Previously: eye contact and puppy slugs (w/e 25 June)
14.56, Friday 25 Jun 2021 Link to this post
Two recommended blog posts from the archives, originally published this week in years past.
1 year ago
Early web videos, eye contact, and anti-attention (22 June 2020).
How about a pair of augmented reality glasses with an app to manipulate everything I see, ensuring that no-one, no matter how charismatic, could hold my gaze for longer than 3.2 seconds?
The idiom of YouTube vlogging is straight to camera, eyes locked, and that was invented back in 2006. Zoom is all about eye contact too. But eye contact is a unconscious engagement amplifier (and also fatiguing) – wouldn’t it be cool to have an anti-eye-contact feature built into the computer?
6 years ago
Filtered for computers and birds (19 June 2015).
“We ask the network: ‘Whatever you see there, I want more of it!’ This creates a feedback loop: if a cloud looks a little bit like a bird, the network will make it look more like a bird. This in turn will make the network recognize the bird even more strongly on the next pass and so forth, until a highly detailed bird appears, seemingly out of nowhere.”
It turns out that, approximately 6 years ago, we saw the announcement of DeepDream (Wikipedia), Google’s computer vision/image generation/dreaming AI technique that created photorealistic images out of thin air, with a tendency to diverge into trippy fractelesque image montages in which bizarre animals could be found worming their way out of the corners: puppy slugs.
The door was opened: the rekindling of AI as something that could operate in the human realm, deep fakes and the undermining of reality… I can’t believe it has only been 6 years. It’s great to look back on some of those early pieces and see what an impact it had.
Personal favourites selected from this week’s On This Day archive spelunking page. This is an experiment to see how to best surface older ideas in the current feed in a meaningful way, and I’m trying it as a regular Friday feature. Keep-going/why-not-try-this-instead feedback welcome.