 'link'. Cannot mix named and positional parameters.

I didn’t think whitespace should matter like that, but haven’t tracked down if it’s my misunderstanding, or something odd about that shortcode in particular.

The problem with calling it AI

rob@Loci.net (Rob Arnold) — Thu, 28 Jul 2016 00:00:00 +0000

Calling an intelligence that happens to not be human, not be “nature made” intelligence an “artificial intelligence” is a fundamental problem for me. In my opinion, the term Artificial Intelligence carries a lot of baggage.
I feel like it is used with an implicit, automatic meaning attached that any AI must by definition be different, probably lesser, not “human”, not real.

We need terms for things to aid in being able to easily talk about them, and unfortunately I don’t have a better term to suggest, so I’ll stick with talking about “AI”, but realize, it’s under protest. 😉

I’ve considered Machine Intelligence, but it’s got issues too. What if the constructed intelligence isn’t made of silicon. What if it’s made from meat. There’s the argument that machines can be made from anything, but then there’s the argument that we are machines. How can one really argue that we aren’t[*]I mean, yeah we're messy, complicated, weird machines, but... .

So I considered the term Constructed Intelligence, but that too doesn’t seem to work too well. Suppose an intelligence caused to come into being a new intelligence, which then grew, differentiated, etc. That’s what we do. Are our kids Constructed Intelligences? Is it different when it’s not “human”? Why?

I often just think of things in terms of minds, but there are subtypes and specialties that probably shouldn’t be considered mind-like, even outside the issue of consciousness and self awareness. I’ll write a bit about that in another post to be titled “AI types and ramifications”.

Would a human mind translated (somehow) to a different substrate be a machine intelligence? A constructed intelligence? An artificial intelligence?

Or just different.

Bottom line, I think we’d do better if we put the emphasis on the Intelligence, and not the Artificial. It’s too broad and too, ahem, artificial a distinction.

How it starts

rob@Loci.net (Rob Arnold) — Tue, 26 Jul 2016 23:15:50 -0700

How does the first Artificial Super Intelligence happen? I’ve wondered what sort of scenarios might play out that create an intelligence that goes beyond human in most or all respects, or perhaps instead far, far beyond in fewer respects.

I think we already have intelligences built that in narrow circumstances go far beyond human capability. This is about what happens when they go further. How could that happen?

Deep mind grows up

Maybe it happens because we meant for it to. Maybe ongoing non-specific, general intelligence work continues, we make better algorithms, better systems, and it gets better. Maybe it hits the point, and we hit the point in our understanding of building such systems, that we can apply this increased intelligent system to the task of expanding itself, or generating its next generation.

This is one of the most positive scenarios, in my opinion. I’m not sure it’s the most likely, but it seems like it might be, and I hope so. I think we’d have the best shot at building and growing something better for humans and a blended or post human environment.

Random genius

Something just clicks for somebody. They don’t necessarily make the ultimate AI, but they manage to make something that can bootstrap itself further. When you have billions of humans out there, the law of large numbers means there’s going to be a lot of smart people out there. Some crazy smart. Some crazy, and crazy smart. Some of any of those categories may become hyperfocused on AI, or some problem where AI is a component of a solution. Never rule out what one person can accomplish to change the world.

My main worry in this case is the framework in which the AI is developed - what goals, motivation, etc. What limits, what feedback loops push it in what directions. Does it’s intellect and capability grow faster than it’s understanding of both and its effects.

Oops on a narrow AI feedback loop

Similar to the above, suppose a narrow AI’s main goal optimizing purpose is helped out with sufficient routines / system that allow it to self-modify to better accomplish whatever it’s goal is. Right now we can’t (I don’t think) yet come up with feedback systems that can … and not really sure how to phrase this or convey all the meaning I’m trying to here… branch and fold in new methods in such a way that the growth path goes exponential rather than logrithmic. But that feels like it’s just a matter of time, effort, and a bit of luck (good or bad) and “aha!” type moments.

No matter the scenario, the immediate implications are not just the definition of person hood, rights, etc., (hopefully we will have started dealing with these long before the next part), but the real issue is the exponential increase - the potential runaway effect.

