Deepfake It's Not
Stop worrying about deepfakes ā§ How not to be stupid about AI ā§ ...Or AGI, for that matter
šĀ On time for your first weekend of the year: a round-up of remarkable storiesĀ at the intersection of technology, business, design, and culture. Three reads and three listens, and like last year: no fluff, just stuff ā”
The past is written, but the future is left for us to write. And we have powerful tools, openness, optimism, and the spirit of curiosity. All they have is secrecy and fear. And fear is the great destroyer āĀ Capt. Jean-Luc Picard
š Reading
Stop Worrying About Deepfakes:
One thing I came to appreciate more [ā¦] is the value of parody and satire in human communication. This is a very old format for making a social critique, often used quite strategically [ā¦] A lot of the internet is just like that. Itās awash in transgressive material that requires you to think a bit to get the message. The problem is, the feed-like nature of social media often causes us to stick to the surface-level message, so then you get these outcries.
Walter ScheirerāNautilus | 11 minutes
How Not to Be Stupid About AI:
We're missing something big to get machines to learn efficiently, like humans and animals do. We don't know what it is yet. I don't want to bash those systems or say theyāre uselessāI spent my career working on them. But we have to dampen the excitement some people have that we're just going to scale this up and pretty soon weāre gonna get human intelligence. Absolutely not.
Yann LeCunāWired | 17 minutes
The Road to Artificial General Intelligence:
There is no point where learning disparate skills āobject detection, speech synthesis, chess playing, ā¦ you name itā suddenly leads us toĀ learn how to learn. Thereās a qualitative jump there, in the same sense as making ever higher skyscrapers wonāt lead us to the moon. Itās not just a matter of scaling what has worked so far. You need a completely different theory.
Alejandro Piad MorfisāMostly Harmful Ideas | 19 minutes
š§ Listening
Why Can No One Think Rationally Anymore?
I think that humans run at the speed that they run at and I think you can overclock humans in the same way that you can overclock technology. You're just playing this game and we're going to move along. What you end up with, if you try and do it too quickly, is you end up with firehosing, which is the problem of overloading people with information.
George MackāModern Wisdom | 97 minutes
The Geopolitical Implications of A.I.
Silicon Valley is very good at imagining an abstracted idea of the future [ā¦] Silicon Valley fails to figure out how to take their technology products and technology innovation and coexist with humans. I think there is a distinct lack of understanding in pretty much every single company, bar maybe one or two [ā¦] We have very little understanding of how technology dovetails with actual human beings.
Om MalikāStuck@Om Podcast | 52 minutes
Humans and AI:
We love generative AI, but if it's context-less, it hallucinates [ā¦] There's questions and answers, but the answers depend upon the context of the question being asked. I'm excited about how we're asking questions about meaning: who's meaning, your meaning, long form, short form, medium form. And I love that, at least right now in late 2023, even if you have a giant token window, it only remembers the beginning and end, like a human being might. And so really, memory matters a lot.
John MaedaāWeaviate | 53 minutes
š Timeless
1ļøā£ year agoāHow to profit off AI
2ļøā£ years agoāThe Economics of Creativity: Who Gets Paid and Why
3ļøā£ years agoā8 Themes For The Near Future (Of Tech)