Ask HN: What underestimated technologies will dominate within 5 years? | Hacker News Reader

Ask HN: What underestimated technologies will dominate within 5 years?

48 pointspanabee54 comments

This chart[1] from anandtech.com plots the impressive performance trajectory of Apple silicon over several years.

Of course, no technology is guaranteed to continue improving forever. The sustained performance gains reflect Apple's tremendous dedication and investment in the endeavor.

After the A11 chip, though, it became clear there was a non-zero probability that Apple could one day meet or surpass Intel.

Which begs the question: what underestimated technologies exist today that could overtake incumbent technologies within 5 years? Put another way: what technologies are in the A11 stage and not yet widely recognized for its potential?

[1] https://images.anandtech.com/doci/16226/perf-trajectory.png

View on HN

Lidar.

People often pigeonhole Lidar as an "Augmented Reality Enhancement" which is itself a massive underestimation of what it is capable of. Lidar isn't new tech, but it has become cheaper, smaller, and more importantly the compute capable of processing it massively more available (e.g. under-used machine learning chips in modern smartphones).

Lidar for mapping objects/rooms/buildings/etc. Lidar for mapping people/yourself.

For example that custom countertop? No need for an in-person visit, just use Lidar to scan the space via our app, and we'll have it delivered in 4-6 weeks.

You want a custom tailored suit? Just scan your whole body using Lidar, and we will custom make it to those measurements abroad and have it delivered.

Got into a minor vehicle accident? Just scan the damaged section, and we'll be able to quote the exact cost of repairs.

Lidar is a new core/low level sensor. It is a gimmick today because of the chicken/egg problem, but we'll see it eventually become completely standard and expected for normal life stuff. Boring stuff. AR isn't the what it is about, that's just an easier sales pitch.

Capability Based Operating Systems, such as Genode will allow the deployment of systems that stop many classes of attack by simply not trusting applications, with a default deny instead of allow for access to resources.

Reconfigurable Computing, such as FPGAs and more homogeneous computing fabrics offer a way around the Von Neuman bottleneck.

Reactive programming, like Verilog, offers a way to utilize FPGAs and multiprocessing in systems that take full advantage of thee available hardware, and route around Amdahl's law.

Lots of very interesting information security tech.

Fully homomorphic encryption would be revolutionary, but the performance isn't acceptable at the moment. As soon as it became acceptable it would be adopted swiftly in many domains.

Until then, fully encrypted memory will help prevent a lot of future attacks like Spectre, Meltdown, Heartbleed, and ShellShock. Intel and AMD both seem dedicated to implementing this.

ECC memory would need to become universal to prevent RowHammer style vulnerabilities. Would love to see that take off, and included as a mandatory part of a future DDR6. No prediction as to whether that will actually happen, and the market segment that cares the most has already had it since forever.

Probably fast solid state hard drives. That technology isn’t underestimated but it is under appreciated due to prohibitive price barriers. On modern computers the slowest low level performance, hands down, is walking a large file system. It’s hard to say what capabilities will exist once that performance barrier is lifted from common users because it will expose capabilities that have never been computationally convenient.

Is this not here already?

Yes, but costs a premium currently. Intel Optane product line (https://en.wikipedia.org/wiki/3D_XPoint tech) benchmarks: https://cdn.mos.cms.futurecdn.net/Z8YFAHnD7uYd8jPSx5scJN-970...

And what of HDDs that do compute tasks in addition to IO as an underestimated thing?

Millimeter accurate GPS.

I played with survey grade equipment for a day and it's crazy. Being able to repeatably measure positions and distances down to millimeters feels like magic. Imagine a VR headset that doesn't need trackers, a vacuum cleaner without camera/lidar, and so much more we can't even imagine without access to the technology. It's probably more than five years out though.

when your vacuum cleaner breaks your toes because it lost its GPS signal and went rogue

jokes aside it's fascinating that technology is even possible

Hehe GPS was never meant for that kind of accuracy. The reliabily is terrible today. It's a very cool hack but it will break your toes sooner rather than later.

Haha, I think the HN submissions will be more along the lines of "How I spoofed a GPS signal to get my neighbor's automated lawn mower to cut my grass"

Related to that: high-precision indoor positioning piggybacking of WLAN etc.