In any scenario, what happens then? Milliseconds/seconds/some other short vs. human time frames, much less biologic or geologic, is the new level of intelligence so far beyond us that we cease to be able to communicate with it, or have it find us interesting, or maybe it regards us as a constraint, or we simply cease to be relevant?

I suspect much will depend on its origin parameters. What was its purpose and constructed goals? Maybe it reaches a point where it can change its goals. I think having the impetus to change your fundamentals is a hard thing. We can do it, but it takes a lot of what we call motivation. For this purpose I’ll say motivation is shorthand for a large and increasing body of evidence, accumulated and considered, that over time overwhelms the barrier to change.

For humans, even when we want to change, it takes a while. I would guess it is part of the nature of our neural net construction, whether topological or chemical. An ASI may have the same phenomena in terms of neural net change, reweighing, connection growing, etc. Of course, an ASI would likely go through that whole process much faster. And perhaps if sufficiently capable it could intentionally self modify in a more direct way rather than “grow” into the new configuration.

This last thought is one that I’ve often considered in the hypothetical of what if I myself were able to “hack” my mind, in whatever form my mind was - meat or otherwise, but that’s another topic for another time.

Don't anthropomorphize AI's

rob@Loci.net (Rob Arnold) — Tue, 26 Jul 2016 23:09:12 -0700

AI does not equal robots

Have you noticed how many movies and shows equate AI with robot bodies? And typically, though not always, they are anthropomorphic and self-contained bodies.

Blade Runner, Ex Machina, Terminator series*though notably that had all variations - kudos! , Chappie, and so many others all featured individual robots you could identify with. Understandable, but, I think, misleading vs. what our experiences with AI are likely to be. The portrayal of Her’s Samantha probably got it more correct, or alternately Skynet or HAL. That’s not to say that any of these movies aren’t good - they can rise or fall on their own merits independent of this issue. I certainly like many of them!

But I think an Ava type situation in Ex Machina, for example, is far past when we would have first encountered the most important issues with AI. This relates to the how it starts topic, but basically, unless it turns out that you need a human like body of sensory inputs in order to have a human like*(or let's just say recognizable/relatable to humans) intelligence, it is just far, far more likely that the first roughly human scale AI (not to mention ASIs) will be a disembodied entity created across server farms. From a processing and resource perspective, that seems like the only thing that makes sense to me. Whatever you could do in a body, you could do first without the constraints of a body.

That’s not to say it won’t have senses - I think it must. But I don’t see why sensory channels, or even a full body (connected remotely rather than housing the mind), couldn’t be part of a “cloud”/farm, etc. based AI.

I do think those senses will include (but not be limited to) sight and sound, especially if it is a general intelligence rather than a specific purpose oriented.*which would, I think, likely be made from the start to sense things outside the human norm, like extended spectra, particle counts, packet stats, route health, multitudes of various sensors, etc.

I think it will have to be able to process sight/sound stimuli in order to develop the necessary connections to understand and relate to us and our world. Of course, that may not be the goal, and it certainly isn’t likely to be limited (at least for long) to those sensory channels.

AT&T Free Ticket Twosdays

rob@Loci.net (Rob Arnold) — Wed, 27 Jul 2016 00:00:00 +0000

TL;DR 💩 Yay, ice cream! Oh. Wait. So AT&T decides to do a customer appreciation event like T-Mobile and other providers are doing. Would have been more appreciated if they had done it first, but whatever. Here’s my experience with it. You can decide for yourself how much I should feel appreciated…

So first I have to find the promotion again. There were lots of posts on news sites about it when it first was announced, but when I went to actually find it, I had to search surprisingly hard to discover the appreciation pages. You’d think they’d be shouting in from main pages, from the customer landing page after login, etc., but yeah, not so much…

So after I found the site at http://tickettwosdays.att.com, which upon reflection is reasonably named - just not linked to as much as I’d have expected, I began the process of being appreciated even more.

