Ask HN: What is your prediction for next 10 years?
Which niche technology today will have high impact in coming 10 years ? share your thoughts in couple of sentences please.
Can you elaborate please?
Don't get me wrong, I love my online banking as much as the next guy, but a lot of technology being produced has a very negative impact - look at social media.
* RISC-V, and a general open source hardware revolution.
* Augmented Reality. I don't see VR going anywhere. AR has broader practical applications.
* Voice UI. I think this still counts as niche. The general interaction model at the beginning of the movie Her seems highly plausible to me. Not so much the AGI.
* Desktop/mobile OS convergence. This seems inevitable. The convenience factor is too high and why would the average Joe keep forking out for 2 or 3 computers when they can have 1 that does everything.
* Software formal verification. This is long overdue and we can't continue to be this irresponsible.
* Cryptocurrency, or something like it. I think the need is there, but the current implementation leaves much to be desired.
* Processors with many cores. We're already up to 18 or so on desktop processors. But I expect many more.
* FPGA. I'm less sure about this. 10 years might be to soon. But if your computer was just full of FPGA's and companies could sell you digital hardware through an app store without having to actually manufacture anything, I'm sure they would love it.
What's interesting is that it will probably be precipitated by a low-level breakthrough in energy storage or transmission. Somewhere today there is a junior researcher, laboring away at some currently un-sexy technology, the butt of more fashionable and trendy peers, who transmutes into the next Edison in less than a decade's time ;)
You just pay for storage and byo from Dropbox, google, whatever. With encryption you’d the be able to use services that are free, free and privacy focused.
I think JAM stack would work well with this and I’m thinking of making some POCs
Maybe: Crypto-currencies face court cases which try to determine what is actually owned legally, potentially undermining them.
Amazon faces a existential legal threat over counterfeit goods. They pay the settlement with AWS revenue, which by that point provides the majority of their income.
Younger generations develop an "in-real-life" counter culture, leading to financial volatility for existing social networks. In general, the attention economy matures, flattening corporate revenues year over year.
The federal government and state AGs prosecute major court cases against parties involved in textbook sales for illegal and anti-competitive behaviors.
A "Trader Joes"-style university emerges with a reasonably high quality, relatively inexpensive, accredited undergraduate education, with all course materials produced under the university brand. It runs year-round, has no tenure, and pays professors relatively well with standard benefits. The new university and its copy-cat competition create economic pressure on traditional academic institutions that have become administratively heavy over the previous three decades and, in some cases, have substantial construction debts that are now under increased interest rate pressure.
Maybe (related): Congress passes, and the president signs, a bill reintroducing bankruptcy discharge of student debt, possibly hidden in a must-pass omnibus spending bill.
Finally, for remaining in-use 32-bit unix-based systems, the 2038 problem is now 10 years closer.
We are at an interesting point hardware wise. You can make an argument that the approx processing power to be equivalent to a human is around 100 TFlops (https://jetpress.org/volume1/moravec.htm) and the cost of that drops about 10x every 5 years (https://aiimpacts.org/wikipedia-history-of-gflops-costs/) and is now of the order of $10k (https://fudzilla.com/news/graphics/47726-nvidia-unleashes-th...) which would make that about $1m ten years ago and $100 ten years in the future. Cheaper hardware may lead to a fair bit of innovation.
One reason I think that is because most of the core capabilities missing from AI systems are things that animals can do. Most of the difference with humans is a matter of degree.
It's still a very challenging list of capabilities of course. The core requirements for AGI are fairly well understood at this point. No one knows exactly how to achieve all of them together but many geniuses are working on it. And the tools such as DL, RL and less mainstream approaches such as SNNs and Sparse Predictive Hierarchies are very powerful. AGI is a real field now and intersects with things like developmental robotics.
It's a long shot but there are so many great minds involved and the rewards are so large that I think making a bet for within ten years is not a bad bet. Sure we have said that before. Things are not the same as before.
That seems wildly improbable to me. It seems far more likely that anyone who thinks they understand them are grossly underestimating what's involved.
Yeah, I know that I can't prove my position...
What I mean is we have lists of capabilities that AGI needs to have and know how it differs from narrow AI. For example: online continuous learning, perception, generalization, transferring skills between domains, etc. These aspects of the _what_ (rather than how) of AGI have been well researched.
Here are my prediction for the next 5-10 years.
1. Autonomous cars will be common, and they will have hive minds as AI progresses. 2. Governments around the world will look different, as AI will supplant politicians as representatives (probably in late decade). 3. Banking systems will crash and bartering will occur in the first half of the decade. It will change to adapt usage of cryptocurrency. 4. Blockchain will be part of the internet infrastructure. AI personas will exist as butlers for people, these personas are actual replica of their personalities. 5. Robotics will replace most jobs such as waiters and cashiers. Some of these robots will be monitored or operated by one or two persons. That job will require higher tech education/degrees.
