5 ms·

Ask HN: What core research are you most excited about?

26 points by icyfox 4y ago

Products and companies usually lag fundamental research by decades or more. What early stage publications or disciplines are you most bullish on for affecting humanity?

axg11 4y ago
Machine Learning: - Image generation - Multi-modal AI: (one algorithm that can process text, video, images, audio, etc) Genomics - liquid biopsy / cfDNA: early cancer detection from a blood test, other diagnostics - fast (<1 hr) cheap (<$20) sequencing Machine learning + biology - AlphaFold++
- OrangeMonkey 4y ago
  Your genomics comment is the same as mine. We are about to enter a world different from what our parents could have imagined. For the better, I think. Look at illumina and their company Grail who has the Galleri blood test. Vial of blood and they can report if you have internal cancers that are not easily detectable otherwise. It has me legit excited!
- hrgiger 4y ago
  Your list plus cellular aging, my other categories: os: hpc/highly tuned single purpose oses/unikernel machines will become more dominant filesystem: cxl on the horizon dax supported systems will become dominant then userspace+vendor shipped software will take over if they can catch up a common standard (intel already does good job) fantasy: hardware cooling: I believe we have a lot of room on having control on the heat, wanted to check alternative applications especially vacuum
  - the_biot 4y ago
    > os: hpc/highly tuned single purpose oses/unikernel machines will become more dominant As much as I like the idea, I'm not seeing this happening at all. What makes you say this will be more dominant?
    - hrgiger 4y ago
      I dont see someone to have a reason refuse a reliable software solution which would speed up their existing application which could be done just by an software upgrade and reduce their bills at no cost, most benefical I think would be the tunning part for an average user. Its also compliment with many existing architectures today including cloud/container orchestration and the baremetal itself. But probably the key is still community/vendor support and standards , I dont know how it would affect maintainer life
yababa_y 4y ago
The elaboration of rulial theory: https://www.wolframphysics.org/ https://www.wolframphysics.org/ It’s such an intriguing idea, and early correspondences seem so strong, that it seems impossible we won’t learn some awesome things about how physical theories relate and arise.
h2odragon 4y ago
I think there's some core research in magnetic control of plasmas / energy transfer from plasma that's been done, and is going to start leaking into more public applications like fusion. The military's "hypersonics" may be a beginning of some of that; bringing out things they've either been sitting on or desperately trying to get working for 30+ yr.
- starwind 4y ago
  really interesting point about hypersonics and plasmas. Developments in aviation have trickled down and led to other developments in technology—tons of work in modern control theory, material sciences, and aerodynamics was originally motivated by trying to build a better plane. I’d be surprised if we didn’t learn a few very practical things about plasmas while working on hypersonic weapons
bluelightning2k 4y ago
Fusion
- icyfox 4y ago
  Are we materially further along on fusion than we were 10 years ago? I know we're making some strides with magnetic / vacuum sealing but I'm unclear on how much novel insight the research has surfaced, versus just iteratively improving our ability to physically prototype it.
  - mecsred 4y ago
    I'd say we're about 20 years away from commercial fusion.
starwind 4y ago
Deep geothermal energy. Here's the best explanatory video I've seen: https://youtu.be/8erbvqFZ9M8?t=50 https://youtu.be/8erbvqFZ9M8?t=50 Caveat, I have no idea if these guys can do it but someone can. At this stage, it's about getting done it accurately and profitably. https://newatlas.com/energy/quaise-deep-geothermal-drilling-questions/ https://newatlas.com/energy/quaise-deep-geothermal-drilling-... https://www.vox.com/energy-and-environment/2020/10/21/21515461/renewable-energy-geothermal-egs-ags-supercritical https://www.vox.com/energy-and-environment/2020/10/21/215154...
- mountainriver 4y ago
  Yeah this is super cool and has the benefit of being always on
- h2odragon 4y ago
  Good link on current geothermal: https://austinvernon.site/blog/geothermalnextsteps.html https://austinvernon.site/blog/geothermalnextsteps.html
morelandjs 4y ago
Advances in additive manufacturing. Think 3D printers on steroids.
- ericb 4y ago
  What are the new advances that have you excited?
abrichr 4y ago
Conscious realism and the cosmological polytope: https://www.youtube.com/watch?v=reYdQYZ9Rj4 https://www.youtube.com/watch?v=reYdQYZ9Rj4 https://arxiv.org/abs/1709.02813 https://arxiv.org/abs/1709.02813 According to Prof. Hoffman, this model of the universe is highly parsimonious in that it can model data from the Large Hadron Collider using a single parameter, while the best incumbent theory (quantum field theory) needs millions. The implication is that everything we see and experience, including space and time itself, are not fundamental. The unit of reality is consciousness. This has far reaching applications into every other scientific and non-scientific human endeavor, from neuroscience to philosophy. It's no exaggeration to say that if he's right (and he claims the math shows that he is) it may be the most important discovery in the history of humanity. Even Albert Einstein appears to have intuited this when he wrote: "Time and space are modes by which we think and not conditions in which we live."
pdimitar 4y ago
OS kernel level observability. I shouldn't need to fire up `strace` for that; I should be able to do something like `observe $my_program $my_args` and have streaming results somewhere in `/proc`. Another thing: CPU-level observability. I want to be able to get statistics about a block of code: which instructions got ran the most, which ones are the slowest etc. Profiling on steroids, basically. If Intel / AMD are not willing to provide that then I'm willing to work with almost any CPU model that does offer that. Anybody knows of such CPUs? I'm also interested in eBPF but sadly can't find the time to dive deep there.
jhj 4y ago
ML-assisted combinatorial optimization and searching in high-dimensional discrete spaces (MILP/SAT/SMT solvers/theorem provers, etc) ML-assisted continuous regime optimization (e.g., in solid/fluid mechanics problems, chemical problems, etc) ML-assisted PDE solvers (e.g., weather simulation, etc)
f0e4c2f7 4y ago
Fusion is a big one. Some of the quantum computing research is pretty wild lately. People essentially writing "code" that gets converted by a quantum computer into new states of matter. https://www.popularmechanics.com/science/a40359762/paired-time-crystals-quantum-computing/ https://www.popularmechanics.com/science/a40359762/paired-ti...
egberts1 4y ago
Malware and WASM/JavaScript