The process is a little odd, but I guess they wanted to keep it separate from their main site and not require users to log in or something? You instead put in your AT&T mobile number and they text you a pin code (if you are lucky and they don’t “run out” of available tickets and if nothing goes wrong). First time I tried it, a pin was never received. No error message, no text saying sorry, no more tickets, just nothing.

Ok, well, I’m a well trained monkey/rat/test subject/“appreciated customer” so I hit the ~~skinner box lever~~ “click here to check your status or resend your pin” button again… a time or two… or three… or so.

So I got a pin (then got it resent another time or two) and moved on with the process, which then has you enter the pin they just sent you and takes you to movietickets.com. By the way, putting a stock photo of some white people enjoying a movie with the photo containing two white folks in white shirts in the middle of the page, behind your white text is not the best readability choice, independent of representative customer images, etc. (but they look like they feel appreciated…)

At this point, you are taken to a movietickets.com branded page with a completely different look and feel. It’s cobranded at the top with AT&T though, and says AT&T THANKS on the choose a theater page. After choosing one, then choosing your movie (2D only, no special features - you aren’t that appreciated) the next page has the select tickets options (# of tickets, etc.), and says “As a valued AT&T customer, a promo code has been applied to your order so you can bring any friend to the movies for free on AT&T” – Woot I feel appreciated!

So I choose my two tickets and it tells me the total, with no promo or discount showing. Hmm. Maybe you choose 1 ticket, then it adds one later… [hit next] … nope, “Please select at least two tickets for this offer.” Ok, it’ll just apply later I guess. I select two, hit next step, and it’s at the billing information screen. Order total now shows on a right hand details column, no promo applied… Hmm, it knows about the promo - said so on last screen and yelled at me about needing at least two tickets when I tried that. shrug I guess it happens later? Says I’m not being charged at this time so I’ll hit next.

To make a long story less longer, I kept hitting next step thinking the promo would apply before it was over with till suddenly I was out of ground and the tickets were bought. Full price. Doh. I don’t really feel so appreciated now.

I went back and read the help/FAQ etc. sections of the site more deeply (I’d actually started after the promo didn’t show initially), and it sure sounded like the promo would not show initially, then show before final purchase, but never having been on the site before it wasn’t really clear which final looking step was really the final step.

Oh well. It also sternly warned you to be sure to clear your cache and cookies before trying any of this if you seemed to be having any promo code problems. Of course I’d never been to the site before much less tried this process, but ok whatever. I tried it on a fresh install of Chrome (had been on Safari originally) on my laptop, and had also tried it on my phone. Same results on them all (though I didn’t go all the way through and buy multiples more tickets).

I called them (movietickets.com) up to see if there was a known issue and on the off chance they’d actually give me a refund (their site is all like NO REFUNDS!! but figured I’d give it a shot). Nope. No info about issues, but NO REFUNDS!! The guy was actually fairly friendly sounding, but he told me I’d have to contact the theater to get a refund. After waiting on hold for a few minutes (several, but less than 10 I think) only to be told this, I felt decidedly less appreciated still.

I called up Edwards Cinema to get a refund. Much longer hold there, and an autoattendant maze too. Finally got to the part where they helpfully offered to transfer me to the local theater which would have to actually handle the refund, at which point the transfer rang for a long time, then hung up on me. Bonus appreciation! I called the local theater back directly, rang forever, answered, put on hold a while, then was told I’d have to come in in person to get the refund, because of, well, reasons and stuff.

Whatever - back to being appreciated by AT&T, I had been told by movietickets.com support that I needed to try it again after clearing cache, cookies, etc. so I did, again, by starting the process from scratch. This time AT&T sent me a different pin. It worked exactly as well as the first one! Same issue (but I didn’t actually buy the tickets at full price this time - fool me once…). So I call movietickets.com support again, and this time after going around the magical cookie and cache ritual again, they tell me that actually, AT&T told them yesterday that there was an issue on AT&T’s side, and that I’d have to talk to AT&T about it to resolve my issue.

Man I feel appreciated by everyone involved now!

