ted (not lasso)
@ted
3310 Following
251530 Followers
5 replies
1 recast
26 reactions
we used to rely exclusively on newspapers, radio, or letters for information. so limited, so slow. no way to “fact check” unless you were there on the frontlines.
now we have endless resources to access information, the ability to fact check in minutes if not seconds, and… the majority of the world just doesn’t. this is egregiously bad on tiktok (more on this below).
i know it is human nature to default to system 1 thinking (lol Daniel Kahnemann reference): fast, intuitive, emotional. i know that we tend to process information in ways that reinforce our group identity. and i know that emotion drives virality.
and so ofc i know that if we keep treating misinformation as an individual problem, this will only get worse.
platforms like groundnews are great, but that’s opt-in so it attracts users who already care about truth (unfortunately that’s magnitudes smaller TAM).
to me it is much more interesting to think of it as a platform problem. with AI, the design space could expand significantly (but obviously comes with its own risks).
two good built-in platform examples:
- twitter community notes on inaccurate tweets cut retweets by at least 50% and increased user-initiated deletions by 80% (but right now it is too slow, should be done within an hour not 12 hours!)
- twitter reminding users to read an article before they retweet actually reduced blind shares by ~33% (a little bit of friction can be a good thing!)
tiktok, however, hasn’t done anything really at all.
some labels for COVID, a politifact and snopes partnership i never saw, and a bullshit STEM feed for “reliable, vetted data” (hint: it was a paid partnership for two science brands to get guaranteed distribution).
the most dangerous feature of tiktok, imo, is its algo-driven echo chambers. it can ensure the content you see is what you want to see, that the comments are what you want to read.
it pushes users deeper and deeper into self-reinforcing narratives. you can scroll for hours and never be exposed to an alternative perspective on a single topic.
when users are exposed to entirely different information ecosystems, there’s zero foundation for shared discourse. this has got to be the most powerful, least visible force driving division today.
SAD!!! 27 replies
12 recasts
127 reactions
24 replies
9 recasts
166 reactions
9 replies
2 recasts
44 reactions
software as content is the most underexplored, highest potential opportunity right now.
what i'm seeing right now in terms of AI-generated videos and apps mirrors exactly what i observed and studied about tiktok's rise to dominance.
on the surface, it may seem like tiktok’s earliest wave was driven by dancing. in reality, it was driven by accessibility: it gave users a clear format, a song, a trend, a shared visual language.
everyone could easily participate. you didn't need to be famous or particularly good dancer. you just needed to show up, try it, and post it — and all you needed was your phone and the tiktok app.
tiktok lowered the barrier to creativity by making experimentation feel intuitive, fast, and social.
since those early days, the creative aperture has widened. we've seen an explosion of new video (and now carousel!) trends proliferate: GRWM videos, DIML videos, how it started vs. how its going, pepe the prawn king stories, man of the year, etc.
with AI, we're seeing the same pattern happen again but now across apps and media.
tools like suno, ohara, runway, replit, veo, etc. have turned content creation into prompt-based creative exploration and play. people are spinning up apps, webpages, music, videos, memes, etc. these are not polished projects, but expressive and creative experiments.
AI is doing for software and media what tiktok did for video content. it’s lowering the bar, speeding up iteration, and turning more people into creators. now anyone can create a software experience that is lightweight, interactive, and shareable without needing technical mastery.
AI is doing for software and media what tiktok did for video: lowering the bar, accelerating iteration, and turning spectators into creators.
and when you combine the two (generative video and generative apps), you start to see something powerful: built-in infrastructure for creativity and distribution.
we’ve already seen glimpses of this with tiktok effects: green screen, face morph, the do-re-mi filter. they were creative building blocks. effects like these sparked entire viral trends because they were easy to use and fun to remix.
AI tools today work the same way. a single model can kick off a wave of content. a new aesthetic, a format, a meme. we saw it with the studio ghibli-style AI videos that took over timelines. we’re seeing it now with suno tracks and animals-doing-olympic-sports becoming go-to sounds or videos for tiktok and reels.
and we also saw this happen on farcaster. remember when @jc4p did the builder alignment chart app? or when @aneri did the hogwarts sorting hat app?
early tiktok creators tested what worked using songs, dances, and effects. today’s AI-native creators are doing the same with prompts, tools, and interactive media.
the platforms are different and the tools are more powerful, but the behavior (experimentation, participation, and distribution) is the same.
we need to lean into existing user behaviors. 8 replies
7 recasts
64 reactions
5 replies
2 recasts
31 reactions
36 replies
51 recasts
312 reactions
10 replies
4 recasts
90 reactions
36 replies
13 recasts
159 reactions
8 replies
6 recasts
92 reactions
2 replies
2 recasts
27 reactions
6 replies
5 recasts
44 reactions
22 replies
15 recasts
128 reactions
14 replies
7 recasts
97 reactions
27 replies
10 recasts
127 reactions
11 replies
3 recasts
118 reactions
3 replies
2 recasts
23 reactions
9 replies
16 recasts
70 reactions
11 replies
3 recasts
125 reactions
12 replies
5 recasts
53 reactions