Hacker News

anonymoosestdnt
Ask HN: What are some promising fields in computer science in the next 5 yrs?

mikewarot5 hours ago

Multilevel capability based secure Operating Systems are due for a resurgence thanks to the LLMs that make it obvious that ambient authority based systems aren't suitable for the modern world, what with persistent internet and mobile code. Genode is supposed to be coming out with a prototype of a capabilities based gui (or at least some sketch of it) at the end of August this year.

It's important to understand what a data diode is, and how to properly use them, along with air-gapped networks. You can have computer security enforced by the laws of physics, if everything is done right.

There are many important advances to be made in the field of high throughput processing. We've relied on CUDA for far too long, things like TinyGrad offer a hint of what's possible. There are also FPGA and similar architectures that can do pipelined processing far outside what's possible in a Von Neumann architecture.

austin-cheney10 hours ago

Transmission management. Right now MuleSoft is blowing up in popularity in the government sector, but it’s ripe for competition from a superior alternative. I know I could do it better and I am sure other people who have written high performance data transmission solutions could likewise do it better.

Performance as a feature is a business idea I don’t see anybody leaping into. Think about custom task automation on demand as a service that also provides benchmarks and logging. I imagine that as something like test automation but as a practical business solution for humans and businesses.

Self-hosting as a preconfigured and packaged product. This is largely blocked by the crazy prices of hardware, but it’s coming.

Location and fleet management. This is more than distribution centers. The military, Amazon, and Walmart have super invested in this like you couldn’t imagine for over 25 years and yet this is still not as mainstream as you would think.

iyamani9 hours ago

Not that I am in the field, but cybersecurity is still promising after all these years. One of the main reasons is that deep inside the leadership of cybersecurity organisations, there is absolutely no trust in AI taking decisions and it will not change for a while. This, combined with malicious actors integrating AI in their tooling, the demand in experienced analysts is still growing.

hn-front (c) 2024 voximity
source