haxel

What about variably-autonomous cars? Take a car with self-driving capability and then have it be manually operated while solitary but increasingly self-driving as the group of similarly automated cars it belongs to grows. There's a lot of benefits to group awareness: share sensors, coordinate behavior for efficiency, instant communication, and so on.

This gives us the best of both worlds: safety in numbers, safety in automation and sensors, and it retains the personal autonomy we're used to. All while not requiring 100% perfect autonomous capability.

If the cars were small and cheap enough, they could be made available via subscription (much like a lease) and treated like a municipal transportation network that's superior to light rail and bus networks.

If the point of this money was to develop a unified approach to self-driving cars, then it could be a very good thing. Assuming the choice of unified approach makes sense.

Variably-autonomous cars operating in swarms (and thus requiring standard communication and sensor packages) seem to be superior than having dozens of car and tech companies each striving for full autonomy. In swarms, the degree of autonomy would be directly proportional to the size of the swarm a vehicle belongs to. Those on the edge of the swarm would contribute sensor input while the interior vehicles would contribute processing power. Thus achieving greater ability than a solitary vehicle, whether autonomous or not.

The advantage of variable-autonomy is that there's much less need to specially-modify roads, change the way we view liability, or have to get all the way to full autonomy.

I wrote more on the topic here: https://medium.com/@SteveHazel/let-s-do-semi-autonomous-cars...

If the $4 billion was spent on developing the standard communication and sensor packages, and promoting them amongst manufacturers, and promoting the development of compact-form-factor vehicles, it might have significant impact.

Not that we know of.

If it does ever exist, how long will it take for us to learn of it? Powerful tools are often kept secret.

And then, how long until the average person can leverage it themselves? The most powerful tools tend not to trickle down, if it can be helped.

What about the political implications of AGI? This concerns me most.

We need to carefully consider who exactly controls such a powerful tool and what they use it for. Right now AGI is capital-intensive and risky: you need large-scale computing, top PhDs, and plenty of time. This means that only the most well-heeled organizations can justify earnest AGI development: big tech, government agencies, etc. Those orgs already have plenty of political power, and it seems to me that achieving AGI (or steps toward it) would amplify that power. Perhaps to the point of being unchallengeable.

Do we want a political landscape where power is even more concentrated than now?

Ah, I meant advertising in the sense that we're seeing tons of advertising for cheap meat, air travel, factory food, new gas cars, and whatnot. A good question to ask might be that if corporations _didn't_ advertise so much, how much less frequently would we buy those types of products/services? Would we still need them? I think the advertisers know what would happen, and so they have to keep it up.

Now, if we decided to throw a few environmental ads into the current mix, I think you're right that they'd have little effect. But we can't forget how much advertising we see for stuff that's harmful in many ways, regardless of how temporarily satisfied we are with our purchases. Because so much of what we buy has serious negative (but hidden) consequences, advertising is causing us to act against our (non-immediate) self-interests.

Does a carbon tax stop the slash-and-burn in Indonesia (for palm oil) and Brazil (for cattle)?

I agree that the full cost of what we buy needs to be included in its price or else we'll keep seeing the same problem, having only kicked it down the road a bit. But there are also many more large problems besides climate change that involve more than carbon dioxide emissions.

You say global economy and then you say far beyond the reach of individual consumers. Yet the global economy is by definition composed of the demands of individual consumers.

You also say that top-down hasn't been working very well and so we may be running out of time. Isn't that a strong signal that we should look at a different approach?

As technologists, we can plainly see the power of our mobile computers and the influence they have over our behavior. Advertising is huge because it works. But what if its influence drives us to disproportionately purchase products or services that contribute to climate change?

If advertisement works, and technology works, and people's behavior is already being directed, then why can't we direct it another way?

I look at the human-caused part of climate change as a billion or so smaller and easier-to-solve (but still hard) problems. Each one of these billions of problems is a person like you and me who acts without truly knowing what they are doing. Not in the sense of faking our way through life, but that we have only a superficial awareness of the consequences of our habitual or culturally-driven actions. This makes climate change an information problem, and technologists do information splendidly.

It's an information problem because we base much of our behavior on feedback loops. When we get a big electricity bill, we take a closer look and perhaps rely less on our heaters or air-conditioner. When we see the odometer on our car hitting big numbers too soon (and repair bills looming), perhaps we look into a job closer to home. There's thousands of examples, many of which have an impact our our environment.

So if we focused on developing a tool whose sole purpose was to give us quantified feedback on the consequences of our actions - much like how we use utility bills - we might be able to make a real dent in the big problem of climate change and others like it. We already have the pieces for a tool like this (Internet, computers everywhere, software libraries) along with the skills of technologists, so what's needed is the vision and the demand.

I recently wrote these ideas in long form: (https://medium.com/@SteveHazel/we-re-drowning-in-low-quality...)

