1 'I'VE SEEN THE FUTURE OF CONSUMER AI, AND IT DOESN'T HAVE ONE'
4 Thursday September 06, 2018 @11:30PM (msmash)
5 from the there-is-no-spoon dept.
7 o Reference: 0102639856
8 o News link: https://tech.slashdot.org/story/18/09/06/205221/ive-seen-the-future-of-consumer-ai-and-it-doesnt-have-one
9 o Source link: https://www.theregister.co.uk/2018/09/05/consumer_ai_ifa_2018_roundup/
12 Andrew Orlowski of The Register recounts all the gadgets
13 supercharged with AI that he came across at IFA tradeshow last
14 week -- and wonders what value AI brought to the table. He
15 writes: I didn't see a blockchain toothbrush at IFA in Berlin
16 last week, but I'm sure there was one lurking about somewhere.
17 With 30 vast halls to cover, I didn't look too hard for it.
18 But I did see many things almost as tragic that no one could
19 miss -- AI being squeezed into almost every conceivable bit of
20 consumer electronics. But none were convincing. If ever there
21 was a solution looking for a problem, it's ramming AI into
22 gadgets to show of a company's machine learning prowess. For
23 the consumer it adds unreliability, cost and complexity, and
24 the annoyance of being prompted. [...] Back to LG, which takes
25 2018's prize for sticking AI into a superfluous gadget. The
26 centrepiece of its AI efforts this year is a robot, ClOi. Put
27 Google Assistant or Alexa on wheels, and you have ClOi. I
28 asked the booth person what exactly ClOi could do to be told
29 "it can take notes for your shopping list." Why wasn't this
30 miracle of the Fourth Industrial Revolution let loose on the
31 LG floor? I wondered -- a question answered by this account of
32 ClOi's debut at CES in January. Clearly things haven't
33 improved much -- this robot buddy was kept indoors.
36 ** Stupid industry fads (Score:5, Funny)
37 (by Spy Handler ( 822350 ))
40 3D printer in every home will fundamentally change human society
41 IoT internet connected belt buckles and toothbrushes will take
43 AI will revolutionize consumer electronics
44 Net PC from Sun will dominate the computer industry (this one is
47 ** Re:Stupid industry fads (Score:5, Insightful)
48 (by ShanghaiBill ( 739463 ))
51 Excessive hype is always followed by a trough of
52 disillusionment. But as the TOD fades, plenty of mature,
53 practical applications are likely to emerge. The
54 technological naysayers are usually even more wrong than the
56 [1]Hype cycle [wikipedia.org]
61 [1] https://en.wikipedia.org/wiki/Hype_cycle
64 (by fahrbot-bot ( 874524 ))
67 > Excessive hype is always followed by a trough of
69 Pro Tip: Get out in front and mention this *before* taking
70 your date home. Better for her to hear it from you than
71 her working it out on her own ... :-)
74 ** Re:Stupid industry fads (Score:4, Insightful)
75 (by CaptainDork ( 3678879 ))
78 If smart phones and tablets are any indicator ...
79 AI, too, is an evolutionary dead end.
80 It's a buzz word with a vacuous definition.
83 (by Q-Hack! ( 37846 ))
86 Not a lot different than back in the 1950's when the
87 trend was to create all manor of odd gadgets to make
88 life easier. Those deemed useful are still around...
89 The rest can be found in junk markets around the world.
90 But hey, the Cracker-barrel's of the future will still
91 need stuff to decorate their walls with.
94 (by CaptainDork ( 3678879 ))
97 In reaction to your sig:
98 I recently re-read "Nineteen Eighty-Four," because
99 my first reading was so long ago.
100 Good read, but what a goddam depressing book!
106 (by magzteel ( 5013587 ))
109 > Excessive hype is always followed by a trough of
110 > disillusionment. But as the TOD fades, plenty of mature,
111 > practical applications are likely to emerge. The
112 > technological naysayers are usually even more wrong than
114 > [1]Hype cycle [wikipedia.org]
115 Back in the early PC days, when you had to hook up a
116 cassette player to load your application, and then another
117 one to load your data, we used to tell people they could
118 store recipes on their TRS-80 personal computer. This was
119 not much of a productivity enhancer. I'm sure based on
120 this experience some people would have thought PC's were
121 useless and had no future.
