He's not wrong, but I've heard that complaint about the tech press in general for years, so I'm not sure its bringing anything new to the table on HN.
But at the same time, "some of these will surely turn out to be revolutionary eventually" isn't a great excuse to treat everything as revolutionary before it has a chance to even be seen in the field.
Here's a short and not very complete list of unsolved meta-problems:
Even when products and systems are revolutionary, there are unexpected negative consequences (e.g. FB and Cambridge Analytica)
All systems can be trolled and abused, and if they can be, they will be (e.g. fake reviews on Amazon etc.)
AI doesn't actually work all that well yet. (Neither Siri nor Alexa truly pass a conversational Turing test, which means there's a lot of guessing about whether or not any novel request will generate a useful response.)
IT systems and products of all kinds are brittle, unreliable, and often downright stupid. Users don't trust updates and feature changes, and often they're right to do so. Given that, why would AI "products" be any better or more reliable?
When we pass the Turing test it means we've got actual AI.
But I'm not sure there'll ever be a clear line. So Duplex kind of passes it in a very narrow context. Whether or not the person at the end of the line was actually fooled or not is a slightly different question. They could have just been humouring what they figured was a weird automated system.
But it's not like Google won't improve exponentially with this. They've now got a basic AI conversation system that they hope people will use and feed it data of actual conversations.
So Duplex v2 will have an expanded system where they can handle ten times the number of scenarios and questions.
The more I think about it, the more impressive it seems. Most attempts at a Turing test are text only where the subject is supposed to be a 13yo immigrant boy. Here Google's jumping straight to voice conversations.
Even with the constrained grammar, neither of them passes the test of reliably producing the same result for the same phrase under good conditions. If the error rate is well into the double digits for simple structured queries using a constrained vocabulary it seems like the Turing test is still pretty far off.
Domain-mismatch is a common problem: something is expanding rapidly, but in the end doesn't work on many or most niches. One size does NOT fit all.
An example is the no-sql movement. People started writing ALL new applications with it, thinking "it's the future". Turns out the regular RDBMS are still the better option 98%+ of the time. "Microservices" are a similar boondoggle. Make sure you really need it.
So instead of debating solutions to local needs with neighbors, we fetishize juicing Googles stock price and enabling them to build things that mathematicians already figured out the theory and limits to years ago
This is nonsense.
It’s one thing to theorize the Higgs and build an LHC
It’s another to build chat bots with our time so we can avoid menial scheduling.
Feed people misdirection about how the economy needs this cause we’re all too dumb to figure out our own lives (translation: elites lose power if we do that) and you keep getting these stupid generational pyramid schemes: we buy in now on some vague promise of a future payout we’ll not be around to collect on
Corporate is the church now. Love it or be ostricized.
Duplex isn't about scheduling or calendars, it's about robots conversing with humans. It's proof of concept, a stepping stone on the path to general artificial intelligence.
I hate talking on the phone. If I can't fill out a web form (login-free please) or send an email to complete some task like that, I'd love to have a robot do it for me. I don't see anything unethical about this as long as the robot performs the task reasonably well, doesn't waste the receptionist's time, etc....
The classic hacker ethic says it's bad for humans to be required to waste time on something a machine could do, and Google has just expanded the list of things machines can do. Of course, under that framework, it would be preferable to automate the receptionist taking the call as well.
I'm eager for this technology to become more general-purpose. For months, I've had a case open with an airline over an item that went missing from my checked bag. The only way to get updates is to call them on the phone. It first goes to an IVR, then a call center in India where the call is screened by a person, and only then to the actual department that can give me useful information. There is no direct line (I asked). The process is obviously designed to frustrate users so they give up. The hassle is arguably not worth the $90 they owe me, but I don't want to let them get away with it. I really wish I could have a robot talk to them.
The trick, I have found, is to treat the other person on the line, as well, a real person. When they say their name, say it back to them, ask them how their day is, chat for a second about how it's their Wednesday (halfway there!), what the weather like where they are (Broncos are going to have a tough game, eh?), etc. Even just 30 seconds of chatter will get you great service. So many people treat them as 'the help', so when you come in and treat them as an equal human, their day just got a little better, and they'll treat you better too, because you treated them well first.
And I mean you really get better service. That direct line to their manager's manager? You have that now, just because you were nice. That bouquet of flowers for your mom/wife? It just got a little larger, because you asked them how their day is going. That bill you are having trouble paying? It's 15% off, because you complained about the snow too. Yeah, it's not a lot. Yeah, it happens maybe 1/10 times. But it is worth 30 seconds.
Besides, you got to make an actual human's day better. No one is too busy for that, for good manners and a smile, even if it's over the phone.
