6 ms·

Request for Technically Risky Projects

20 points by astro-codes 6y ago

What technical projects do you think should exist in the world? What hard problems exist in your field? Where do you think the frontiers of science/technology lie? I'm interested in learning about hard technical problems you think are worth solving.

manx 6y ago
Creating a "Wikipedia of Arguments", one can refer to in public discussions. It turns out that this is a way harder problem than I thought. I'm in this space for some time now and am happy to talk about it. Relevant links: https://github.com/canonical-debate-lab/paper/blob/master/README.mediawiki https://github.com/canonical-debate-lab/paper/blob/master/RE... https://www.societylibrary.org/ https://www.societylibrary.org/
- jka 6y ago
  Let's hope that a well-constructed, well-behaved and well-participated argument forum maintained over long enough timescales can lead towards better outcomes for everyone. Best of luck with the project. You might be aware of it already, but just in case, the most progress I've seen in the area of web-based argument mapping has been Arguman ( https://github.com/arguman/arguman.org https://github.com/arguman/arguman.org ). They're also following an open source model and there could be opportunities to co-operate and/or share ideas.
  - manx 6y ago
    Thanks, I'm aware of it. There are quite a lot of tools in this space already. Unfortunately, many are unaware of each-other and therefore making the same mistakes and learnings over and over again. Looking at all these tools makes it clear how difficult it is to create one that actually "works". Most of the researchers at the canonical debate lab created their own tool in the past and are now discussing how a next-generation tool should look like.
    - jka 6y ago
      Cool. In some ways, that sounds like the software industry in a nutshell, to be honest :) Can you share a sense for what any of the core unsolved problems are? (I'd be happy to read a mailing list / previous discussions if there's too much background to be worth communicating in a comment here) I've noticed that Arguman seemed to run into community management and spam issues. More broadly speaking there seems to be a societal challenge in getting people to trust and feel invested in honest debating and to accept best-known truths.
      - manx 6y ago
        The core unsolved problems in my opinion are: - The argumentation data model - An attack/manipulation-proof community curation system (maybe similar to StackOverflow) - A user-interface that can be used by anyone who is interested to contribute, but powerful enough to work with the argumentation structure There is a weekly zoom meeting every monday that is recorded and uploaded to youtube: https://www.youtube.com/playlist?list=PLwMyf-sRX2_Hqw-h9ba_StFt5jOKlpBcF https://www.youtube.com/playlist?list=PLwMyf-sRX2_Hqw-h9ba_S...
        jka 6y ago
        Late reply here, but thank you - I'll try to catch up on a few of those recordings. Edit: small clarification
- gitgud 6y ago
  That is a brilliant paper, I expected it to be centralised around logical fallacies (which is how I view arguments/debates) but it goes into so much more detail around the nature of online debate. It's interesting to dive into the factors that influence online debates; reputation, trolling, anonymity, echo-chambers, the hive mind, language barriers, commercial interests.... There's so much more to an online discussion than just the argument presented...
HelpEveryone 6y ago
Spectacles that run 24/7
rapnie 6y ago
Decentralized search in p2p / federated networks. Privacy-respecting decentralized online identity.
- ragerino 6y ago
  I am sure you know about Solid. Just in case: https://solidproject.org/ https://solidproject.org/
  - rapnie 6y ago
    Thanks. Yes, I do. There is a lot of work in the field of decentralized identity, but status is still far from a 'solved problem'. Work being done by Rebooting the Web of Trust [0] (see their github repo's) and Identity Foundation [1]. Besides Solid there's a bunch of interesting, related specs evolving like Verifiable Credentials [2], Self Sovereign Identity (I hope for impls without the blockchain). [0] http://www.weboftrust.info/ http://www.weboftrust.info/ [1] https://identity.foundation/ https://identity.foundation/ [2] https://www.w3.org/TR/vc-data-model/ https://www.w3.org/TR/vc-data-model/
_alex_ 6y ago
The "cold-start" problem of new social network sites
- silentsea90 6y ago
  Apologies for the condescension, and with all due respect, building a social network is a bit neo Silicon valley frothy.
  - cweagans 6y ago
    the idea of building another facebook is just...ugh. however, a friend of mine has been working on a community site for fishing enthusiasts for years. it's _extremely hard_ to bootstrap a community site of any kind, even if it's not "social networking" per se.
    - muzani 6y ago
      I think the idea is to build a not-facebook. I'm working in social media too, and there's a lot of things that FB does poorly. The goal is usually media, centered around yourself. For some people, it could be their line of work, or happenings in their neighbourhood. FB, prioritizing growth, is like the BuzzFeed of social media, but we're starting to see why that doesn't work.
aciswhat 6y ago
Fusion
silentsea90 6y ago
Way outside my field, but a platform for global climate management that links current actions to future outcomes and accountability to nations and companies would be very cool, assuming it is taken seriously by parties involved (which is an even harder challenge). This is technically hard, but harder still from a people perspective.
runawaybottle 6y ago
No one seems to trust vpns. I wonder if a VPN company almost needs to be law technology. All of our attempts at privacy is technically sound but lacks the backing of solid lawyers ready to hold it down. That’s got to be worth something.
- ta17711771 6y ago
  So the implied trust due to legal backing through same ownership as a law firm? Interesting.
- steviekong 6y ago
  VPN's run through a company will always be subject to monitoring. I don't see a way around this other than something similar to TOR or mixnets.
sapientiae3 6y ago
A Laser Mosquito Turret for homes, that uses a directional microphone to home in on the target.I don’t believe that it would be difficult to build the laser part, but the difficulty would be in making it safe for home use.
- toomuchtodo 6y ago
  https://photonicsentry.com/ https://photonicsentry.com/
  - csomar 6y ago
    Isn't that a safety-hazard. The laser is probably strong enough to make you blind if it hits you wrongly in the eye.
    - toomuchtodo 6y ago
      One would assume you’d endeavor to keep humans out of the kill box when the system is active.
    - muzani 6y ago
      It's for industrial uses. There are more dangerous things, like chainsaws and microwaves. You either design it to be safer (microwave) or keep humans out of the way (chainsaw). It sounds like parent commenter is working on the former.
  - giantg2 6y ago
    Awesome
qlk1123 6y ago
I have two regarding embedded software stack, 1. Memory Safety C-language literally dominates embedded software stack, while every on knows it is not memory safe and terrible memory bugs will be found but not yet discovered. 2. Compile time The whole GNU toolchain take minutes if not hours to make, which makes CI/CD painfully long.
p1esk 6y ago
Making large transformer based models like Jukebox [1] efficient enough so that they can be trained on a single 8x GPU machine to the same level of quality as the original model, in the same amount of time. [1] https://openai.com/blog/jukebox/ https://openai.com/blog/jukebox/
Dirak 6y ago
An alternative to the html model for building websites. The html model is antiquated for building web apps, and inconsistent across various browsers x platforms. The direction of the spec imo is becoming an overengineered mess that is increasingly controlled by the Google (see portals and web packages). With the arrival of webgpu, I predict we'll begin to see end to end frameworks for building web apps that own the entire pipeline of application state management to drawing pixels on the screen.
literallycancer 6y ago
Typed language with extremely fast compile times suitable for use by neural nets.