122 And then floppy disks and spreadsheets were invented.
127 [1] https://en.wikipedia.org/wiki/Hype_cycle
131 (by jythie ( 914043 ))
134 It is really difficult to say if the naysayers or
135 hypesters are more often right or wrong. One problem with
136 looking back at negative guesses is we only really
137 remember the ones that turned out to be wrong since the
138 evidence is in modern use today, while all the naysayers
139 that we right, well, the things they were right about
140 faded into obscurity.
143 (by ShanghaiBill ( 739463 ))
146 You only count as a "true" naysayer if you are negative
147 about an overhyped trend with groupies and fanbois, not
148 about an obviously stupid idea.
149 The naysayers were right about the Segway, but that was
150 an easy target, since it reached peak hype before it
151 had even been shown to the public.
152 Other tech failures were Iridium, Zune, Pebble,
153 Juicero. But none of these were hyped as world changing
161 (by ShanghaiBill ( 739463 ))
164 > In the long term only 1/20 companies really make it.
165 Success of a technology is rarely correlated with the
166 success of particular companies. Silicon Valley is
167 littered with plaques marking the graves of
168 semiconductor pioneering companies. Few of them
169 survived. Yet semiconductors have been the greatest
170 technological success since fire was tamed.
171 For another example, look at aviation. It took 66 years
172 to go from Kitty Hawk to the Sea of Tranquility. Yet
173 how many airlines made money during those years? Almost
180 (by atherophage ( 2481624 ))
183 Prognosticators have been wrong before. While it is easy to
184 poke fun at the unusual who knows, perhaps in a few years
185 dental floss will come with AI. The thought of not having AI
186 floss will be unthinkable.
189 ** Re: (Score:3, Interesting)
190 (by Anonymous Coward)
193 As much as I am a nerd, I blame "nerds" for this. There is
194 this whole new fad of being a "techie", watching Big Bang
195 Theory, owning a Tesla, and generally being absolutely
196 ignorant about real science, technology and math while
197 "pretending" to be a nerd. I used "pretending" but there may
198 be some legitimate attempt but it is hard to tell if someone
199 is a fake nerd or just a stupid nerd. I think this trend
200 partly follows from women trying to follow the (tech) money
201 and then men trying to follow the women.
205 (by AbRASiON ( 589899 ) *)
208 I don't know that there's a lot of these people but they
209 do exist, for certain yes. The 'watching big bang theory'
210 is the kicker, once someone admits watching that, you know
211 they're very unlikely to be a 'proper nerd' for lack of a
213 Considering they only have partial skills in technology
214 then, we can likely guess, if they work in the industry,
215 they're probably higher on the ladder than us and paid
216 more though :/ like most management / consultant types.
221 (by JMJimmy ( 2036122 ))
224 The thing no one can consider is time.
225 "AI" being jammed into things now is probably lame, awkward,
226 and of very limited use. Much like computers were back in the
227 punch card days with devices that. Less than 100 years later
228 we've got computers in our pocket. We are in the early days
229 of AI - we'll look back on it decades from now as we do with
230 things like: [1]https://www.youtube.com/watch?...
232 This article is just another example of someone who can't see
233 past their nose to the road ahead and the million differen
238 [1] https://www.youtube.com/watch?v=Sp7MHZY2ADI
242 (by AHuxley ( 892839 ))
245 Good for a few workers over the decade of hype.
249 (by m00sh ( 2538182 ))
252 > 3D printer in every home will fundamentally change human
254 > IoT internet connected belt buckles and toothbrushes will
255 > take over the world
256 > AI will revolutionize consumer electronics
257 > Net PC from Sun will dominate the computer industry (this one
259 I don't know about home but it plays a big part in
260 manufacturing. There are very specialized and successful
261 medical companies that use 3d printing.
262 Don't know about belt buckles but fitbit, apple watch, garmin
263 has been worth billions of dollars and fundamentally changed
264 the way a lot of people do things.