If that information is useless to us, it's due to inadequate tools: software, hardware, algorithms, and philosophy. Facebook and Google are both able to extract value from an individual's information (even isolated) because they have both the software and incentive. Both can look at word frequencies and activity over time to develop behavioral prediction models that are then used to target ads at you.

I'd say the ability to make relevant predictions is a form of power.

Why is it a forgone impossibility for an individual to possess tools to capture similar value from their own information? Does self-knowledge hold no value? In my experience, changes in behavior (such as to better align with your own goals/interests) are often triggered by a clear and quantified message from a trusted source. Why not make this source your own information and your own computer?

Sure, what Facebook has is lower quality and harder to process than we might prefer, but there's nothing stopping us from collecting and structuring our own information in more appropriate ways, then layering on interfaces and algorithms, and then simply doing better than anyone else can. Because it's ours.

So the We is We hackers, who can design and build new information-contexts with different value propositions and power dynamics. And it's also We thinkers who can put together the reasoning behind such efforts. And We as regular people, who want more joy and less pain, with less effort.

Yeah, it does seem like more a problem of philosophy or perspective than a lack of advanced technology. We have all of the ingredients for a variety of dishes, but it seems we're settling into one basic recipe: centralized lockdown.

@s73v3r, Self-knowledge can be tremendously rewarding too.

Smartphones aren't going anywhere, but their value proposition is not the only one available. If we were to jump-start our creativity through tight constraints like the X-prize idea, I expect that we'd end up combining the best of our current technology with the best of the new. Or maybe the worst of both...

It's more like complaining about license plate readers, cabin voice/video recorders, hackable cars, GPS tracking, and being increasingly dependent on a centralized service for repairs and maintenance. The "good old days" were actually different.

Yes, there are some benefits to all of those "features", but it seems prudent to consider and expose their risks as fervently as we're promoting the rewards. Perhaps the current balance isn't ideal over the long term, and we'll only see improvement by actively exploring the whole picture (whatever it is).

Have you considered that less control for the user is more control for someone else?

The more we do on computers that are opaque to us, and yet interconnected on the Wild West of the Internet with OS's with known vulnerabilities (and potentially backdoors) and privacy-hostile apps, the deeper and more detailed our activity and communications can be tracked. Are those good conditions for a healthy balance of power? Perhaps citizens of developing countries will face less-tolerant political climates than we're fortunate to have.

The long-term consequences of a lack of control over our smartphones, which as per the presentation deck we tend to have with us all the time and use very frequently, and also have tremendous sensor capacity, seem to be worth some thought.

I'd set it up as a behavioral mirror. Naturally I have a horse in this race :)

Inputs would be the actions you take in life (choices you've made) and how they relate to your goals (which depend on actions). Output would be an interactive visualization of past actions. The purpose would be to help improve the quality of your choices (actions you take) to better achieve your goals. Algorithms would work with the accumulated information to pull out correlations, patterns, and in general accelerate the process.

No Internet required, and so much greater security, privacy, and trust than usual. Which is freeing in many ways.

Smartphones and security don't really go hand-in-hand anyway.

However, I am encouraged by the fact that people paid many thousands of dollars in the '80's and 90's to purchase PC's. They were much harder to use, much less powerful, and yet because these PC's represented either a competitive advantage or a novelty, the price was paid. No Internet, but a heckuva lot more control and trust than now.

If a similar value proposition pops up again, somehow de-commodifying or replacing the standard PC design (smartphone or otherwise), I suspect we'll see a similar willingness to pay a premium.

The next X-prize should be for the best implementation of a smartphone OS/GUI where the "smartphone" has no I/O whatsoever except for the touchscreen and a charging/debug port.

If that couldn't crank up our creativity, nothing could.

Do you think that redecentralization will also depend on the public Internet or perhaps instead focus more on alternative networks?

I'm speaking of value, which I view as the balance of risk vs reward. Your view may differ. On one side, it seems to me that the reward or benefit we're receiving isn't anywhere near proportional to the power increases we've seen. On the other side, what are the risks of exposing so much personal information?

Receiving information seems much more valuable than giving it. After all, information has a price. If giving is not so great, shouldn't we be a little concerned about the consequences of trading away so much of it? Information asymmetry strikes me as the root of many power imbalances.

So, if we want more _value_ out of our "pocket supercomputers", it seems we have to consider the overall value proposition. Then perhaps we can figure out how to do better.

Those examples benefit their owners, providing competitive advantage on a national or corporate scale.

What sort of example might benefit you as an individual?

What if our information stayed private and we used our "supercomputers" to do all sorts of data-mining, or AI, or data-science on it? Same stuff the big guys are doing, but for ourselves. Smaller, higher quality datasets.