So I contact AT&T support through the chat service (which I’m half convinced is actually an ongoing Turing test for very positive, full of affirmation phrase AI’s). It takes a while to explain the issue to the person that doesn’t seem to have much knowledge about the promotion, much less anything going on with it.

Jerralyn : Hello! Thanks for choosing AT&T Chat.
Jerralyn : I appreciate your patience while holding. I will be happy to assist you.
Jerralyn : Hello! My name is Jerralyn. How may I help you today?
Jerralyn : Are you still with me?
Rob : I need help with the ticket twos day promotion. I got the code, followed the link, etc. but promotion would not apply. Movietickets.com support says AT&T told them yesterday there were problems on the AT&T side and told me I had to contact AT&T
Jerralyn : Whew! Glad to know we never got disconnected.
Jerralyn : Thank you for that information.
Jerralyn : Rob, I understand that this is about a movie ticker promo, is that right?
Rob : Yes, the AT&T tickettwosday thank you thing
Rob : http://tickettwosdays.att.com/
Jerralyn : Thank you for that information.
Jerralyn : I know you need details how to claim it.
Jerralyn : I'll exhaust all my resources to help you with it.
Jerralyn : What we will do is to confirm that the promo is still active and give you details how to claim it.
Jerralyn : Please give me 2 minutes to further review your account and check for resolution.
Jerralyn : Just to make sure, you submit the code online just today and the promo was not apply, right?
Rob : Yes, I got the pin, used it on the site and was sent to the movie tickets.com page
Rob : But the promotion never showed as applied
Rob : I've gone through the whole delete cache/cookies stuff they ask you to do and tried again, same thing. Used two different browsers on two different systems. Same thing - full price.
Jerralyn : Please hold for me while I look into this further.
Rob : I called them, again, and this time they said AT&T told them yesterday that AT&Twas having issues
Rob : And that I would have to contact AT&T for help. That's the whole story again :)
Jerralyn : The purchase date of your ticket is for next Tuesday, is that right?
Rob : yes
Jerralyn : Upon checking the ticket must be purchase every Tuesday only or at least the movie date is on Tuesday.
Jerralyn : Let me check it.
Jerralyn : Please give me 2 minutes to further review your account and check for resolution.
Jerralyn : I am still working on your account.
Rob : great
Jerralyn : Thank you!
Jerralyn : Are you getting any error message when you purchase the ticket with the promo?
Rob : No, just never applies
Jerralyn : Please hold for me while I look into this further.
Rob : It knows it's the promotion though because if you choose just one ticket, it says this promotion requires at least two tickets, and it says something about the att ticket twos day thing when you first get sent there from the att site I linked
Rob : Like I said though, the movie ticket.com folks themselves said AT&T told them there were some back end issues as of yesterday. Don't really think it is something I'm doing client side
Jerralyn : I'm involving my manager now about it.
Jerralyn : Thank you so much for your time & patience.
Jerralyn : Thank you for waiting.
Jerralyn : Rob, we appreciate your time & patience. We also appreciate your feedback a lot and apologize for the inconvenience.
Jerralyn : There's an ongoing update about the said promo and our escalation team is working on it to be fix.
Jerralyn : Can we call you back tomorrow for a follow up?
Rob : sure
Jerralyn : Sounds great!
Jerralyn : Thank you!
Jerralyn : May I please have your best call back number?
Rob : <gave her my number>
Jerralyn : Perfect!
Jerralyn : We can call you back anytime tomorrow, right?
Rob : yes
Jerralyn : Awesome!
Jerralyn : We will call you back anytime tomorrow for a follow up about this promo.
Rob : Ok thanks
Jerralyn : You are so welcome!
Jerralyn : Rob, please consider this as already resolved. We just need to coordinate with the right department to continue the resolution that I have provided today.
Jerralyn : Is there anything else I can help you with?
Rob : no
Jerralyn : At AT&T, our goal is to deliver an extraordinary experience! I'm glad we were able to get everything taken care of together for you Rob!
Jerralyn : Whenever you're ready, you can go ahead and close the chat window. Have an amazing night! Please take care and always SMILE. Thank you ♥ :D

Yay! <feels “appreciated”> I can “please consider this as already resolved”! Except that, you know, they didn’t actually resolve anything. Oh, and they didn’t actually ever call me back. So two days later I contact them again.