265 I don't know about NetPC but what about the cloud? The hype
266 that we would all put all our stuff in the cloud blah blah
267 actually materialized. There are many companies who own no
268 hardware except the dev la
272 (by lokedhs ( 672255 ))
275 Net PC was not from Sun. I should I know, I worked for them
276 during that era. What they had was JavaStation, which was a
277 neat idea but ahead of its time. That concept is now realised
278 by the Chromebook. Net PC was a Compaq thing, if I recall
279 correctly. However, Wikipedia tells me it was Oracle, so
280 perhaps the Compaq device was called something else.
286 (by bobbied ( 2522392 ))
289 > Example: Apple will go under...any day now....since 1984
290 But they've been totally correct in not predicting the "Year
291 of the Linux Desktop" has come.
292 You win a few and lose a few.
295 (by jythie ( 914043 ))
298 *nod* to expand on this.... true, Apple never did go
299 under. But look how many computer companies started up
300 around the same time and did. It is fun to look at the
301 successes and compare them to the naysayers who were
302 wrong, but the ones who were right, well, their
303 predictions did not leave much to talk about today.
309 ** Re:Now With AI! (Score:5, Informative)
310 (by ShanghaiBill ( 739463 ))
313 > Gee, I could have sworn we already HAD the AI craze back in
314 > the late 80s. Or was it early 90s?
315 It was the 1980s. It had faded long before 1990.
316 But there was an earlier AI craze in the 1960s, based on
317 perceptrons. That faded by 1970.
318 The 1980 AI hype cycle was driven by "expert systems" and
320 The latest cycle started in 2006 with the publication of the
321 [1]seminal paper on deep learning [sciencemag.org], and has
322 so far lasted far longer than any previous AI hype cycle.
327 [1] http://science.sciencemag.org/content/313/5786/504
330 (by CaptainDork ( 3678879 ))
334 AI had an unambiguous definition that eroded under stress
335 because the industry came to the realization that the "I"
336 part (intelligence) used the human mind as the high bar.
337 The second epiphany came when no one could fabricate an AI
338 that would simply refuse to cooperate if Facebook was
343 (by Pseudonym ( 62607 ))
346 In the 90s it was all "knowledge-based systems" and in the
347 noughties it was all "intelligent agents".
350 (by ShanghaiBill ( 739463 ))
353 > In the 90s it was all "knowledge-based systems" and in
354 > the noughties it was all "intelligent agents".
355 Yes, but those generated far less hype than what
356 happened in the 60s, 80s, and teenies.
357 The big things in the 90s and noughties were the web
363 (by CWCheese ( 729272 ))
366 Thanks for the Lisp reference! I fondly remember learning
367 Lisp in an AI class during college in the 80s. Actually
368 enjoyed programming Lisp because it could be so terse and
369 do so much very rapidly. However, we really had no good
370 applications to use for it, other than having an
371 application learn the best way to win a chess game. I
372 chose not to pursue AI as a career and haven't suffered
376 (by ShanghaiBill ( 739463 ))
379 > I chose not to pursue AI as a career and haven't
381 Learning Lisp would not have helped you. Modern AI uses
382 mostly Python based libraries such as Tensorflow and
383 PyTorch. C++ is used for performance critical stuff.
384 Nobody uses Lisp for AI anymore. It was a dead end.
388 ** Fifth Generation (Score:2)
389 (by mcswell ( 1102107 ))
392 Fueling the hype in the 1980s AI cycle was the Japanese
393 Fifth Generation project, for which a stated goal was to
394 leapfrog the West's computer technology and skills. People
395 like Edward Feigenbaum and Pamela McCorduck used the FUD
396 generated around this project to call for increased
397 funding, claiming in their 1983 book 'The Fifth
398 Generation: Japan’s Computer Challenge to the World' that
399 "America needs a national plan of action, a kind of space
400 shuttle program for the knowledge systems of the future."
406 (by Torodung ( 31985 ))
409 > Cats on the blockchain, anyone?
410 Well, at the very least, every zig should be on the
411 blockchain. Don't know about Cats.
414 ** You can stop reading at "Orlowski" (Score:4, Interesting)
415 (by serviscope_minor ( 664417 ))
418 Andrew Orlowski of The Register is basically a professional
419 dickhead. His main goal seems to be to be as obnoxious and
420 ignorant as possible presumably with the goal of trolling the
421 readership. He's pretty much the reason I stopped reading the
422 Register because of the constant streem of utter bullshit from
425 ** Re:You can stop reading at "Orlowski" (Score:4, Interesting)
426 (by starless ( 60879 ))
429 And also appears to be climate change denier....
430 (at least for some of his Register articles.)
