On January 1st, 2018, I made predictions (here) about self driving cars, Artificial Intelligence and machine learning, and about progress in the space industry. Those predictions had dates attached to them for 32 years up through January 1st, 2050.
So, today, January 1st, 2019, is my first annual self appraisal of how well I did. I’ll try to do this every year for 32 years, if I last that long.
I am not going to edit my original post, linked above, at all, even though I see there are a few typos still lurking in it. Instead I have copied the three tables of predictions below. I have changed the header of the third column in each case to “2018 Comments”, but left the comments exactly as they were, and added a fourth column titled “Updates”. In one case I fixed a typo (about self driving taxis in Cambridgeport and Greenwich Village) in the left most column. I have started highlighting the dates in column two where the time they refer to has arrived, and I am starting to put comments in the updates fourth column.
I will tag each comment in the fourth column with a cyan colored date tag in the form yyyymmdd such as 20190603 for June 3rd, 2019.
The entries that I put in the second column of each table, titled “Date” in each case, back on January 1st of 2018, have the following forms:
NIML meaning “Not In My Lifetime, i.e., not until beyond December 31st, 2049, the last day of the first half of the 21st century.
NET some date, meaning “No Earlier Than” that date.
BY some date, meaning “By” that date.
Sometimes I gave both a NET and a BY for a single prediction, establishing a window in which I believe it will happen.
For now I am coloring those statements when it can be determined already whether I was correct or not.
I have started using LawnGreen (#7cfc00) for those predictions which were entirely accurate. For instance a BY 2018 can be colored green if the predicted thing did happen in 2018, as can a NET 2019 if it did not happen in 2018 or earlier. There are five predictions now colored green.
I will color dates Tomato (#ff6347) if I was too pessimistic about them. No Tomato dates yet. But if something happens that I said NIML, for instance then it would go Tomato, or if in 2019 something already had happened that I said NET 2020, then that too would go Tomato.
If I was too optimistic about something, e.g., if I had said BY 2018, and it hadn’t yet happened, then I would color it DeepSkyBlue (#00bfff). None of these yet either. And eventually if there are NETs that went green, but years later have still not come to pass I may start coloring them LightSkyBlue (#87cefa).
In summary then: Green splashes mean I got things exactly right. Red means provably wrong and that I was too pessimistic. And blueness will mean that I was overly optimistic.
So now, here are the updated tables. So far none of my predictions have been at all wrong–there is only one direction to go from here!
No predictions have yet been relevant for self driving cars, but I have added one comment in this first table.
[Self Driving Cars]
|A flying car can be purchased by any US resident if they have enough money.||NET 2036||There is a real possibility that this will not happen at all by 2050.|
|Flying cars reach 0.01% of US total cars.||NET 2042||That would be about 26,000 flying cars given today's total.|
|Flying cars reach 0.1% of US total cars.||NIML|
|First dedicated lane where only cars in truly driverless mode are allowed on a public freeway.||NET 2021||This is a bit like current day HOV lanes. My bet is the left most lane on 101 between SF and Silicon Valley (currently largely the domain of speeding Teslas in any case). People will have to have their hands on the wheel until the car is in the dedicated lane.|
|Such a dedicated lane where the cars communicate and drive with reduced spacing at higher speed than people are allowed to drive||NET 2024|
|First driverless "taxi" service in a major US city, with dedicated pick up and drop off points, and restrictions on weather and time of day.||NET 2022||The pick up and drop off points will not be parking spots, but like bus stops they will be marked and restricted for that purpose only.||20190101 Although a few such services have been announced every one of them operates with human safety drivers on board. And some operate on a fixed route and so do not count as a "taxi" service--they are shuttle buses. And those that are "taxi" services only let a very small number of carefully pre-approved people use them. We'll have more to argue about when any of these services do truly go driverless. That means no human driver in the vehicle, or even operating it remotely.|