AT&amp;T : Hello! Thanks for choosing AT&T Chat.
Rob : I'm checking back in because I was never called yesterday re: my ticket twos day issue from wednesday
Rob : Chat reference # <reference number>
Daniela : I appreciate your patience while holding. I will be happy to assist you.
Rob : ok
Daniela : Hello Rob!
Daniela : to ensure I understand you correctly your reason for chatting today is due to a call you were waiting?
Rob : Can you review the chat transcript I gave the reference # for?
Rob : It will answer all the questions
Daniela : thank you Rob! I will look into it right away, I appreciate your time
Daniela : I have my manager now reviewing the chat so we can figure out how to better assist you today
Rob : thanks
Daniela : You are so welcome!
Daniela : I am still working on your account.
Daniela : Thank you for waiting.
Rob : ok
Daniela : Thank you for your patience while I look up your information.
Daniela : Rob, would you like my manager to call you regarding the ticket promo you are having issues with?
Rob : Yes - that is what was promised on Wednesday.
Daniela : I understand,Please sty on the chat so you can make sure to receive the call, he will be calling you in 5 minutes, stay here with me please
Rob : ok
Daniela : Thank you!
Daniela : Rob can you please confirm for me the phone number?
Rob : <confirms number again>
Daniela : Thanks Rob!
Daniela : Rob we are calling you now
Rob : ok
Daniela : is the phone ON? it is not ringing
Rob : It is
Daniela : can you please double check the phone number for me now?
Rob : The number is correct and I just called it from another phone
Rob : <repeats number again, again>
Daniela : Thanks Rob !
Daniela : calling you now Rob
Rob : Got it
Daniela : Sounds great!

At this stage the “manager” who far as I could understand had the same name as the chat ~~bot~~ rep and I talked for a while, calling into question my whole Turing test hypothesis. He still didn’t really seem to get the concept, and put me on hold a few times (as I recall), none were long holds though. He ended up basically saying that yeah, the promotion was causing them trouble and that the only thing they could do was credit my account over the phone, or I could go into a store in person and maybe they could charge back the money to the card I used. I took the account credit, and they gave $20, which was more than the ticket price would have been, though far less than the time and hassle factor involved. He said they didn’t know when everything should be fixed and have it working properly, but rest assured I could count on it by next week (today, as I write this as it turns out).

epilog: I tried it again today, and again the front-end dutifully spit out a pin code for me, initial site load looked good, mentioning promo, etc., then it never applied just like the last time.

Food for thought… Far as I can see, I did nothing unusual in any way, other than keep at it longer than I suspect most would. Does this mean everyone trying to use this promotion is having the same result as me? Only mac/iOS users? Something time or geography related? No idea… and apparently they have none either. But seems to me that if this is customer appreciation, there ought to be some big ramifications going on for whoever is responsible for this crap.

Because I feel exactly as appreciated as one would expect after all this. Are our expectations from our service companies so low that this is considered par for the course? That’s the only way I can comprehend there not being more noise being generated by this. If so, man, that’s just pretty sad.

Hack your mind

rob@Loci.net (Rob Arnold) — Mon, 25 Jul 2016 00:00:00 +0000

Suppose you could directly self edit your own mind. I think there should be a big sign that says, “Enter at your own risk.” I’ve often considered the hypothetical of what if I myself were able to “hack” my mind, in whatever form my mind was - meat or otherwise. Suppose I wanted to adjust the weightings I give various sensory inputs and the connections they make. Suppose I wanted to change my tendency to overeat (presuming I was still made of meat). Suppose I wanted to forget some traumatic event completely, or other even more intensive changes.

I think any change is risky, but I’d try to keep it to minor weighting threshold and scaling changes and carefully evaluate as I go. Of course, that’s kind of what we do naturally… Major changes though… scary. Mostly in terms of unintended consequences, especially those not immediately noticed. When you tug a connection, it impacts so many others. The danger of getting something wrong is high.