434 (by TJ_Phazerhacki ( 520002 ))
437 So, Walt Mossberg for a new generation? Shutup!
440 ** AI in a Toaster! (Score:2)
444 Red Dwarf has already shown why this is a BAD Idea.
445 https://www.youtube.com/watch?v=lhnN4eUiei4
448 (by Revek ( 133289 ))
451 Please learn basic html K, thanks.
454 (by AbRASiON ( 589899 ) *)
457 How about slashdot stop being entirely backwards with that
465 (by Applehu Akbar ( 2968043 ))
468 But I do like being able to verbally ask my phone to navigate
469 to a contact, without having to squint at a screen in the
470 sun, and get turn by turn directions. Digital assistants have
471 slipped into a place in my life where they do a few useful
472 things. As time goes on, this set will grow larger.
473 But I know: "If it works, it's not AI!" "If it's AI, it won't
478 (by Anonymous Coward)
481 If Sony's Aibo lives up to the demos I have seen - that would be
482 one big application. AI as a pet.
483 I also use AI (maybe more ML) all the time with photo sorting,
484 image recognition, etc. It is already in the home.
486 ** OP must be joking... (Score:4, Insightful)
487 (by JoeDuncan ( 874519 ))
490 ... because consumer AI is *ALREADY* ubiquitous and all around
492 From the face detection in your phone, to the fuzzy logic
493 controllers in washing machines, to the ant colony algorithms
494 being used to route network traffic, to finding directions with
495 google maps, to Netflix and Amazon's recommendation algorithms,
496 to OCR for cheques and mail, to NEST thermostats, to robot
497 vacuum cleaners and lawn mowers, to expert systems in medical
498 diagnosis... (I could keep going)
499 AI in consumer products is literally *already* ALL around us.
500 Saying that consumer AI "has no future" is like looking around
501 at the world today and saying "personal cars have no future" -
502 it's completely idiotic because to anyone with half an ounce of
503 perception that future is ALREADY here.
504 It's like looking at a forest and claiming there are no trees
507 (by grahamsz ( 150076 ))
510 Yeah it seems like it is a natural fit in optimizing the
512 Even though I don't routinely use my phone as an alarm clock,
513 it still knows when i'm likely to get up and if I plug it in
514 at bed time it'll do a good job of figuring out when i'm
515 likely to get up and adjusts its charging rate to be done
516 about an hour before then. Yet if I plug it in a 3pm then
517 it'll assume i want as much charge as possible and charge as
518 fast as it can. It's not rocket science, but it's useful.
519 Do I need a dishwasher with
522 (by JoeDuncan ( 874519 ))
525 > Do I need a dishwasher with a screen that I can talk to?
526 Nope, but I'm willing to bet it has an embedded fuzzy
527 logic controller in it to control water levels.
531 (by mcswell ( 1102107 ))
534 "Do I need a dishwasher with a screen that I can talk to?"
535 Printers have a screen. You can't talk to it (at least
536 you're not supposed to--when aggravated, I've been know to
537 do so, and not kindly). But try to decipher what's on that
538 screen. I claim that printers are not any easier to use
539 than they were in 1984 (which is when I got my first dot
540 matrix printer). You (ok, I) *still* can't figure out
541 what's wrong with them, despite the screen.
548 (by JoeDuncan ( 874519 ))
551 > None of the tings you mention actually contain any real
552 > artificial intelligence in the sense of being able to
553 > making decisions in the face of unknown circumstances and
556 Roombas have to be able to adapt to unknown obstacles and
557 uncertain sensory input (could get blocked, partially
559 Embedded fuzzy logic controllers (also used in anti-lock
560 brakes) have to be able to maintain a steady output signal
561 given uncertain input (wear and tear on the mechanics,
562 grit...) that can vary wildly in an unknown manner.
563 OCR systems need to be able to tell the difference between
564 a cheque and unknown things, like night club flyers, and
565 they deal with hand written
570 (by AHuxley ( 892839 ))
573 Re "face detection" is not AI. Its a really big and fast
574 database. Filled with faces the police know about and random
575 people walking past CCTV.
