The Knowmads Podcast (@KnowmadsPodcast) header
The Knowmads Podcast (@KnowmadsPodcast) profile picture

Podcast by @garellaprachi and @bhavaytyagi | We talk about Science, Philosophy, Mathematics and everything else.

Joined August 2023
New Episode with @karch_andreas out now!
2
4
1
7
1,104
The Knowmads Podcast retweeted
I am back on Knowmads Podcast! Last time was about strings, this time about AI. Thanks for having me on!
1
5
16
876
The Knowmads Podcast retweeted
New Episode with @karch_andreas out now!
2
4
1
7
1,104
The Knowmads Podcast retweeted
When you start a PhD, they tell you: “Find a problem you’re passionate about, go deep, become an expert, and you’ll be fine.” Five years later, you enter the job market and discover that the one thing nobody is hiring for is your expertise.
19
37
3
712
29,465
The Knowmads Podcast retweeted
Thanks to @bhavaytyagi and @garellaprachi for inviting me to be the first ever return guest on the @KnowmadsPodcast I get to rant for an hour about AI in science … to appear soon
1
3
2
42
2,521
Presenting, The Return of the Karch. We really value the quality of our conversations and what we bring to the world. Can’t wait to share this next episode with everyone.
Thanks to @bhavaytyagi and @garellaprachi for inviting me to be the first ever return guest on the @KnowmadsPodcast I get to rant for an hour about AI in science … to appear soon
1
3
176
The Knowmads Podcast retweeted
Only way to summarise this past summer: The Summer AI Turned Pretty
🤖 Made with AI
2
3
1
10
433
The Knowmads Podcast retweeted
Replying to @rperezmarco
For someone who knows so much about mathematics, you seem remarkably unfamiliar with the concept of a logical argument. :)
1
1
11
375
The Knowmads Podcast retweeted
Replying to @TheOnesAncient
That’s only half the story. Simply stated, as training datasets grow, we lose more control over the learned weights, and increasingly complex emergent behaviors arise. Effectively taking the model outside convex hull!
1
1
4
404
The Knowmads Podcast retweeted
I agree, the “convex hull” argument seems rather uninformed and therefore pessimistic.
Replying to @Frances01896069
people who understood the NS solution told me that it contains 10 field-shaping breakthroughs as part of the solution. The idea that the models only produce results in the convex hull of human knowledge (whatever that means) is misguided.
1
3
1
31
8,060