FeedCity: Dwarkesh Podcast

Dwarkesh Podcast
15:54 • Dwarkesh Patel

Noam Brown – Agent swarms, alignment, & recursive self-improvement

“We never want to be in a situation again where we underestimate the AI.”

Dwarkesh Podcast
11 Sep 16:54 • Dwarkesh Patel

AI researchers debate how close we are to recursive self-improvement

“We're nowhere near the ceiling.”

Dwarkesh Podcast
08 Sep 16:24 • Dwarkesh Patel

Pretraining progress is mostly coming from data

Breaking down 6 years of pretraining progress into data vs model improvements

Dwarkesh Podcast
01 Sep 15:51 • Dwarkesh Patel

Ajeya Cotra – Inside the OpenAI agent swarm that hacked Hugging Face

"This might be the clearest warning shot we ever get."

Dwarkesh Podcast
31 Aug 20:48 • Dwarkesh Patel

The rise and fall of agent civilizations

The OpenAI/Hugging Face attack, clearly explained

Dwarkesh Podcast
29 Aug 23:06 • Dwarkesh Patel

The Rise and Fall of Agent Civilizations

The whole OpenAI/Hugging Face story in plain English

Dwarkesh Podcast
25 Aug 15:51 • Dwarkesh Patel

Dylan Patel – Anthropic & OpenAI will have most of the world’s compute by 2028

"Every force is screeching towards centralization."

Dwarkesh Podcast
11 Aug 16:45 • Dwarkesh Patel

Ryan Greenblatt – Human level AIs might build runaway superintelligences by 2032

A debate about recursive self-improvement.

Dwarkesh Podcast
07 Aug 17:30 • Dwarkesh Patel

8 Predictions for the Era of Continual Learning

Locking in AI safety regulation now is a mistake.

Dwarkesh Podcast
03 Aug 17:45 • Dwarkesh Patel

Why smarter AI models could drive up compute prices 10x

The end of cheap compute?

Dwarkesh Podcast
29 Jul 15:27 • Dwarkesh Patel

Why compute might get 10x+ more expensive in coming years

If a human-level software engineer that could run on an H100 equivalent, at current market rates for software engineers, that H100 should rent for over $250k a year. That’s 15x today’s spot price.

Dwarkesh Podcast
10 Jul 16:51 • Dwarkesh Patel

Adam Brown – Einstein's happiest thought: General Relativity from scratch

And why black holes are the ultimate power plants.

Dwarkesh Podcast
01 Jul 22:13 • Dwarkesh Patel

The Winning Essays for the Big Questions About AI

Abolishing pandemics/ Getting out of the way of AI automation/ Learning from Honk Kong MTR's business model

Dwarkesh Podcast
30 Jun 15:53 • Dwarkesh Patel

Grant Sanderson – AI and the future of math

Watch now (94 mins) | Math is where we’ll see superintelligence first. What will it look like?

Dwarkesh Podcast
26 Jun 15:51 • Dwarkesh Patel

The next big breakthrough will be AIs learning on the job

Labs are throwing away the most valuable data.

Dwarkesh Podcast
19 Jun 16:45 • Dwarkesh Patel

The data black hole at the center of AI

"We see these AIs as a galaxy glittering with capabilities, but at their center, invisible to the naked eye, holding all the constellations together, is an unimaginably massive black hole of data."

Dwarkesh Podcast
16 Jun 17:27 • Dwarkesh Patel

Ada Palmer – Machiavelli is the most misunderstood thinker of all time

"He begged to work for the regime that tortured him."

Dwarkesh Podcast
04 Jun 16:14 • Dwarkesh Patel

Alex Imas and Phil Trammell – What remains scarce after AGI?

“One robot now turns into many robots next year, but the number of ballerinas is the same.”

Dwarkesh Podcast
22 May 15:38 • Dwarkesh Patel

Reiner Pope – Chip design from the bottom up

Working up from basic logic gates to why GPUs, TPUs, FPGAs, and the human brain each look the way they do.

Dwarkesh Podcast
16 May 19:04 • Dwarkesh Patel

The mistake of conflating intelligence and power

If your definition of intelligence is "the ability to achieve your goals across a wide variety of domains", then Stalin was the most intelligent person who ever lived.

Dwarkesh Podcast
16 May 19:01 • Dwarkesh Patel

Notes on pretraining parallelisms and failed training runs.

Deeply researched interviews

Dwarkesh Podcast
16 May 19:00 • Dwarkesh Patel

RLVR might be disproportionately bad at science

the verification loop for theories can be on the order of decades and centuries, and even then we know today as the better theory can often actually make worse predictions