576 Re "fuzzy logic controllers in washing machines" A set amount
577 of power, water, weight of laundry is not AI. Just good
578 programming within set limits.
579 Re "'finding directions" with maps that are created and set.
580 Re "recommendation algorithms" that is set by past people
581 buying things and another person showing the same interests.
582 More to do with collecting lots
585 (by JoeDuncan ( 874519 ))
588 > ..."face detection" is not AI. Its a really big and fast
589 > database. Filled with faces the police know...
590 ...and just HOW do the faces "police know" get matched to
591 this database? Explain without reference to AI.
592 > ..."fuzzy logic controllers in washing machines" A set
593 > amount of power, water, weight of laundry is not AI.
594 No it isn't, but you're a fool if you think your washing
595 machine is that simple these days. It DOES take fuzzy
596 logic to adapt to things like wear and tear on the
597 machine, arbitrarily changing water pressures and
598 temperatures, etc... and still maintain consistent
600 > "'finding directions" with maps that are created and set.
601 ...and using AI algorithms to find the best path.
602 Blah blah blah... you get the point. You've deliberately
603 downplayed the AI aspect
610 (by JoeDuncan ( 874519 ))
613 You are clearly uneducated, Troll.
614 If you actually wish to enlighten yourself, I'd start
615 here: [1]https://en.wikipedia.org/wiki/... [wikipedia.org]
620 [1] https://en.wikipedia.org/wiki/Artificial_intelligence
625 (by hazem ( 472289 ))
628 > From the face detection in your phone, to the fuzzy logic
629 > controllers in washing machines, to the ant colony algorithms
630 > being used to route network traffic, to finding directions
631 > with google maps, to Netflix and Amazon's recommendation
632 > algorithms, to OCR for cheques and mail, to NEST thermostats,
633 > to robot vacuum cleaners and lawn mowers, to expert systems
634 > in medical diagnosis... (I could keep going)
635 When I took an AI class a few years ago, one of my favorite
636 things the professor said was, "What we called 'AI' yesterday
637 is simply the algorithm for how we do a thing today."
640 ** AI's Strength (Score:2)
641 (by thePsychologist ( 1062886 ))
644 AI (i.e. machine learning/neural networks) is really good at
645 optimizing stuff, so its natural strength shows when you have
646 hundreds of thousands of entities in a system. Examples are the
647 electricity grid, playing Go, and a department store's
649 In our individual lives, AI seems more like another drop in the
650 bucket of too much technology, and I think one day we'll realize
651 that less is more when it comes to the stuff in our homes.
653 ** Getting concerned myself (Score:1)
654 (by SuperKendall ( 25149 ))
657 I was looking at new fridges recently as a friend was asking for
658 a recommendation, and it's alarming how trying to find a fridge
659 without a screen is getting to be like trying to find a cell
660 phone without a camera... it really limits your options.
661 The only way they could make fridges any worse is the if screens
662 also played CNN constantly when not in use, like in an
663 airport... you can absolutely see subsidized ad-fridges coming
670 Seems like only the highest and lowest-end fridges lack
671 screens these days (as well as ice/water in the door,
672 something else I could do without).
675 (by mcswell ( 1102107 ))
678 Come to my house. The refr *has* an ice/water dispenser in
679 the door, but it hasn't worked for over a year. I think
680 the tube to the water dispenser is frozen, and if it gets
681 thawed, it just freezes up again. Same with the water
682 dispenser on the refr nearest my office at work.
683 As for the ice dispenser on our refr, we never used it, so
684 I took it out and got lots more room in the freezer. If we
685 want ice cubes, we make them in trays, like the 1960s.
692 (by sheramil ( 921315 ))
695 > What the hell does a fridge need a screen for?
696 You can connect it to a webcam inside the fridge and see
697 if the light goes out when you close the door.
701 ** I remember a time... (Score:2)
702 (by Dallas May ( 4891515 ))
705 My uncle was a computer scientist for a National Lab. He retired
706 15 or so years ago. I remember just after my grandmother first
707 got internet, he didn't have it at his home yet because he
708 didn't believe it was safe -this was probably 1997 or 98, and I
709 remember him talking to me about how disappointed he was with
710 the internet. "It was supposed to be this great thing. It's
711 useless. It'll never amount to anything."