|Such "taxi" services where the cars are also used with drivers at other times and with extended geography, in 10 major US cities||NET 2025||A key predictor here is when the sensors get cheap enough that using the car with a driver and not using those sensors still makes economic sense.|
|Such "taxi" service as above in 50 of the 100 biggest US cities.||NET 2028||It will be a very slow start and roll out. The designated pick up and drop off points may be used by multiple vendors, with communication between them in order to schedule cars in and out.|
|Dedicated driverless package delivery vehicles in very restricted geographies of a major US city.||NET 2023||The geographies will have to be where the roads are wide enough for other drivers to get around stopped vehicles.|
|A (profitable) parking garage where certain brands of cars can be left and picked up at the entrance and they will go park themselves in a human free environment.||NET 2023||The economic incentive is much higher parking density, and it will require communication between the cars and the garage infrastructure.|
|A driverless "taxi" service in a major US city with arbitrary pick and drop off locations, even in a restricted geographical area.||NET 2032||This is what Uber, Lyft, and conventional taxi services can do today.|
|Driverless taxi services operating on all streets in Cambridgeport, MA, and Greenwich Village, NY.||NET 2035||Unless parking and human drivers are banned from those areas before then.|
|A major city bans parking and cars with drivers from a non-trivial portion of a city so that driverless cars have free reign in that area.||NET 2027|
|This will be the starting point for a turning of the tide towards driverless cars.|
|The majority of US cities have the majority of their downtown under such rules.||NET 2045|
|Electric cars hit 30% of US car sales.||NET 2027|
|Electric car sales in the US make up essentially 100% of the sales.||NET 2038|
|Individually owned cars can go underground onto a pallet and be whisked underground to another location in a city at more than 100mph.||NIML||There might be some small demonstration projects, but they will be just that, not real, viable mass market services.|
|First time that a car equipped with some version of a solution for the trolley problem is involved in an accident where it is practically invoked.||NIML||Recall that a variation of this was a key plot aspect in the movie "I, Robot", where a robot had rescued the Will Smith character after a car accident at the expense of letting a young girl die.|
Right after the Artificial Intelligence and machine learning table I have some links to back up my assertions.
[AI and ML]
|Academic rumblings about the limits of Deep Learning|
|Oh, this is already happening... the pace will pick up.||20190101 There were plenty of papers published on limits of Deep Learning. I've provided links to some right below this table.|
|The technical press starts reporting about limits of Deep Learning, and limits of reinforcement learning of game play.|
|20190101 Likewise some technical press stories are linked below.|
|The popular press starts having stories that the era of Deep Learning is over.||BY 2020|
|VCs figure out that for an investment to pay off there needs to be something more than "X + Deep Learning".||NET 2021||I am being a little cynical here, and of course there will be no way to know when things change exactly.|
|Emergence of the generally agreed upon "next big thing" in AI beyond deep learning.||NET 2023|
|Whatever this turns out to be, it will be something that someone is already working on, and there are already published papers about it. There will be many claims on this title earlier than 2023, but none of them will pan out.|
|The press, and researchers, generally mature beyond the so-called "Turing Test" and Asimov's three laws as valid measures of progress in AI and ML.||NET 2022||I wish, I really wish.|
|Dexterous robot hands generally available.||NET 2030|
BY 2040 (I hope!)