20 points by astro-codes 6y ago

manx 6y ago

Creating a "Wikipedia of Arguments", one can refer to in public discussions. It turns out that this is a way harder problem than I thought. I'm in this space for some time now and am happy to talk about it. Relevant links: https://github.com/canonical-debate-lab/paper/blob/master/README.mediawiki https://github.com/canonical-debate-lab/paper/blob/master/RE... https://www.societylibrary.org/ https://www.societylibrary.org/

jka 6y ago
Let's hope that a well-constructed, well-behaved and well-participated argument forum maintained over long enough timescales can lead towards better outcomes for everyone. Best of luck with the project. You might be aware of it already, but just in case, the most progress I've seen in the area of web-based argument mapping has been Arguman ( https://github.com/arguman/arguman.org https://github.com/arguman/arguman.org ). They're also following an open source model and there could be opportunities to co-operate and/or share ideas.
- manx 6y ago
  Thanks, I'm aware of it. There are quite a lot of tools in this space already. Unfortunately, many are unaware of each-other and therefore making the same mistakes and learnings over and over again. Looking at all these tools makes it clear how difficult it is to create one that actually "works". Most of the researchers at the canonical debate lab created their own tool in the past and are now discussing how a next-generation tool should look like.
  - jka 6y ago
    Cool. In some ways, that sounds like the software industry in a nutshell, to be honest :) Can you share a sense for what any of the core unsolved problems are? (I'd be happy to read a mailing list / previous discussions if there's too much background to be worth communicating in a comment here) I've noticed that Arguman seemed to run into community management and spam issues. More broadly speaking there seems to be a societal challenge in getting people to trust and feel invested in honest debating and to accept best-known truths.
    - manx 6y ago
      The core unsolved problems in my opinion are: - The argumentation data model - An attack/manipulation-proof community curation system (maybe similar to StackOverflow) - A user-interface that can be used by anyone who is interested to contribute, but powerful enough to work with the argumentation structure There is a weekly zoom meeting every monday that is recorded and uploaded to youtube: https://www.youtube.com/playlist?list=PLwMyf-sRX2_Hqw-h9ba_StFt5jOKlpBcF https://www.youtube.com/playlist?list=PLwMyf-sRX2_Hqw-h9ba_S...
      - jka 6y ago
        Late reply here, but thank you - I'll try to catch up on a few of those recordings. Edit: small clarification
gitgud 6y ago
That is a brilliant paper, I expected it to be centralised around logical fallacies (which is how I view arguments/debates) but it goes into so much more detail around the nature of online debate. It's interesting to dive into the factors that influence online debates; reputation, trolling, anonymity, echo-chambers, the hive mind, language barriers, commercial interests.... There's so much more to an online discussion than just the argument presented...