(I have 5 more but it's not related to tech, lol).
- Moore's law is already dead, but in 10 years the implications of that will be fully realized
- Global warming won't be curtailed, rich countries will learn to cope with extreme weather patterns, people from poor countries will emigrate
- Cancer treatment will improve dramatically for a large variety of relatively niche cancers, but there still won't be a common cure
- Humans will still not go to Mars
- Soft AI will become prevalent for many low-level automation tasks, but general AI will stay in science fiction
- there will be significant headway made in resolving general relativity and quantum mechanics, likely by finding more situations where they disagree but not resolving those discrepancies
- We'll find exoplanets likely suitable for life, but too far away to reach them.
- Fusion power will just be "10 years away"
- China's chip industry will be fully caught up to the U.S.'s, Intel chips will still be the best, but only by a few percentage points and won't be economical at a large scale.
- Building custom ASICs will be far more common, and will be an important way of improving performance, as well as differentiation in the market (e.g., Adobe will produce their Photoshop chip).
- AI will get very good at analyzing PDFs and HTML tables. There will be apps that you can give access to your email that will be able to make very detailed itemized reports of your spending habits, and make recommendations on how to save money or get more for it.
- Dominant forms of trust will form on the internet. I am inclined to think that traditional media will take up this role, but it's possible that others will fill this space too.
- Even though humans will not go to Mars in 10 years, there will be a credible plan to put humans there (possibly taking another 10 years).
- Probes to Titan will be launched, and credible plans to land something on the moon will be well underway.
- Genomics in criminal law continues it's rapid growth, making unsolved rape cases a thing of the past.
- China's rapid growth in creating patents starts causing internal issues; they face their own first internal patent war, causing significant internal economic issues, forcing them to re-evaluate the value they ascribe to patents.
I am not bullish on self-driving cars - I think that will take longer than people think (more than ten years to become mainstream).
I think EV combined with ride calling (Uber/Grab etc.) will grow considerably.
Doesn't matter elaborate
But it's more than that. I've never believed the "To the moon!" predictions, because I've never believed that the real use value was there. It seems that people are investing on belief, but the actual real world use is... almost non-existent.
What's the daily volume of Bitcoin transactions that are purchases of other things, rather than investments in Bitcoin? [Edit: Divide that by the existing number of Bitcoin. That gives you the fraction spent in a day. What's the same figure for the US dollar?] Is Bitcoin really a currency if (relatively) nobody uses it to buy things? No, at that point it's an investment. [Edit: Or a store of value. But the value is fluctuating too much for it to be a very good store of value...]
But if there's no fundamental value, and it's not being used as a currency, what's the rational basis for its value? This makes it seem like a bubble.
Now, I don't actually have data for volume of purchases made in Bitcoin. I'm going on my impression, and I could be wrong.
As time goes forward the computational substrate will mature. Huge amounts of compute will be available for less and less money through easy to use APIs.
The number of people working on AGI will only increase. People will keep taking cracks at it until they get it. They will get it much sooner than anyone thought. The only reason it hasn’t been done already is because there was no computational substrate for it to incubate in so to speak. Now that the computational tinderbox has been created, all that is needed is the smallest of sparks. It will happen soon and unfortunatley it will be very unpleasant. I am being completely honest when I say that this is the sole reason why I have not decided to have children and when I implore anyone who reads this to not bring anyone into the world. The future will be very bleak for future generations.
The economics of AI demand the casting off of human life as we know it. The world as we know it now only sustains is because we as humans are the exclusive source of intelligent signal processing. When we are superseded in that ability, we will be lost to evolution. It’s impossible to know exactly how the transition will happen but one of the few things that can be relied on is that human society will become extremely fragile. Human society persists because we are the most dominant life form. When machines take over, we may be given token existences or a society pen to live in but it will be vestigial. There is no force of evolution or economics that ensures that each successive generation or “species” of machine intelligence will accommodate humans.
Many different machine entities will compete and fight, and in that way evolution will continue much as it has before. We may linger for a time in that world, but we do not have a divine right to exist in that world as we do in this one. We will be vestigial and ultimately lost. And all of this is just the economics/evolutionary aspect. I have not included here the possibility of a “terminator” outcome, or even mentioned the other countless grotesque possibilities.
Nobody will believe any of this. My opinion is universally dismissed. But it’s right.
Right now is the best it will ever be more or less, for humans. But we could preserve this state of affairs if we chose to.
We don't have any revolutionary substrate in the pipeline other than silicon in the next 10 years. A 50-100 years? maybe.
If everyone listened to you, there would be no future generations. Have you thought this through? "Let's end the human race, because something bad might happen" doesn't exactly sound very solid reasoning.
p.s. What did your girlfriend/wife say about your decision?
I'm somebody's girlfriend/wife and I do not want children (in general, and also because of future economic uncertainties) and I also don't like my existence being relagated to "someone who might want babies".