Where I want a person is when I want to bypass processes completely or I have a big exception that warrants a human. So if I’m calling your call center, you may as well send me up to tier 3 or higher because I’m going to be a pain.
And given I did support for a couple years myself, sure I’ll try to make their day a bit better where I can and try to have all my info ready and to be as calm as possible. Because I know it can take a while, so I usually have an hour or two set aside for these calls and can just wait and not have to hurry anyone.
For actually resolving a customer service issue, sure I'd rather talk to a human. I do try to be nice to them, and it does, indeed often result in better service.
For getting an update on what's going on with the request I made three months ago, I want a website, or push notifications by email.
(Imagine a shopkeeper with a thick scottish accent). "We've got an opening for 4pm wednesday." "I'm sorry, I could not understand you. Please repeat that."
Beyond the “sharing” misnomer, those were proven concepts with existing industries. So far what's been proved is that if you're willing to pour billions of dollars into a company you can produce a better app than the incumbents, and that you can see short-term gains if you're willing to break the law and/or subsidize heavily.
> if you showed google maps to someone from 2000, they would think it is magic.
By 2000 the most likely reaction would have been “oh, it's like MapQuest but faster”. Google Maps reflects a lot of evolutionary improvements but the big change was the rise of the advertising model meaning that you didn't have to pay a third-party for GIS software or a subscription service to get annual map updates.
> By 2000 the most likely reaction would have been “oh, it's like MapQuest but faster”.
Depends on context. Sure, maybe if you showed it to them on desktop with basic operations (e.g., turned off traffic display, etc.)
Show using it on mobile for navigation with “Ok, Google, navigate to...” with turn-by-turn navigation and real-time, traffic-based route adjustment with voice promoted and confirmation, and it's at least as far beyond 2000s MapQuest as that version of MapQuest is above a dead-tree book of maps.
Sometimes the consequences take a while to work through. What will Amazon do after it kills bricks and mortar? How many customers will it have when most of the population is barely scraping by with gig economy now-you-see-it-now-you-don't income?
https://www.epi.org/press/uber-drivers-earn-the-equivalent-o...
So, not a win really. At least not right now.
Look at all the money we spend on medical/pharma research. Most of it doesn't pan out, but we're basically ok with that since the rewards are pretty large when it does.
Unlicensed minicabs and illegal sublets had already been invented. Putting an app on them is not exactly innovative.
When my home is smart enough to cook all my meals and do dishes and fold laundry, then I'm ready to buy into it.
- When I'm cooking and my hands are dirty and I want to set a timer
- When I'm cooking and my hands are dirty and I want to change music/lower volume/raise volume/call someone (or do a measurement conversion)
- Turning off all the lights in my house when I'm already in bed and forgot whether I left some on
- Laying in bed with the gf and asking random questions out of curiosity
- Asking how long it will take to get somewhere if my gf and I are pondering some place to eat but are at the dinner table without our phones
- Get home after having a few beers and dive into the messy chicken wings with both my hands, and then realize I want to turn on the TV, play a show or change the current one on my Shield.. perfect use for the GH
- Since I already look at my phone way too much (potentially straining muscles in my neck), I can leave my phone in another room and still get quick answers to deep life questions without tilting my neck downwards
They're all __super__ minor of course, but honestly, I'm so used to the convenience now it actually feels weird to have to do some of these things 'manually'. I do understand there are serious concerns about privacy that may or may not offset the benefits, definitely plenty of room for discussion on that.
But, I see how useful it can be when your hands are busy! My home has an ancient early 2000 smart home installed, which can be updated at a pretty big price, so I don't bother. I guess I am just used to my "inconvenience."
Yes, assuming you're already sitting at your desk, with your desk calculator in hand.
Also, this might be a generational gap issue here, but desk calculator? Not smartphone? (And if I'm at my desk, my goto is usually "ipython<ENTER>x+y<ENTER>" anyway.)
I seriously doubt that even if we still lived in a world where normal people actually owned calculators. This is just "No wireless. Less space than a nomad." again.
And just because I don't particularly enjoy IOT of Wireless Headphones, does not mean that I am a nomad, mate!
1. Dec'15 SpaceX performs first ever orbital-class rocket landing on land.
2. Apr'16 SpaceX performs first ever orbital-class rocket landing on a ship.
3. Mar'17 SpaceX reuses first orbital-class rocket.
4. Feb'18 SpaceX launches Falcon Heavy, the highest payload capacity of any currently operational launch vehicle.
5. May'18 SpaceX launches Falcon 9 Block 5, set to give USA direct human access to space (in late 2018) for the first time since Mar'11.