HelpEveryone 6y ago

Spectacles that run 24/7

rapnie 6y ago

Decentralized search in p2p / federated networks. Privacy-respecting decentralized online identity.

ragerino 6y ago
I am sure you know about Solid. Just in case: https://solidproject.org/ https://solidproject.org/
- rapnie 6y ago
  Thanks. Yes, I do. There is a lot of work in the field of decentralized identity, but status is still far from a 'solved problem'. Work being done by Rebooting the Web of Trust [0] (see their github repo's) and Identity Foundation [1]. Besides Solid there's a bunch of interesting, related specs evolving like Verifiable Credentials [2], Self Sovereign Identity (I hope for impls without the blockchain). [0] http://www.weboftrust.info/ http://www.weboftrust.info/ [1] https://identity.foundation/ https://identity.foundation/ [2] https://www.w3.org/TR/vc-data-model/ https://www.w3.org/TR/vc-data-model/

_alex_ 6y ago

The "cold-start" problem of new social network sites

silentsea90 6y ago
Apologies for the condescension, and with all due respect, building a social network is a bit neo Silicon valley frothy.
- cweagans 6y ago
  the idea of building another facebook is just...ugh. however, a friend of mine has been working on a community site for fishing enthusiasts for years. it's _extremely hard_ to bootstrap a community site of any kind, even if it's not "social networking" per se.
  - muzani 6y ago
    I think the idea is to build a not-facebook. I'm working in social media too, and there's a lot of things that FB does poorly. The goal is usually media, centered around yourself. For some people, it could be their line of work, or happenings in their neighbourhood. FB, prioritizing growth, is like the BuzzFeed of social media, but we're starting to see why that doesn't work.

aciswhat 6y ago

Fusion

silentsea90 6y ago

Way outside my field, but a platform for global climate management that links current actions to future outcomes and accountability to nations and companies would be very cool, assuming it is taken seriously by parties involved (which is an even harder challenge). This is technically hard, but harder still from a people perspective.

runawaybottle 6y ago

No one seems to trust vpns. I wonder if a VPN company almost needs to be law technology. All of our attempts at privacy is technically sound but lacks the backing of solid lawyers ready to hold it down. That’s got to be worth something.

ta17711771 6y ago
So the implied trust due to legal backing through same ownership as a law firm? Interesting.
steviekong 6y ago
VPN's run through a company will always be subject to monitoring. I don't see a way around this other than something similar to TOR or mixnets.

sapientiae3 6y ago

A Laser Mosquito Turret for homes, that uses a directional microphone to home in on the target.I don’t believe that it would be difficult to build the laser part, but the difficulty would be in making it safe for home use.

toomuchtodo 6y ago
https://photonicsentry.com/ https://photonicsentry.com/
- csomar 6y ago
  Isn't that a safety-hazard. The laser is probably strong enough to make you blind if it hits you wrongly in the eye.
  - toomuchtodo 6y ago
    One would assume you’d endeavor to keep humans out of the kill box when the system is active.
  - muzani 6y ago
    It's for industrial uses. There are more dangerous things, like chainsaws and microwaves. You either design it to be safer (microwave) or keep humans out of the way (chainsaw). It sounds like parent commenter is working on the former.
- giantg2 6y ago
  Awesome

qlk1123 6y ago

I have two regarding embedded software stack, 1. Memory Safety C-language literally dominates embedded software stack, while every on knows it is not memory safe and terrible memory bugs will be found but not yet discovered. 2. Compile time The whole GNU toolchain take minutes if not hours to make, which makes CI/CD painfully long.

p1esk 6y ago

Making large transformer based models like Jukebox [1] efficient enough so that they can be trained on a single 8x GPU machine to the same level of quality as the original model, in the same amount of time. [1] https://openai.com/blog/jukebox/ https://openai.com/blog/jukebox/

Dirak 6y ago

An alternative to the html model for building websites. The html model is antiquated for building web apps, and inconsistent across various browsers x platforms. The direction of the spec imo is becoming an overengineered mess that is increasingly controlled by the Google (see portals and web packages). With the arrival of webgpu, I predict we'll begin to see end to end frameworks for building web apps that own the entire pipeline of application state management to drawing pixels on the screen.

literallycancer 6y ago

Typed language with extremely fast compile times suitable for use by neural nets.