26 points by icyfox 4y ago

Products and companies usually lag fundamental research by decades or more. What early stage publications or disciplines are you most bullish on for affecting humanity?

axg11 4y ago

Machine Learning: - Image generation - Multi-modal AI: (one algorithm that can process text, video, images, audio, etc) Genomics - liquid biopsy / cfDNA: early cancer detection from a blood test, other diagnostics - fast (<1 hr) cheap (<$20) sequencing Machine learning + biology - AlphaFold++

OrangeMonkey 4y ago
Your genomics comment is the same as mine. We are about to enter a world different from what our parents could have imagined. For the better, I think. Look at illumina and their company Grail who has the Galleri blood test. Vial of blood and they can report if you have internal cancers that are not easily detectable otherwise. It has me legit excited!
hrgiger 4y ago
Your list plus cellular aging, my other categories: os: hpc/highly tuned single purpose oses/unikernel machines will become more dominant filesystem: cxl on the horizon dax supported systems will become dominant then userspace+vendor shipped software will take over if they can catch up a common standard (intel already does good job) fantasy: hardware cooling: I believe we have a lot of room on having control on the heat, wanted to check alternative applications especially vacuum
- the_biot 4y ago
  > os: hpc/highly tuned single purpose oses/unikernel machines will become more dominant As much as I like the idea, I'm not seeing this happening at all. What makes you say this will be more dominant?
  - hrgiger 4y ago
    I dont see someone to have a reason refuse a reliable software solution which would speed up their existing application which could be done just by an software upgrade and reduce their bills at no cost, most benefical I think would be the tunning part for an average user. Its also compliment with many existing architectures today including cloud/container orchestration and the baremetal itself. But probably the key is still community/vendor support and standards , I dont know how it would affect maintainer life

yababa_y 4y ago

The elaboration of rulial theory: https://www.wolframphysics.org/ https://www.wolframphysics.org/ It’s such an intriguing idea, and early correspondences seem so strong, that it seems impossible we won’t learn some awesome things about how physical theories relate and arise.