715 (by bobbied ( 2522392 ))
718 > My uncle was a computer scientist for a National Lab. He
719 > retired 15 or so years ago. I remember just after my
720 > grandmother first got internet, he didn't have it at his home
721 > yet because he didn't believe it was safe -this was probably
722 > 1997 or 98, and I remember him talking to me about how
723 > disappointed he was with the internet. "It was supposed to be
724 > this great thing. It's useless. It'll never amount to
726 > Yeah, he was wrong.
727 Was he? Was he really?
728 How much of the internet is truly useful and how much is just
729 trash? Judging by my inbox, the number of E-mail in my inbox
730 the ratio 1s more than 10 to 1 SPAM to worth while messages
731 (And that's AFTER the SPAM filters.)
732 I find that this ratio pretty much governs the whole of the
733 internet.. Where 1/10th of it is actually something of use
734 and the rest is just useless junk.
735 So he's not that wrong.
741 (by mcswell ( 1102107 ))
744 And here you (and I) are.
748 ** I heard... (Score:1)
749 (by Hentai007 ( 188457 ))
752 AI is turning frogs gay.
755 (by AbRASiON ( 589899 ) *)
758 That's actually not true, the frogs are only gay for pay.
761 ** Nobody buys something because of AI (Score:3)
762 (by Laxator2 ( 973549 ))
765 I did not see any example where someone says: "I did not buy
766 that product because it lacked AI".
767 I did not hear from anyone that they need AI so they are going
768 out of their way to buy it. In its current form AI is good for
769 pattern recognition in some cases, for example, face
770 identification in photos.
771 The only customers are corporations with massive collections of
772 personal data to analyze, but not individual consumers.
773 I believe AI has been over-hyped and pushed in areas where it is
774 not usable in its current form (like self-driving cars) and we
775 start to see the backlash.
776 I've already seen stories saying that the medical diagnoses made
777 by IBM's Watson are just plain wrong. More examples will follow.
780 (by m00sh ( 2538182 ))
783 > I did not see any example where someone says: "I did not buy
784 > that product because it lacked AI".
785 > I did not hear from anyone that they need AI so they are
786 > going out of their way to buy it. In its current form AI is
787 > good for pattern recognition in some cases, for example, face
788 > identification in photos. The only customers are corporations
789 > with massive collections of personal data to analyze, but not
790 > individual consumers. I believe AI has been over-hyped and
791 > pushed in areas where it is not usable in its current form
792 > (like self-driving cars) and we start to see the backlash.
793 > I've already seen stories saying that the medical diagnoses
794 > made by IBM's Watson are just plain wrong. More examples will
796 What about Google home and Alexa?
797 How do you recognize pedestrians in self-driving cars without
799 IBM Watson was wrong quite a bit but it won jeopardy.
802 ** First they ignore you, ... (Score:1)
803 (by Maxo-Texas ( 864189 ))
806 First they ignore you, then they laugh at you, then they fight
809 This field is moving so fast compared to the 90s.
811 ** It's a dead end because it's not very good anyway (Score:2)
812 (by Rick Schumann ( 4662797 ))
815 So-called 'AI' is over-hyped and under-performing.
817 ** Another AI winter? (Score:2)
818 (by OneHundredAndTen ( 1523865 ))
821 The AI bubble seems to be starting to deflate. It may not pop,
822 but it will likely carry on shrinking. Most people already know
823 that Alex and co. are little more than gimmicks, good for party
824 games, grins and giggles, and little more. The AI community
825 seems to be making the same mistakes they made in the late 60s
826 and 70s. The second AI winter is nigh.
828 ** how do you see non-existent things ? (Score:2)
829 (by bingoUV ( 1066850 ))
832 If Consumer AI doesn't have a future, how can that non-existent
834 In an alternative interpretation, the author has seen the future
835 of Consumer AI and so of course it exists. But the future of the
836 future of Consumer AI doesn't exist. I.e. Future of Consumer AI
837 doesn't have one - where "one" stands for future.
838 Any other interpretations ?
840 ** It's not "Consumer AI" (Score:1)
841 (by themusicgod1 ( 241799 ))
844 Since the consumer is not control of it.
846 It's Anti-Consumer AI if anything
git://git.nikiroo.be / gofetch.git / blob