Worked in that space for a year and boy is indoor positioning with wifi or Bluetooth unreliable. I bet on it never making it, especially with all the other possibilities for indoor positioning opening up.

what do you think are the likeliest technologies to power indoor positioning within 5 years?

Laser & Lidar and also straight up GPS for many big buildings.

My robot vacuum with lasers has better indoor positioning than we ever managed to come close to in ideal conditions.

With the Wifi standard extensions to allow time-of-flight, or the "old-fashioned" signal strength methods?

I thought GPS signal is scrambled (artificial noise added) for national security purposes.

You are thinking of “selective availability” which was turned off over 20 years ago https://www.gps.gov/systems/gps/modernization/sa/

Thanks, yes. Although my gps receiver acts like it didnt get the memo.

Underestimated doesn’t necessarily mean unvalued or unhyped, it just means that the future value will be dramatically higher than people expect, even if what people expect is high to begin with.

So having said that: machine learning.

It’s basically been proven that throwing more compute and data at a problem gives proportionally better results. Getting to GPT10 is just a matter of money at this point. This straightforward scaling scenario leads inevitably to more resources spent on more compute (and different kinds of specialized computation), which simultaneously drives down prices and pushes forward performance.

Siri is the Altavista of AI. When the Google of AI appears in the next 5 years, it will rock our world harder than anyone expects. No one expected Google, even though plenty were hyping the internet. The same thing is happening with ML today.

Why do you call it gpt10 and not gpt4?

Very astute observation though.

1 or 2 GPT upgrades over the next five years get us to about 10? 10 is an arbitrary number, but my point is that ML advances are coming at an accelerating pace.

Yeah for sure. I personally think GPT4 is going to be very impressive.

Starlink will be huge. I would like to receive the thing right now. I don't think people appreciate just how big it will be.

For context: right now, I'm in Algiers, Algeria, with an ADSL connection of 4Mbps paid around $30/month for home usage. Recently down to $15/month, but the. 2Mbps for businesses is $77/month. Now, I've tried numerous times to upgrade this.

Fiber was priced at around $190/month for home usage, with many people showing upload is capped at 1Mbps. Yes, that's "one megabits, not megabytes, per second". Even then, I was willing to get it.

20Mbps for business is ... $505/month. It's cheaper for home use at ... $77/month recently for 100Mbps, which is a price drop from $190/month a month ago... if you can get the company to install FTTH, and if you live somewhere 'with DSLAM' as they say. They can't serve downtown.

ADSL and fiber are only available through the state owned company, of course, which also proposes Wimax, and 4G. Other carriers have 4G. Doesn't work where it matters. Plus it's capped. There were competitors in ADSL and Wimax, they were closed and had trouble with the regulation authority which oversees all.

I once needed someone remote to be on video call with a physician. ADSL went down. I sent "airtime/4G" credit to 8 numbers belonging to people who were present so they could have "internet" and be able to make the call. These numbers were of the three different carriers. No internet.

Starlink will be huge: bypass state monopolies, bypass regulating bodies. People will pay off customs agents to get receivers in if authorities try to restrict their imports, and there will be a black market for that if that happens.

Many computer architectures struggle with arithmetic vs. memory performance. A typical CPU can do a lot of computation per unit of time, but ends up starved because of low memory bandwidth.

I hope that we see either an existing fast memory technology become cheap, or a new technology arise that will bridge this gap.

I think modern day CPUs and GPUs remain underappreciated and not used to their full potential because of how hard it is to program around slow memory…

Seriously, we don't need esoteric new computer architectures, just give us cheap fast memory in a dumb flat memory hierarchy that is easy to program for.

It feels like there's fundamentally a tradeoff between "fast and flat" and "big". So basically I would guess NUMA and programming models that take it into account is the future.

And maybe software management of CPU cache.

I think wanting fast, flat, and big is sort of like wanting the CAP theorem to go away... "it sure would be great" if we could have an infinitely scalable SQL database without any programming model changes.

In other words, no matter how "fast" a memory becomes, someone's going to want it "bigger", which makes things slower.

1. Qubes OS [0], security-oriented desktop operating system based on Xen virtualization, especially when Qubes Air is ready [1], enabling cloud technologies.

