Apart from a new JavaScript framework, Machine Learning, BlockChain, etc. Is there any new field or emerging tech that you think will make an impact in coming future or are we reaching a saturation point ?
A new JavaScript framework is not "the next big thing". It's the next thing that everybody's going to jump on as a bandwagon, but that's not the same thing at all.
The actual next big thing? Not a language. Not some hardware change ( I mean, ARM servers could be big, but only the server people will care. It won't be the Next Big Thing.) Not block chain - that's close to played out. Not machine learning - that's too far along. It's what's now, not what's next.
I see two possibilities. It might be quantum computing. I doubt it, but it might be. It also might be wearable VR. Again, I kind of doubt it, but I doubt it less than I do quantum.
That's not entirely fair to JS frameworks. There hasn't been a major new framework in years, with Angular, Vue and especially React dominating the ecosystem. React in particular presents a unique way of thinking about and composing UIs and the model has leaked into plenty of other ecosystems, for example SwiftUI which borrows heavily from it.
The reason new frameworks can't dislodge the big 3 is because they aren't bringing new ideas. Any framework that changed UI development as much as React did would absolutely be worth noting.
What's your opinion about Svelte? Has it made any big contributions to the ecosystem? Isn't the syntax and in general the way of creating components more intuitive with Svelte in comparison to for example React?
Quantum computing will revolutionize some areas. I doubt that it will revolutionize computer science as a whole. I think that there are many more categories of problems that it won't change than categories that it will.
Because the algorithms that offer "quantum supremacy" all are effectively analog algorithms, and rely on phase shifts to do the heavy lifting. Any small phase shifts get amplified quickly and swamp the usefulness.
All known "quantum error correction" algorithms are focused on not flipping qubits, and ignore phase.
For instance, using Grover's algorithm to crack a 128 bit key requires 2^64 quantum operations in sequence. To make this happen your phase shift and interaction between the qubits has to be low enough that it doesn't accumulate to the level sufficient to flip a single qubit in all that time.
You'd need signal to noise ratios of about 380 dB to make this happen. This would be 220 dB more!!! than detecting the signal from the Voyager (10^-20 watts) probe right next to a SuperNova explosion (10^40 watts)
I think we are hitting a decade of tech winter. Not seeing anything substantial in the 2020s. VR will be limited to gaming & entertainment. Not sure any other platform or tech is emerging at this point.
IMHO Apple’s AR/VR device will be another incremental but major milestone for mainstream VR.
Stable Diffusion is already making waves in digital art. I expect more specialized models fairly soon.
Free data and services we used to take for granted like YouTube, farming software, and car seat heaters are becoming forcefully monetized. Will we someday refer to these as the good old days when search engines were “free”?
This is peak HN bubblism. No, people don't actually care about security or privacy if one looks at their revealed preferences and not what they say they value (see how many still use Google for example). People will continue to use computers as long as it serves a need for them.
Apple pushes the security angle as a big differentiator. Source: https://www.apple.com/privacy/
The title of that page says "Privacy.That’s Apple.".
I think as people get more knowledgeable of tech they will care more about security or privacy. It is already a big selling point of messaging applications.
Maybe some more practical and less costly formal verification?
There are some signs Rust is heading in this direction.
Efforts like Dafny are also pretty interesting, but still far from mainstream.
I think a large codebase (20-40 KLOC) is about the maximum one can routinely verify in Dafny before stuff gets too hard. And coding speed is pretty slow.
CUE(lang) and that we finally have a proper language for configuration and schemas. I think it will be significant in DevOps and anywhere we use Yaml or json for config.
I checked out Dhall back when I also found CUE. Definitely do not like it. The syntax is too different from existing configuration languages to see mass adoption. CUEs philosophy is much better outlined and reasoned, and comes from the same place as containers, kubernetes, and Go.
Another benefit of CUE over Dhall is that I can use CUE from Go. Very few people use Haskell, Go is an industry standard at this point
Personally I'm really hoping for better private decentralized apps and networks.
A way to escape the google silos and protect your private data from mass surveillance, but without losing the usability. Things like locutus, solid, etc.
Any examples or categories? I think blockchain is interesting but don't see the real-world usefulness to be honest, other than bitcoin. Most applications can just use a database.
The actual next big thing? Not a language. Not some hardware change ( I mean, ARM servers could be big, but only the server people will care. It won't be the Next Big Thing.) Not block chain - that's close to played out. Not machine learning - that's too far along. It's what's now, not what's next.
I see two possibilities. It might be quantum computing. I doubt it, but it might be. It also might be wearable VR. Again, I kind of doubt it, but I doubt it less than I do quantum.