Assuming positive intent rather than sarcasm, so here's why they're game changers:
1.2.3. landing and reusing orbital rockets will deliver a huge increase in launch frequency and a huge reduction in cost, both acting to increase humanity's practical access to space. It's the difference between building a plane that flies once and building a plane that files a hundred times.
4. Falcon Heavy allows us to launch higher payload missions that was possible before. Especially since nobody assembles spacecraft in orbit, this raises the cap on how massive a spacecraft humanity can operate.
5. Giving America the ability to put humans in space without relying on the Russian Soyuz is a political and strategic victory for the US.
On a lighter note: we've put people on the Moon before we've put wheels on the luggage[2]. Exploration before comforts.
[1] http://www.lettersofnote.com/2012/08/why-explore-space.html
[2] https://betafactory.com/what-came-first-wheeled-luggage-or-a...
Also, the SV reality distortion field hasn't infected all the tech industry. In the companies I worked, nobody was convinced we were changing the world, we're just writing some tools for cash.
The senior editor...
Most of the things that has big impacts are actually things we skip through. For example, we have a pretty decent camera in our pockets at all times. We take pictures constantly. The "selfie" revolution, one that has big impacts in how we perceive ourselves and want other to perceive us is enabled by having an small quality camera around. At some point passing from "terrible quality" to "good enough quality" made a jump in that. Sure, not everyone takes selfies. But it has been a big change in photography in the last 10 years or so. So big we have a new name for something it exist before.
I'm pretty sure that being able to track a run and share it on social networks is a big motivator to do exercise, for example.
I seems weird to me talking to AI, and even weirder that "the future" is calling to a shop with an AI instead of using some sort of web shop/app/whatever to get an appointment.
But who knows...
For those of us that keep a skeptical mindset when evaluating new technology, none of this comes as a surprise.
Those are harder to approach and solve, and you need more patience for results. Now we see complaining that the nifty gadgets aren't coming like they used to. Hmm.
For example, Housing isn't a problem, it's an investment opportunity. The people donating to your campaign wouldn't like it if their investments were pushed down in value.
An app can't fix your economy being dominated by bad actors.
crime- Is strongly correlated with poverty, which has been rapidly reduced globally thanks to technological innovation
Housing- Will be solved by tech making remote work viable and thus reducing demand in cities. Automation and some form of income stipend could also allow people to live in more affordable places rather than cities.
Your view of "tech" is pretty stunted if you only think web apps.
Let’s ignore the shattering impact of incredibly convenient video streaming, content filtering via image tagging, electric cars, transportation or food on demand, having contactless payment available to anyone for pennies…
MTurk pays offensively little, sure, but thanks to ‘Silicon Valley’ anyone can work now, no matter how much discrimination or how remote they are — and buy groceries with it. If the accusation is “building”, Amazon built it. If you want it to be more lucrative, you can either legally enforce minimum wage, or start tasks that pay more.
Shelter, like food, is a problem we solved a while ago, as long as you can afford it: inequality is a problem, but you can’t blame AirBnB, SpareRoom and countless others for allowing people to find others with a lifestyle, a schedule that fit them so that they can save on rent. Let’s not talk about how internet empowered people to compare mortgages.
Crime: Well, yes, Silicon Valley has done a lot for Law enforcement, but expectedly and Thank God, not as a B2C business model.
I’m honestly more confused arguing that article than I would argue with a flat-earther. At least a flat-earther is probably right when they say that you have not personally checked yourself for what you claim.
Are you being ironic? The steam engine literally reshaped the world. I think you express exactly what the author is talking about: Acting as if just by announcing it, it has already succeeded.
I've used Evernote since 2012 - when I started it really wasn't that useful. But now, after over 6 years of using it daily, I have a huge body of content personal to me which I find tremendously useful. This only came about because I didn't jump to the next shiny alternative.
Another thing he criticised was the Internet of Things and other gadgets: I agree with the author on that one. I hope it dies a quick death. If there's a tradeoff between an internet-connected device that lasts 1-2 years and a not-connected device lasting 20 years, I would always chose the longer-lasting one.
Currently most companies want to be Facebook or Google and collect all the data they possibly can while putting out a mediocre/bad product with no security that only works when there is an internet. No consumer actually wants this.
Instead of killing the industry, we need a new generation of companies that put out products that function the same or ideally better than their analog competitors. Data collection, such as usage information, and internet connectivity will be what sets these products apart from their traditional competitors. However, the emphasis should be on sending that data to the owner of the product, instead of sending it to the cloud for the companies benefit. Optimizing energy usage, learning usage patterns, etc will be great differentiators. My products working together will be an even bigger differentiator
Aggregation of that data can be useful, and could be opt in. I'd love to know that I seem to use way more energy than other people around me. Perhaps I have a bad AC, or my insulation is particularly bad.