If our smartphones worked like this, and we'd benefit more from better information, then we'd also have a pretty good reason to develop advanced information-centric UIs and focus a bit more on security and trust.

So information has no value? Why do we say that information is power? I suppose that idea could be just as wrong as saying we have a supercomputer in our pocket, but the idea that information is power is much, much older.

Information isn't being hoovered up at the cost of billions of dollars without a very good reason for doing so.

Still, it seems a matter imagination. If our personal computers have tremendous (and accelerating) power, there's few actual limits to where we can take them. The deck even has a slide that talks about pulling components off the shelf and playing lego with them.

I think it's prudent to consider whether the average person is, all things considered, empowered by our Internet-connected computers. If so, great. If not, or not enough, then how?

"Everyone gets a pocket supercomputer" - Slide 8

I see this idea repeated so often, but it's unfortunate that we don't also have the _value_ of a supercomputer in our pocket. The sole purpose of a supercomputer is to advance the interests of its owner, who has exclusive control over it. Whether the purpose is prediction, or simulation, or to advance the state of the art, the benefit goes to the owner.

Yet we seem to have less and less control over our smart-phones. With so much information about each of us being siphoned off through the Internet, it's easy to wonder whose interest they serve.

With all of the advances in computing power, you'd think we'd put a bit more imagination into capturing more value for each individual smart-phone user, and less into centralized capture and analysis of our digital activity.

The author seems to ignore the degree of power Facebook, Zuckerberg, and friends gain from a vast quantity of detailed information about people's lives.

I can't imagine Zuckerberg would trade the unlimited upside (optionality) of real power and influence for stable, predictable income. Upside is why Facebook's market cap as of today is $230 billion.

Those are my observations, maybe I run in different circles than you do.

You seem to be saying there's only one way for technology to be - the way it already is - and that everyone had better get on board with it, or else.

I'm saying there's also another way for computers to be that lets everyone receive more value with less effort and, if they choose to take it that far, ease into the way technology is now. But they may not ever need to. This would accomplish the same goal you are espousing, but in a practical way rather than by decree.

I'm taking the beginning and moving it way back, to the point where an Uzbek villager seeing such a computer for the first time can use it as well as an expert within a day or two, because it operates on principles he's already familiar with and doesn't require classroom instruction.

As a SysAdmin, you have the keys to the kingdom, the world of technology is at your fingertips. You're in a position of power in a technological world. As a long-time software guy, the same is true for me. Yet our expertise only puts us on thick side of a power-law distribution; a fine place to be, but the vast majority of people have nowhere near our expertise and simply do not possess the comfort-level and skills to extract the value we do. Whether it is to pull down a nice salary or to change the world with software, at this point in time we are a minority.

Beginnings are, pretty much by definition, largely mysterious and chaotic, intimate, and disorganized. I have a tremendous respect for those qualities - without them we'd not have computers or even ourselves to talk to. We wouldn't exist.

Overall, in my view, the feminine represents the creative process - the ability to discover and establish new directions. It's the way to achieve any kind of change. Is there anything more valuable?

The idea though is that by making the beginning of a computing experience more closely align with where we actually are as individuals, the ramp-up is much less steep than it is now. I say it's steep because of my observations of people who are not long-time techies - just regular folk who are afraid of computers and abhor detail. People taking classes and using command lines are already well along their journey.

I agree that women would appreciate and especially benefit from a more balanced computing experience, but I also think it would help everyone else too, including techies like myself and many here.

It seems to me that modern computers are already severely polarized and it's limiting our views of what they can do for us.

Well, I just described exactly what it is and how it works, followed by a concrete example. So I'm not sure that's vague.

When the person taking the action is recording the action, and the information is kept private, then it's what I consider to be high quality information, even at very low levels of detail. Contrast this with the quality of information that ad networks get (it's lesser). Given that low-quality information the usual situation, I completely understand why you think that it takes a large amount of computing power to make useful predictions from it. I've found that the situation is greatly simplified when dealing with high quality information.

Too bad the post was flagged.

My intention was to spark a discussion on the high-level ideas, but it seems that many are more interested in the implementation. That's fair. The challenge in this case is that "what it is" matters less than "where it's going" or "why it exists".

In a nutshell: First choose a thread of action - narrow your view to one set of things you already do. Then when you act, you record that you took that action. Add more threads-of-action as you wish, organize as needed. That's the beginning of Benome. This pattern remains throughout as the fundamental.

This simple pattern allows Benome to become an interactive visualization that combines your past behaviour with probabilities for you future behaviour. A very simple example is that if your chosen thread of action was when you buy tomatoes, and then you choose to share that thread with someone who sells tomatoes, they'd know when you were ready to buy. But that same pattern can be used for many more applications than simple commerce. You can leverage a detailed awareness of your own behaviour into many personal advancements.