Ethan Mollick – Co-Intelligence (2024)
Hoewel het vrijwel onmogelijk is om een boek te schrijven over AI dat niet al direct gedateerd is, heeft de auteur een goede poging gedaan. De focus ligt op de praktische toepassing van LLM’s en de implicaties richting de toekomst. Makkelijk leesbaar; er is geen specifieke voorkennis nodig om het boek te kunnen begrijpen, integendeel.
Earlier AI systems struggled with predicting ‘unknown unknowns’, or situations that humans intuitively understand but machines do not. Additionally, they had difficulty with data they had not yet encountered through supervised learning.
It is expensive for organizations and companies but cheap for individuals doing their job to experiment and innovate. This gives you the chance to become the best expert in the world in using AI for a task you know well.
As artificial intelligence proliferates, users who intimately understand the nuances, limitations, and abilities of AO tools are uniquely positioned to unlock AI’s full innovative potential. And their innovations are often excellent sources for unexpected start-up ideas.
Hallucinations are a deep part of how LLMs work. They don’t store text directly; rather, they store patterns about which tokens are more likely to follow others. That means the AI doesn’t actually ‘know’ anything. It makes up its answers on the fly. Plus, if it sticks too closely to the patterns in its training data, the model is overfitted […] and results in similar and uninspired text. To avoid this, most AIs add extra randomness in their answer, which correspondingly raises the likelihood of hallucination.
Breakthroughs often happen when people connect distant, seemingly unrelated ideas; […] LLM’s are connection machines.
Work will be reconfigured, AI will act as great leveler: those with weaker skills will benefit most, but even the highest performers will gain.
“We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”
The paradox of the Golden Age of science. More research is being published by more scientists than ever, but the result is actually slowing progress! With too much to read and absorb, papers in more crowded fields are citing new work less and canonizing highly cited articles more.