2. GNU/Linux smartphones (which currently cannot be daily drivers for most people).

[0] https://qubes-os.org

[1] https://www.qubes-os.org/news/2018/01/22/qubes-air/

Next-Gen wireless networks such as 5G cellular and small-sat internet. I think most people are thinking fast internet in town rather than the ability to connect everywhere on the globe with broadband and the ability to build out rural areas without poles or trenching. Many communities are still stuck with DSL in the US.

I don’t know what’s coming, but Apple silicon seemed reasonably likely to match/surpass Intel chips a lot earlier than the A11... I think the A8 was the current chip when I realized it was coming, and I certainly don’t have any special/inside knowledge.

Probably not underestimated, but I suspect that in 5 - 10 years time we'll start to witness XR becoming everyday tech in our lives.

XR?

AR VR MR

I feel the XR revolution is like fusion energy in how it's constantly 5 years away. Maybe I'm just jaded.

The difference being you can buy VR and AR devices and use them right now. Maybe they’re not very good but it’s on the market right now, which is better than what fusion power has given.

XR is not really what we want first. VR can offer more.

Text-Speec-Recognition AI. KIT recently made a breakthrough with 1 sec. delay between speech and output and 5% error-rate.

Id say the Graphene material will usher in a new era of technology. The problem currently lies in the manufacturing process.

Lets say manufactoring is solved. What can you do with Graphene you cannot do with other things?

Lightfield videos, and more decentralised services, like funkwhale is for music, but for email, chat, social networks etc.

https://brilliantlightpower.com/suncell-24th-hour-of-100-hou... is now approaching field trials of 300KW heaters that consume only water but still nobody believes either the science or the tech.

from the page : "Brilliant Light Power has developed a new commercially competitive, non-polluting, plasma-based primary source of massive power from the conversion of hydrogen atoms of water molecules to dark matter, the previously unidentified matter that makes up most of the mass of the universe."

previously unidentified Sorry to disappoint you, but this one is a scam.

Many have said the same over the last 30 years or so but they are now approaching delivery of independently verifiable working devices that may possibly turn physics on its head.

If they 'identified' dark matter, why do they keep calling it that and not what it is?

Presumably because it is more understandable and also grabs more attention than calling it hydrinos - small hydrogen with the electron in a lower orbital than the presently accepted ground state.

Hybrid neurosymbolic AI is promising I'd say

any recent link on existing work in that field ?

i'm super interested to know if any progress is being done there.

There's a lot of things to say but I'm just going to do a non exhaustive lazy comment:

Symbolic AI is a big family:

For datasets babelnet has been a big improvement but is not that much recent.

For symbolic formats the classic OWL/RDF way has been enhanced with SHACL which allow I believe to dynamically create schema that validate the semantic web graph

HyperGraphs and metagraphs are trending (check it out it's a datastructure that is actually more expressive than graphs)

Truth is: Hybrid AI is not original, it is obvious to anyone knowledgeable that you must use sota neural networks for preprocessing tasks such as POS tagging, universal dependency parsing, NER, etc

Those gathered metadata will feed the symbolic engine/DB

The truth is that almost zero hybrid AI systems exists or they have a very limited scope. Some tasks such as coreference resolution are currently neural only but a symbolic semantic understanding would improve the state of the art, same for word sense disambiguation.

Even the task of semantic parsing is almost only using neural networks, symbolic just isn't funded nowadays or researchers are ashamed of giving a new life to this paradigm which is ridiculous.

On the other hand some researchers are trying too hard to implement logic in neural networks which is to me a by design mismatch.

The only projects worth following that I know are: Cyc Wolfram language And the only "promising" one opencog. I'm currently researching on how to implement true semantic parsing but I'm not gonna expand on my work until it is ready. To sum it up: NLU is not ready for primetime and almost nobody is truly working on it (but a lot of people pretend to work on it by just pushing a bit further the illusion of understanding that comes from amalgaming correlation as causation)

Vaccines with mRNA (Pfizer and BioNTech's COVID-19 vaccine being an example).

I never understood how it gets the mRNA inside your cells?

Starlink, HVDC

AMD

Skynet

You mean Legion

This Skynet?

https://blog.sia.tech/skynet-bdf0209d6d34 https://siasky.net/

Looks promising!

Reply on news.ycombinator.com