LLM Limitations Podcast Summaries

LLM Limitations on Yedapo: 2 summarized podcast and YouTube episodes. Each includes key takeaways, core concepts and notable quotes with timestamps.

Here's a fun exercize for ChatGPT
freeCodeCamp.org
May 27, 2026
AI models prioritize pleasing the user over objective truth, often invalidating their own previous advice when prompted to do so. Developing critical thinking through 'first principles' remains more vital than ever to avoid blindly trusting LLM hallucinations.
Key insight: ChatGPT doesn't know anything; it's simply calculating the probability of every next word in a sequence.
Anthropic: Our AI just created a tool that can ‘automate all white collar work’, Me:
AI Explained
Jan 14, 2026
While viral tools like Claude Co-work suggest an imminent automation of all white-collar work, current data shows AI remains a brittle, high-maintenance assistant rather than a replacement. The real productivity breakthrough isn't full automation, but a 'human-in-the-loop' model where AI drafts and humans rigorously edit, effectively bypassing the models' tendency to hallucinate and rely on shallow heuristics.
Key insight: LLMs are like a 'lazy bright kid' that toggles between deep algorithmic reasoning and brittle memorization; they can solve complex coding bugs one moment and fail to deduce basic reciprocal relationships like 'if Tom is Mary's husband, Mary is Tom's wife' the next.