Vitalik Buterin pfp
Vitalik Buterin
@vitalik.eth
The amount of AI-generated deepfakes that we have seen so far is honestly far lower than what I or most people would have predicted we would see at current capability levels. https://twitter.com/GarrisonLovely/status/1729977755826315473
29 replies
14 recasts
118 reactions
Venkatesh Rao ☀️ pfp
Venkatesh Rao ☀️
@vgr
I think major public figures would respond too quickly to credibly debunk and mid-tier ones aren’t worth the effort. I recall when Pelosi fake circulated it was more “look what’s possible” than people getting taken in. You need a weirdly isolated, non-leaky target to influence with high reach/low detectability
1 reply
0 recast
2 reactions
chick3n.eth
@chick3n
Do deepfakes actually have that much power? How many people are deceived by it?
1 reply
0 recast
1 reaction
Sam (crazy candle person) ✦ pfp
Sam (crazy candle person) ✦
@samantha
You probably don’t use TikTok, but imo its very common. People usually use one artist to sing another famous artists’ song, and it gets like hundreds of thousands of likes before it gets taken down. I haven’t seen many use cases for ill intent, but I sense it’s coming
0 reply
0 recast
3 reactions
Mac Budkowski ᵏ pfp
Mac Budkowski ᵏ
@macbudkowski
Do you have any ideas why?
1 reply
0 recast
3 reactions
Idan Levin 🎩 pfp
Idan Levin 🎩
@idanlevin
There might be some bias here because the high-quality deep fake is so good it goes unnoticed, and also people are ashamed to report they got scammed, so there will always be an underestimation of this trend
1 reply
0 recast
1 reaction
Yassine Landa pfp
Yassine Landa
@yassinelanda.eth
The calm before the storm, elections will be a catalyst... Still sticking to my predictions for 2025.
0 reply
0 recast
1 reaction
Nico Gallardo 🍄 pfp
Nico Gallardo 🍄
@nicnode
how sure are you about that? 👀
0 reply
0 recast
0 reaction
:omer pfp
:omer
@omer
we probably underestimate the distance between an “almost convincing” deepfake and a “very hard to tell apart & convincing” one, and it takes some more time to produce the latter
0 reply
0 recast
0 reaction
shazow pfp
shazow
@shazow.eth
I forget where I heard this (some podcast) but apparently there are voice cloning agencies, basically have voice actors all over the world who can imitate a catalogue of public figures on demand. Not sure how much work they get but they'll probably be impacted more than most.
0 reply
0 recast
0 reaction
Tarun Sachdeva pfp
Tarun Sachdeva
@tarun
1. Not sure we have a reliable way to measure without understanding private social networks (eg WhatsApp forwards) where network trust is higher so misinformation spreads easily 2. In public spheres - Even “real” videos are being questioned, esp in a polarized world. In short - Consumers are getting smarter.
0 reply
0 recast
0 reaction
tinyrainboot 🎩 pfp
tinyrainboot 🎩
@tinyrainboot
just wait...
0 reply
0 recast
0 reaction
pfp
@na
a totally exhausted world holds little tendency of crime
0 reply
0 recast
0 reaction
borged.eth pfp
borged.eth
@borged.eth
This could be because the bottleneck for disinformation is distribution not generation. https://www.aisnakeoil.com/p/is-ai-generated-disinformation-a
0 reply
0 recast
1 reaction
wizard not parzival pfp
wizard not parzival
@shoni.eth
Haven’t heard of any related fraud yet
0 reply
0 recast
1 reaction
Raven Lunatic pfp
Raven Lunatic
@ravenlunatic
Personally I think it's a combination of people are lazy and also don't know it's that cheap and easy. I just don't know the ratio. Once they figure it out, it's going to be interesting.
0 reply
0 recast
0 reaction
borg pfp
borg
@borg79
💯
0 reply
0 recast
0 reaction
!196525
@
Turns out, most people in the world are good. Who knew!
0 reply
0 recast
0 reaction
ant pfp
ant
@0xant
maybe it's just that we haven't encountered an event in our daily life that requires that much of economic incentive to do a complete high-quality deepfake video
0 reply
0 recast
0 reaction
Mr.Four pfp
Mr.Four
@004
it's a long way.....
0 reply
0 recast
0 reaction