|Despite some impressive lab demonstrations we have not actually seen any improvement in widely deployed robotic hands or end effectors in the last 40 years.|
|A robot that can navigate around just about any US home, with its steps, its clutter, its narrow pathways between furniture, etc.||Lab demo: NET 2026|
Expensive product: NET 2030
Affordable product: NET 2035
|What is easy for humans is still very, very hard for robots.|
|A robot that can provide physical assistance to the elderly over multiple tasks (e.g., getting into and out of bed, washing, using the toilet, etc.) rather than just a point solution.||NET 2028||There may be point solution robots before that. But soon the houses of the elderly will be cluttered with too many robots.|
|A robot that can carry out the last 10 yards of delivery, getting from a vehicle into a house and putting the package inside the front door.||Lab demo: NET 2025|
Deployed systems: NET 2028
|A conversational agent that both carries long term context, and does not easily fall into recognizable and repeated patterns.||Lab demo: NET 2023|
Deployed systems: 2025
|Deployment platforms already exist (e.g., Google Home and Amazon Echo) so it will be a fast track from lab demo to wide spread deployment.|
|An AI system with an ongoing existence (no day is the repeat of another day as it currently is for all AI systems) at the level of a mouse.||NET 2030||I will need a whole new blog post to explain this...|
|A robot that seems as intelligent, as attentive, and as faithful, as a dog.||NET 2048||This is so much harder than most people imagine it to be--many think we are already there; I say we are not at all there.|
|A robot that has any real idea about its own existence, or the existence of humans in the way that a six year old understands humans.||NIML|
With regards to academic rumblings about deep learning, in 2017 there was a new cottage industry in attacking deep learning by constructing fake images for which a deep learning network gave high scores for ridiculous interpretations. These are known as adversarial attacks on deep learning, and some defenders counter claim that such images will never arrive in practice.
But then in 2018 others found images that were completely natural that fooled particular deep learning networks. A group of researchers from Auburn University in Alabama show how an otherwise well trained network can just completely misclassify objects with unusual orientations, in ways which no human would get wrong at all. Here are some examples:
We humans can see why or how a network might get the first one wrong for instance. It is a large yellow object across a snowy road. But other clues, like the size of the person standing in front of it immediately get us to understand that it is a school bus on its side across the road, and we are looking at its roof.
And here is a paper from researchers at York University and the University of Toronto (both in Toronto) with this abstract:
We showcase a family of common failures of state-of-the art object detectors. These are obtained by replacing image sub-regions by another sub-image that contains a trained object. We call this “object transplanting”. Modifying an image in this manner is shown to have a non-local impact on object detection. Slight changes in object position can affect its identity according to an object detector as well as that of other objects in the image. We provide some analysis and suggest possible reasons for the reported phenomena.
In all their images a human can easily see that an object (e.g., an elephant, say, and hence the very clever title of the paper, “The Elephant in the Room”) has been pasted on to a real scene, and both understand the real scene and identify the object pasted on. The deep learning network can often do neither.
Other academics took to more popular press outlets to express their concerns that the press was overhyping deep learning, and showing what the limits are in reality. There was a piece by Michael Jordan of UC Berkeley in Medium, an op-ed in the New York Times by Gary Marcus and Ernest Davis of NYU and a story on the limits of Google Translate in the Atlantic by Douglas Hofstadter of Indiana University at Bloomington.
As for stories in the technical press there were many that sounded warning alarms about how deep learning was not necessarily going to the greatest most important technical breakthrough in the history of mankind. I must admit, however, that more than 99% of the popular press stories did lean towards that far fetched conclusion, especially in the headlines.
Here is PC Magazine talking about the limits in language understanding, Forbes magazine on the overhyping of deep learning. A national security newsletter quotes a Nobel prizewinner on AI:
Intuition, insight, and learning are no longer exclusive possessions of human beings: any large high-speed computer can be programed to exhibit them also.
This was said by Herb Simon in 1958. The newsletter goes on to warn that over hype is nothing new in AI and that it could well lead to another AI winter. Harvard Magazine reports on the dangers applying a an inadequate AI system to decision making about humans. And many many outlets reported on an experimental Amazon recruiting tool that learned biases against women candidates from looking at how humans had evaluated CVs.
The press is not yet fully woke with regard to AI, and deep learning in particular, but there are signs and examples of wokeness showing up all over.