The biggest issue right now is that companies are putting out insecure products that are completely broken if the parent company of the product are left out of the loop. The products need to work locally, with no reliance on the parent company, and only rely on a company server for true enhancements that can not be calculated locally.
Or perhaps you are running too many always-on IoT devices :-) SCNR
These things always look inevitable in hindsight. But it takes some real cleverness to get them in foresight.
- Nikola Tesla, 1926.
https://kottke.org/18/04/nikola-tesla-predicted-the-smartpho...
Useless, all that technology expressing itself in making glowing fish. The tech behind it however is incredibly useful. But wow, the existential dread watching them swim around under a black light was considerable.
Or on the other shelves, looking and seeing dog or cat "calming collars" that use pheromones tailored specifically for individual species. I mean, its funny at how useless and useful it is at the same time. Is your dog a worrywart? Use pheromones to make him stress free! But you wonder some day if some bright tech baron will start to make human calming collars, or bracelets, or what have you...
Maybe not big tech per se, but the writer of this article might have a failure of imagination if they only see the uselessness.
I do think though that AI is the last "fundamental" change information technologies will bring though. From here out, the future is in the convergence of other hard sciences with advanced IT. Each "Big thing" has been a fundamental building block. By themselves small, but they get significant as they combine.
There are wast continents of mostly unexplored or abandoned territory in IT. Moreover, some of them are well documented. Some of them were documented in 1970s. The fact that IT professionals today don't know about them and can't imagine anything more significant than "AI" for pizza ordering, is unsettling. Just to name a few things Alan Kay often talks about:
- Automated system integration or tools for generalized interconnection of applications available to end users.
- WYSIWYG for the Web.
- Constraint-based problem-solvers accessible to mere mortals.
- Agent oriented programming with UI simple enough for normal people to use.
- Dynamic simulations in "normal" software (not CADs)
I am absolutely sickened by the fact that the most common response to this list is "nobody needs this". As far as I'm concerned, each of those things is needed orders of magnitude more than a haircut-scheduling AI.
For your example, automated system integration is the combination of existing technologies. That's mostly what I'm implying. We have breadth, now the work is on the depth.
Speech synthesis has improved, but are you saying that Duplex is a massive step forward in speech recognition and natural language processing too?
Also imagine it opened up on Google Cloud, in a product similar to AWS Connect.
Finally, imagine this: They've always had a huge amount of data for human voices to train their AI models. Here's one more; being able to call anyone on the planet and get them talking in a way that is natural, then using the data on the other end of the phone to further train the accuracy of the voices they use. Kind of like ReCaptcha, but for voices.
And dear God, spam calls are already bad enough, why do we want to make them even more realistic? I fail to see even one application where I would prefer to talk to a more realistic robot.
I can also google up information about light-bulbs when I'm standing at the store wondering if I should get X or Y. Or if some product doesn't specify if it's gluten free or not and the ingredients look a little iffy.
There are plenty of tiny improvements. It doesn't revolutionize my life, but having my old Nokia would definitely be a downgrade.
Sure, I didn't have a fancy app to look at the timetable for the bus, but there was a wap-page I could use and frankly the apps are about as frustrating but now because touch interactions so often go wrong without real buttons.
And I use both vim and ssh daily, rarely wishing I had an new app to replace them with.
I don't think we'll see another time period like that. But that's not really a bad thing. You can only discover really important things once and then they eventually become part of everyday life. A more interesting question is how to sustain our standard of living when we can no longer expect high economic growth from new innovations.
Generally I consider something ethically questionable when a person is harmed in some way. Can someone explain how a receptionist might be harmed by unknowingly booking an appointment with a robot instead of a person? The real person shows up. The business gets an appointment. The receptionist does their job. Where does the ethical question come in?
Context matters. Deception, in and of itself, is not inherently harmful.
It makes literally 0 difference in the life of the person booking the appointment if it's a robot or a human telling them the desired time and date.
High flyers = people that own an Android phone and get haircuts.
Although we won't ever have a time machine that travels backwards in time, there are a lot of commonplace things that were once science fiction.
Tesla electric cars and electric car infrastructure
The sharing economy where we rent instead of buy things hopefully will stop or at least slow our rabid consumption
These are huge changes in the way we will live our lives.
Maybe that's how the "Darmok and Jalad" language came about.
Absolutely irrelevant to the topic. I already ignore the hype, but I think I'll also ignore this author's hyperbole.
Don’t read the article.
http://www.subzin.com/quotes/M106664b34/Erik+the+Viking/Hy-B....
On a non-sarcastic note the toy that’s barely usable by hobbyists is the early adopter segment, after the innovators. There’s a reasonable explanation at businessdictionary.
http://www.businessdictionary.com/definition/early-adopters....