h2odragon 4y ago

I think there's some core research in magnetic control of plasmas / energy transfer from plasma that's been done, and is going to start leaking into more public applications like fusion. The military's "hypersonics" may be a beginning of some of that; bringing out things they've either been sitting on or desperately trying to get working for 30+ yr.

starwind 4y ago
really interesting point about hypersonics and plasmas. Developments in aviation have trickled down and led to other developments in technology—tons of work in modern control theory, material sciences, and aerodynamics was originally motivated by trying to build a better plane. I’d be surprised if we didn’t learn a few very practical things about plasmas while working on hypersonic weapons

bluelightning2k 4y ago

Fusion

icyfox 4y ago
Are we materially further along on fusion than we were 10 years ago? I know we're making some strides with magnetic / vacuum sealing but I'm unclear on how much novel insight the research has surfaced, versus just iteratively improving our ability to physically prototype it.
- mecsred 4y ago
  I'd say we're about 20 years away from commercial fusion.

starwind 4y ago

Deep geothermal energy. Here's the best explanatory video I've seen: https://youtu.be/8erbvqFZ9M8?t=50 https://youtu.be/8erbvqFZ9M8?t=50 Caveat, I have no idea if these guys can do it but someone can. At this stage, it's about getting done it accurately and profitably. https://newatlas.com/energy/quaise-deep-geothermal-drilling-questions/ https://newatlas.com/energy/quaise-deep-geothermal-drilling-... https://www.vox.com/energy-and-environment/2020/10/21/21515461/renewable-energy-geothermal-egs-ags-supercritical https://www.vox.com/energy-and-environment/2020/10/21/215154...

mountainriver 4y ago
Yeah this is super cool and has the benefit of being always on
h2odragon 4y ago
Good link on current geothermal: https://austinvernon.site/blog/geothermalnextsteps.html https://austinvernon.site/blog/geothermalnextsteps.html

morelandjs 4y ago

Advances in additive manufacturing. Think 3D printers on steroids.

ericb 4y ago
What are the new advances that have you excited?

abrichr 4y ago

Conscious realism and the cosmological polytope: https://www.youtube.com/watch?v=reYdQYZ9Rj4 https://www.youtube.com/watch?v=reYdQYZ9Rj4 https://arxiv.org/abs/1709.02813 https://arxiv.org/abs/1709.02813 According to Prof. Hoffman, this model of the universe is highly parsimonious in that it can model data from the Large Hadron Collider using a single parameter, while the best incumbent theory (quantum field theory) needs millions. The implication is that everything we see and experience, including space and time itself, are not fundamental. The unit of reality is consciousness. This has far reaching applications into every other scientific and non-scientific human endeavor, from neuroscience to philosophy. It's no exaggeration to say that if he's right (and he claims the math shows that he is) it may be the most important discovery in the history of humanity. Even Albert Einstein appears to have intuited this when he wrote: "Time and space are modes by which we think and not conditions in which we live."

pdimitar 4y ago

OS kernel level observability. I shouldn't need to fire up `strace` for that; I should be able to do something like `observe $my_program $my_args` and have streaming results somewhere in `/proc`. Another thing: CPU-level observability. I want to be able to get statistics about a block of code: which instructions got ran the most, which ones are the slowest etc. Profiling on steroids, basically. If Intel / AMD are not willing to provide that then I'm willing to work with almost any CPU model that does offer that. Anybody knows of such CPUs? I'm also interested in eBPF but sadly can't find the time to dive deep there.

jhj 4y ago

ML-assisted combinatorial optimization and searching in high-dimensional discrete spaces (MILP/SAT/SMT solvers/theorem provers, etc) ML-assisted continuous regime optimization (e.g., in solid/fluid mechanics problems, chemical problems, etc) ML-assisted PDE solvers (e.g., weather simulation, etc)

f0e4c2f7 4y ago

Fusion is a big one. Some of the quantum computing research is pretty wild lately. People essentially writing "code" that gets converted by a quantum computer into new states of matter. https://www.popularmechanics.com/science/a40359762/paired-time-crystals-quantum-computing/ https://www.popularmechanics.com/science/a40359762/paired-ti...

egberts1 4y ago

Malware and WASM/JavaScript