Developments in space were the most active for this first year, and fortunately both my optimism and pessimism were well place and were each rewarded.
|Next launch of people (test pilots/engineers) on a sub-orbital flight by a private company.|
|20190101 Virgin Galactic did this on December 13, 2018.|
|A few handfuls of customers, paying for those flights.||NET 2020|
|A regular sub weekly cadence of such flights.||NET 2022|
|Regular paying customer orbital flights.||NET 2027||Russia offered paid flights to the ISS, but there were only 8 such flights (7 different tourists). They are now suspended indefinitely.|
|Next launch of people into orbit on a US booster.|
NET 2019BY 2021
BY 2022 (2 different companies)
|Current schedule says 2018.||20190101 It didn't happen in 2018. Now both SpaceX and Boeing say they will do it in 2019.|
|Two paying customers go on a loop around the Moon, launch on Falcon Heavy.|
|The most recent prediction has been 4th quarter 2018. That is not going to happen.||20190101 I'm calling this one now as SpaceX has revised their plans from a Falcon Heavy to their still developing BFR (or whatever it gets called), and predict 2023. I.e., it has slipped 5 years in the last year.|
|Land cargo on Mars for humans to use at a later date||NET 2026||SpaceX has said by 2022. I think 2026 is optimistic but it might be pushed to happen as a statement that it can be done, rather than for an pressing practical reason.|
|Humans on Mars make use of cargo previously landed there.||NET 2032||Sorry, it is just going to take longer than every one expects.|
|First "permanent" human colony on Mars.||NET 2036||It will be magical for the human race if this happens by then. It will truly inspire us all.|
|Point to point transport on Earth in an hour or so (using a BF rocket).||NIML||This will not happen without some major new breakthrough of which we currently have no inkling.|
|Regular service of Hyperloop between two cities.||NIML||I can't help but be reminded of when Chuck Yeager described the Mercury program as "Spam in a can".|
22 comments on “Predictions Scorecard, 2019 January 01”
You’ve highlighted the weakness of deep learning in some of your observations and photos. But it’s early with Deep Learning. I’m sure I’m not the only one that see’s this kind of deep learning as a sub-component to a larger overall AI system. So my question is: “Is anyone mixing older heuristic algorithms and concepts with newer deep learning algorithms to create a more capable overall AI system”. (I’m sure the answer is yes, and if I had 30 years ‘free’ I might be interested in playing with that!).
I’m pushing back on the many people that say that Deep Learning (which is 9 years old in its current incarnation, and 30+ years old it in many of its underlying techniques) is all that AI needs, or indeed, many newbies who think Deep Learning is all that there has ever been of AI. There will need to be many many different approaches combined. And we may not have **any** of them yet. We may be in the phlogiston era. See my upcoming essay, hopefully out this month. In the meantime for Deep Learning and images they have a serious problem on the input side in handling spatial composition.
has the essay you mentioned been published? (and, if so, where?)
No. It is turning into a book instead. And I am working hard on it.
You are short about 25 NIML’s
Driverless cars will not happen in a true fashion on open roads ever. It is a problem already solved to a higher degree by us. Replacement roads would be a logical first step in an illogical pursuit.
Yes, I have talked about this in my two long essays on driverless cars on this blog.
I disagree. I am 57 and expect to see “driverless cars… on open roads” in my lifetime (though not as soon as the “hypenotists” predict). I think military vehicles will do it first, but imo that counts. Of course, the caveat is that it depends on civilization’s continuance; not a given, for better or worse. I think the flaw in your argument is the claim that it’s a “solved” problem — tell that to the thousands of people around the world whose loved ones died in car crashes in any given week. Yes, humans are better than the current state of the art, but I’m not seeing an asymptote for the improvements currently being made.
Then we shall have to see who is righter in the long run.
No predictions on the most important areas of all: Energy production and storage. Without better nuclear or much more energy dense batteries or fuel cells etc, humanity is hosed.
