Loops just happened to launch right before the US election,...

Loops just happened to launch right before the US election, and I've noticed some undesirable content

I will remove any content that violates our ToS and Community Guidelines, but we have a lot more work to do

Here are a few of my ideas:

- New mod tools
- New privacy/safety feature
- Moderation Appeals
- Insight from industry experts
- Feedback from at risk groups
- Parental feedback and tools

Feedback and boosts appreciated!

Loops will only succeed with a vibrant community, you in?

#loops

Like 9 November at 11:35 | Open on mastodon.social

7 comments

Cairo Braga

@dansup this might be tangential but may I suggest 2FA also for the security part?

9 November at 11:51 | Open on toot.cairobraga.com

dansup

@cairobraga working on this

9 November at 14:31 | Open on mastodon.social

GolfNovemberUniform

@dansup may I suggest adding people with straight sexual orientation as a risk group? Based on my experience, I would say it's necessary.

9 November at 12:52 | Open on infosec.exchange

Darren du Nord

@dansup Definitely in. Enjoying it, and contributing ideas and bug reports along the way.

9 November at 13:17 | Open on c.im

PEXAVC

@dansup what are your thoughts on some on-device models that can help detect harmful content? Thinking of creating a Swift Package to handle a suite of these cases, if you'd like, for the iOS variant.

10 November at 1:05 | Open on mastodon.social

dansup

@pexavc I would love that! I’ve been looking into MLkit for on device NLP

10 November at 1:10 | Open on mastodon.social

PEXAVC

@dansup Nice, I'll take a look at that too. I have been learning MLX, another Apple ML framework, and it works quite well for LLMs. Will see how that can be used with abusive language detection or fact-checking. Vision wise, I'll try to find properly licensed CoreML models to bundle with it all.

10 November at 1:13 | Open on mastodon.social