Understanding The Past Year in AI: 200th Episode Special!
Limitless: An AI Podcast
3 DAYS AGO
Understanding The Past Year in AI: 200th Episode Special!
Understanding The Past Year in AI: 200th Episode Special!

Limitless: An AI Podcast
3 DAYS AGO
In this milestone 200th episode, the hosts reflect on the past 14 months of AI development, analyzing their own transcripts to identify key trends and investment themes that have shaped the landscape.
The hosts celebrate the validation of their early theses, particularly that compute and memory are foundational to AI growth. They highlight massive returns from memory chip stocks like Micron and SK Hynix, and the successful prediction of Cursor's acquisition by SpaceX. The discussion then shifts to emerging trends, including the rise of specialized AI applications in sectors like drug discovery and finance, the potential for training models in space, and the growing importance of local, hyper-optimized small models to reduce costs. They also note a significant shift in podcast mentions, with Anthropic surpassing OpenAI and a move from vague concepts like AGI to concrete topics like robotics. Looking ahead, the hosts are most excited about how intelligence will be turned into real-world value, citing robotics and energy solutions as key areas, while also acknowledging the rising geopolitical risks from China's potential to close-source its AI models.
00:00
00:00
Massive growth in AI companies and shift to AI-generated code
01:50
01:50
Compute, energy, and chips are critical
03:45
03:45
Memory is expensive and crucial for AI.
04:59
04:59
Cursor was acquired by SpaceX for $60 billion
05:34
05:34
Valor Atomics achieved a 13x multiple
07:46
07:46
Energy needs driving nuclear investment
09:33
09:33
Anthropic mentions quadrupled, surpassing OpenAI in episode references.
13:08
13:08
Google's vertical stack includes TPUs and Gemini
13:49
13:49
AI competition shifts to OpenAI and Anthropic
15:33
15:33
Specialized AI could achieve scientific breakthroughs.
17:47
17:47
Intelligence implementation is the key challenge
20:15
20:15
Hyper-optimized small models will become more abundant.
22:45
22:45
China will close-source its AI models.
24:57
24:57
AI can democratize innovation