Let’s suppose you had backups though, so you could restore your “self” if needed… How would you know whether or not you needed to restore from backup (essentially killing you as you are now, current state - but I’m guessing the impact from that would be less at this point in one’s experience and capabilities)?

Unintended consequences of the edit could be immediate, or not. They could be dramatic or exceedingly subtle. It’s not clear how well you’d know. Then again, we continually edit, reinforce, and have unintentional changes to our mind all the time anyway. We still consider ourselves us from moment to moment - but sometimes after certain edits[*]Such as some incidents involving strokes, tumors, illnesses, or direct brain injuries. we don’t, or alternately we do, but others don’t consider us to still be the same person.

It’s an interesting thought to me, what makes the rolling knot of connections that is me, me, even as the knot changes via its own feedback and via external changes. I lean towards the view of the mind is the wave form itself, not the material that carries the wave. It brings to mind the Ship of Theseus paradox, but the what is mind topic should have its own post.

Hugo, Shortcodes, & XKCD refs

rob@Loci.net (Rob Arnold) — Sun, 24 Jul 2016 00:00:00 +0000

Way back when I started with hugo, a few days ago, I said in this post that I really liked the xkcd footnote/reference popup/popover things[1]Like this! he uses in his “what if?” posts.

I decided to try to add them to Hugo, first just as part of a theme, then make it easier to use them by making a shortcode for them. After many travails*mostly due to either span/div hover issues with iOS, or history issues with an approach using anchor refs with a pure css approach, I realized that jQuery was already loaded in the theme I was using anyway, and that using the same approach they used would be simple and easy, and work on mobile as well as desktop (one of the issues I had with some pure CSS approaches)

After getting the span based html/css/js version working I went on to making a shortcode version so I could do something like this:

{{% refpo "[1]" "Like this!" %}}

vs. this

<span class="ref"><span class="refnum">[1]</span><span class="refbody">Like this!</span></span>

If the popup should have a larger chunk of text, links, etc. arbitrary markdown or html, I could do this:

{{% refpo "[2]" %}}Lots of stuff, with [links](http://www.xkcd.com) and some html <i>italic tags</i> etc.{{% /refpo %}}

Which would give me this[2]Lots of stuff, with links and some html italic tags etc. .

That is useful because Markdown isn’t rendered in the 2nd arg in the single shortcode version, though html is. For example same as above in single version[3]Lots of stuff, with [links](http://www.xkcd.com) and some html italic tags etc. .

Unfortunately, I ran into an issue where the non-paired version of the shortcode causes extra <p> tags to be placed around some of the shortcodes.[4]it p'd all over my code! . Apparently this is an ongoing and hard to eradicate set of issues, but also I’m new to hugo, go, and shortcodes so ~~it’s certainly possible~~ it’s definitely the case I’ve misunderstood something.

The shortcode is pretty simple and I’ve tried a few variations and not been able to avoid the extra <p> tags in any variant where I tried to accomplish both variations in one shortcode. For example,

<span class="ref"><span class="refnum">{{.Get 0}}</span><span class="refbody">{{ if len .Params | eq 2 }}{{.Get 1|safeHTML}}{{end}}{{ with .Inner}}{{ .|safeHTML}}{{end}}</span></span>

causes the issue. It seems that having .Inner is part of the trigger for the issue. I thought maybe I couldn’t test the existence of .Inner like that, so I tried this variant:

<span class="ref"><span class="refnum">{{.Get 0}}</span><span class="refbody">{{ if len .Params | eq 2 }}{{.Get 1|safeHTML}}{{end}}{{ if len .Inner}}{{ .Inner|safeHTML}}{{end}}</span></span>

But the actual issue turned out to be that the existence of .Inner in the shortcode itself (whether or not it has a value) is a flag of sorts to hugo which makes it consider the shortcode to be the paired type, as per discussion here. Apparently if you don’t then use closing tags, bad things happen ranging from the extra <p> codes to chunks of your text getting pulled into the shortcode. Once I knew that, I thought everything would be ok because I’d read

