Ask HN: What is the next hype (after ML)? | Hacker News Reader

Ask HN: What is the next hype (after ML)?

18 pointstw101010 comments

Everywhere I look I see stories of ML classes filled to the brim. To some this is an indication of an ML goldrush. But to me it's a sign that supply is about to meet demand real fast. And this makes me want to look into the future, to figure out what is going to explode in interest next.

So, HN, what piece of technology do you think is going to grow in popularity next? 3D printing seems to have died off a bit. Internet of things is a bit of a dud. The obvious answers are things in the decentralized sphere, like ipfs and bitcoin. But are there any less obvious regions that you have your finger on the pulse of, that you predict will grow in the near-ish future (say, 3-5 years), which not many of us are aware of?

View on HN

I feel like there's money in some lower end cloud.

At the moment, AWS/GCP/Azure just kill you with network egress charges. Consider that the cheap dedicated server companies (Hetzner, OVH) typically include about $4k/month or more of "free bandwidth" as compared to the big three.

I imagine Digital Ocean, Linode, OVH, and the like will add the cloud features that are missing (VPC, object storage, load balancing, DB as a service, SQS, Cloudwatch-like, etc) and start picking away at them. Or perhaps some newcomer.

Basically, I feel like a commodity cloud rush is coming.

E-bikes and other kinds of electrified vehicles. They have reached a "tipping point" because batteries and engines are so good that even out of shape people can use them to commute to work. What is left is regulation. When the EU and the US makes 1000 W engines and speeds >25 km/h street legal, then there will be few reasons to drive cars left.

I think your right - you see e-bikes increasing the practical commute distance from say 5 miles to 10 miles in a lot of cities

just got my first e-bike - its an electric cargo bike (babboe). Great for getting the kids around - to be honest most riding doesnt need to electric - but its good for the hills or a head wind

AI, Cloud, Robotics. Will continue to dominate the narrative in the four year time frame.

I've heard something like a quarter of venture capital is being allocated to consumer robotics. And it feels like society is reaching a tipping point. Where the hardware is becoming dependable enough that mainstream America is ready to allow robots in their homes en masse.

One interesting way to enter if you are a startup is to design a crucial component of the consumer robotics supply chain. Perhaps a remote sensing kit. Or modular battery system. But beyond the hype. There is enough substance here to get excited about this space and its transformative power.

Edge computing. Run computationally intensive tasks on the end user's machine instead of the cloud.

Basically, what we were doing before cloud took off.

Cloud computing is great and all. But the recent privacy and centralization issues are real things. There is also the issue of internet bandwidth in rural and poorer communities, where cloud is not a viable option.

I think ML obviously has a lot of applications, but I also think things like "Neural Nets" and "self-driving cars" have great marketing. But in general I think software innovations are usually initially limited by hardware, so as more low powered convolutional accelerator chips come on the market you will see more and more of the technologies in self-driving cars move into smaller, more mobile, and more naturally behaving robots. And robotics is made up of numerous fields: vision, planning, sensors, sensor fusion, perception, firmware, application software, etc. So any of those will come along for the ride, ML is great but it can't interact with the world without UI and software developers and all sorts of other software specializations.

I believe IoT will have a comeback soon coupled with blockchain technologies...

... or in layman terms, small devices that use P2P networks with cryptography to send notifications nobody can erase. Great for billing.

The next big thing will be unsupervised spiking neural networks. http://rebelscience.blogspot.com/2017/09/unsupervised-machin...

My guess would be on Blockchain and related technologies. Like FinTech companies trying to make cryptocurrencies common

In line with @tyingq, the "cloud" will become a commodity, and immutable infrastructure will automatically reprovision itself to run on the cheapest set of inventory purchased across multiple clouds.

Reply on news.ycombinator.com