One of the things I love about AI is its ability to help me slowly morph and regain control over my algorithmic feeds. I've analyzed about 1000 videos on my favorite topics, and YouTube is slowly becoming much higher quality because of it.
The channels floating to the top rank highly in terms of measured Bayesian/scientific language. Still surprised to see Karpathy currently #1 with 10 videos analyzed.
This is the kind of demonstration of AI's worst capabilities - elimination of mistakes. We need to make mistakes in order to learn and this kind of tech while cool only makes true failure more common. And that's true of many domains with similar examples.
What's the actual batting average on doomers? It has to be incredibly low right? From Malthus onward these guys have the most horrendous predictive capabilities. So why are they put on podcasts?
.@So8res thinks it’s more likely than not AI will wipe out humanity by 2036. How does that belief affect his choices in life? He doesn’t save for retirement, but beyond that, he says, “I sort of do what I can and then I live a good life.”
We talked through all the ways we should try to lower those odds. Watch here: youtube.com/watch?v=n8xpeNhj…
Other things about 1 in 4 people believe:
Chemtrails, sun revolving around the earth, fake moon landing, the living and dead can communicate, Clairvoyance, astrology, witches are real, reincarnation, ancient aliens
Why should society entertain the fringe?
AI hit critical mass with tech nerds. I surveyed my entrepreneur group, and people barely even use AI as a chat app.
I built a beautiful CRM and I am struggling to get people to use it over Gmail "because it's another app".
AI still requires power users.
It will never not hurt me that the internet has almost entirely killed people's joy at dancing in public. It's one of the most fundamental expressions of being human and it's basically dying off because of weirdos.
To be honest considering how easy it is to "vibe code" these days there's still surprisingly less major incidents than I would have thought by now. Sure people point to API key leaks but those embarassing human flaws have happened with devs in the past.
There's a fair bit of truth to this. The vast majority of people in the top of their field are either people who ship/action or people who inspire others via action.
They're usually above average but rarely the top intellectual in their field.
GCP can be brutal in the bills it can rack up. Vibe code responsibly jesus. Keep very strict limits and absolutely do not push publicly if you don't fully understand what you're doing. Or at the very least haven't thought of the consequences in the worse case scenario.
We live in this beautiful period where people attempt to flex their ability to detect AI. I fear it will only last for a few more years at most though.
And then what? Many will still consume but a decent minority will aggressively move to support in person art/events imo.
These scenarios come far far after AI has completed the more terrifying prospect AI brings which is the end of all socialization. Facebook et. al have basically publicly stated it.
AI's potential is not just a sexbot, but a weapon ready to rewire your brain's reward centre.
There are over 2,000 cover versions of "Louie Louie" and they're mostly forgotten to time. AI remains a tool, what survives will still be original ideas. It's highly unlikely the most popular creations with AI don't come with a heavy dose of originality.
We’re in a world where copying is trivial: but no one really cares much about copycats of the same stuff (games, videos, even social media posts)
Creating new things that stick is what remains hard. A novel but fun game with a captivating story (eg recently: Twisted Tower for me); an interesting video, a good book…
Yes AI is getting more impressive, but yet I’m just not that impressed with a lower quality copy of the original