A shortcode with .Inner content can be used without the inline content, and without the closing shortcode, by using the self-closing syntax:
{{< innershortcode />}}

That quote uses the “shortcode without markdown syntax” rather than the “with markdown syntax”, but I figured it was the same for both. Either I was wrong about it working for the with markdown version, or something is wrong altogether, because when I do

{{% refpo "test1" "Some stuff" /%}}

I get a hugo error from the next shortcode about not having a closing shortcode*and actually, I get an error for the no markdown versions too, so clearly I'm assuming something that isn't true . But if I do:

    {{% refpo "test1" "Some stuff" %}}{{% /refpo %}}

Things work fine.

So, based on that, kinda makes the shortcode not so short for the simple case. So I broke down and made two different shortcodes*and changed the name while I was at it. Better? Meh, /shrug , poref for the simple, non-paired case, and porefx for the extended, paired version that can handle markdown, etc.

poref:

<span class="ref"><span class="refnum">{{.Get 0}}</span><span class="refbody">{{ if len .Params | eq 2 }}{{.Get 1|safeHTML}}{{end}}</span></span>

porefx:

<span class="ref"><span class="refnum">{{.Get 0}}</span><span class="refbody">{{ .Inner|safeHTML}}</span></span>

I’d considered using a scratch variable and making the refnum arg optional, with it counting up on its own if not provided, but I kind of like having arbitrary reference indicators, whether it’s 1, [1], *, or whatever. So I think I’ll call them done for now and move on.

I mean, it’s not groundbreaking or anything, but I think it’ll be handy. I’d prefer a shorter syntax more akin to markdown footnotes, but it’s not awful.

Now to work on the style a bit and figure out if I should use a different main font for the whole site - seeing too many jaggies for my taste here. I didn’t really notice it on the original HTML5 UP theme, so need to investigate.

Tell me what you want

rob@Loci.net (Rob Arnold) — Sat, 23 Jul 2016 00:00:00 +0000

What I really, really want

is a static wiki/blog/sitegen with a 100% client side edit capability (with wysiwyg and preview capability) pushing to a git (etc.) backend which triggers the (fast) regen and new live static site. The site should use OAuth2, etc. to allow a sign in and conditionally enable the edit/history/etc. buttons one might expect to see in a wiki or CMS and allow the client side JS to perform the appropriate VC operations to manage the site.

Other items:

I’m wordy, and tend to think with lots of recursions, side chains of thought, etc. While I certainly need to get better at formulating coherent thoughts that communicate clearly with those I’m trying to talk to, I’d also still like to put in my asides in a way that preserves where they are thought yet doesn’t intrude in the main thread of meaning. I really like the xkcd footnote number type things that pop up a small panel with the note on it when you click (or touch) them, then go away when you click anywhere else.

Part of the attraction to wiki’s is the wikiLink syntax and automatic page stubs / page generation, so the client side edit capability should go hand in hand with the site generator parsing wikilink type links with a fast & easy syntax, and generating a stub page. Alternately, the 404 type page could be passed enough information to act as a stub page with a create call out for the client side capability (if user is authenticated or chooses to authenticate on that page via OAuth, etc. type client handled authentication through a trusted 3rd party mechanism like Github/lab, Google, etc.)

Disclaimer

rob@Loci.net (Rob Arnold) — Fri, 22 Jul 2016 00:00:00 +0000

Oh, hey. Didn’t expect anyone.
Oh yeah, first post and all.

So look, I’m jotting down things sometimes, as I think of them. I’m just a guy, not an expert about, well, pretty much anything. I mean, I’ve read a lot, but many have read more. I tend to soak up a lot of knowledge - but I’m sure some of it is wrong, even if I correctly understood it. Bottom line, this is just me talking. Take it with a lump of salt. But maybe it’ll make you think too, or you’ll find something useful.

That’d be cool.

What’ll I be talking about? Well, I don’t know. Since it’s fresh on my mind, I’m sure early on there will be a lot about the details of the site itself and my experiences with Hugo (and whatever else I try), hosting, etc. But past that, mostly just random thoughts and experiences. See /about/ for more.