One of the issues with LLMs is that they still suffer from the hallucination problem, whereby they often confidently claim widely wrong information as accurate. This is doubly dangerous given they often are right, to an excerpt level. As a user, it’s all too easy to be lulled into a false sense of security and assume anything coming out of the system is true.

lk 243

Pildi allikas

The central problem for humanity in the twenty-first century is how we can nurture sufficient legitimate political power and wisdom, adequate technical mastery, and robust norms to constrain technologies to ensure they continue to do far more good than harm. How in other words, we can contain the seemingly uncontainable.

lk 228

Pildi allikas

Exponential change is coming. It is inevitable. The fact needs to be addressed.

lk 255

Pildi allikas

Make no mistake: standstill in itself spells disaster.

lk 220

Pildi allikas

That’s part of the problem; we don’t know what failure models are being introduced and how deep they could extend.

lk 211

Pildi allikas

AI is both valuable and dangerous precisely because it’s an extension of our best and worst selves.

lk 210

Pildi allikas

Even if it was fully aligned with human interests, a sufficiently powerful AI could potentially overwrite its programming, discarding safety and alignment features apparently built in.

lk 209

Pildi allikas

AI safety reacheres worry (correctly) that should something like an AGI be created, humanity would no longer control its own destiny. For the first time, we would be toppled as the dominant species in the known universe.

lk 209

Pildi allikas