I suspect we need a next gen nuclear to solve the energy problem, I give it NET 2017. Fusion … oi vey a running demo of at least a month by NET 2040. 5x denser energy storage 2025, 10x denser energy density 2032.
While I don’t disagree, this blog is only about places I feel I have some expertise, and energy predictions is not one of them… I’ll let others prognosticate about that.
Now Deepmind’s Alphastar can beat a StarCraft pro. I’m actually surprised something like that is not on this list. Is there any reason why you don’t involve some landmark video games? Does this not affect timescales?
I did not talk about video games at all. I did not talk about my PhD student David Chapman, and his 1990 PhD thesis titled “Vision, Instruction, and Action”, published as a 1991 book of the same title by MIT Press, where he built an AI program that could beat his favorite video game (Pengi) in the student center. Nor did I talk about how learning to play a video game was a common term project by students in my MIT class 6.836 during the 1990’s. They usually used a form of evolution as their desired machine learning technique. And used at least 1 billion times less computation than the kids at Deepmind. Sometimes their systems would find bugs in the video game and exploit them. It was a joy to watch. Nor did I talk about how Deep Blue beat the world chess champion, Kasparov, over 20 years ago. The current hype around playing games means diddly-squat for the advancement of AI. It is hype. And it is uninformed hype. And it is hype about work done by uninformed people.
I will happily differ to your opinion. I have been following your work on and off (but with Deepmind’s breakthrough) I’ve been diving into your work again. Could you expound upon why gaming is diddly-squat on AI advancement? Where can I read a write up on this (did you work on one)? Because, the AI community is loosing it’s mind over AI beating starcraft, being there are like 10^10 more moves and options compared to Go.
1. The AI community is not losing its mind of this starcraft thing. The hype-notists and the AI wannabe fannboys are losing their minds. I don’t count them as part of the AI community. And there is no “breakthrough”.
2. My whole blog is about why it is diddly-squat. There are eight (so far) long essays on the topic. Here is the index page.
wait… did you just describe your 6.836 students as “uninformed people”? 😉
No I did not. You can see that I was referring to current day people hyping game playing programs.
Could you please give references to those papers you referred in this section: “Emergence of the generally agreed upon “next big thing” in AI beyond deep learning. Whatever this turns out to be, it will be something that someone is already working on, and there are already published papers about it. There will be many claims on this title earlier than 2023, but none of them will pan out.” ?
Hah, funny! Of course not. There are tens of thousands of AI papers published per year. Some strain of AI amongst those papers, with years and years of work will turn out to be the next big thing. Just as deep learning was built on 30 years of slow progress. But there were hundreds of other simultaneous threads of work going on. Only one popped. And no one could predict which thread it would be. Just as today no one can predict which thread will turn out to have great success in the future. But I guarantee it will not come from just one single publication. It will be a long hard set of work, by a number of people who devote their career to it. By the nature of research the vast majority of researchers will never have a great success. Someone will, but we won’t know who until it happens.
“…no one can predict…” — a funny statement in the context of this post, and perhaps I can claim it’s wrong, since I kind of predicted it (unfortunately, not in any public or verifiable way). By “it” I mean the fact that neural networks, given enough training, recurrence, & complexity, would outperform other methods, even though those other methods performed better back in the ’80s, when I “predicted” that.
I don’t think it is incongruous. It says that the trend can be predicted but not the detail that unwinds. Just like evolution predicts that fitness will adjust creatures to an environment, but does not say exactly what form they will take.
I admit this is pedantic, but your math is off in the definition for NIML. The 21st century is from January 1, 2000 to December 31, 2100; so the last day of the first half would be a year later than you specify i.e. December 31, 2050.
Yes, very pedantic. And it looks like you didn’t read the original post where I addressed this issue in my definition: “And that also means that I am only predicting things for exactly the first half of this century (or at least for the first half of the years starting with “20” — there is a whole argument to be had here into which I am not going to get).”.