Hacker News Reader

Comment by jagrsw | Hacker News Reader

Some first ideas coming to mind:

Engineering Level:

  Solve CO2 Levels
  End sickness/death
  Enhance cognition by integrating with willing minds.
  Safe and efficient interplanetary travel.
  Harness vastly higher levels of energy (solar, nuclear) for global benefit.

Science:

  Uncover deeper insights into the laws of nature.
  Explore fundamental mysteries like the simulation hypothesis, Riemann hypothesis, multiverse theory, and the existence of white holes.
  Effective SETI

Misc:

  End of violent conflicts
  Fair yet liberal resource allocation (if still needed), "from scarcity to abundance"

View on HN

The problem with CO2 levels is that no one likes the solution not that we don't have one. I highly doubt adding AGI to the mix is going to magically make things better. If anything we'll just burn more CO2 providing all the compute resources it needs.

People want their suburban lifestyle with their red meat and their pick-up truck or SUV. They drive fuel inefficient vehicles long-distances to urban work environments and they seem to have very limited interest in changing that. People who like detached homes aren't suddenly affording the rare instances of that closer to their work. We burn lots of oil because we drive fuel inefficient vehicles long distances. This is a problem of changing human preferences which you just aren't going to solve with an AGI.

Assuming embedded AI in every piece of robotics - sometimes directly, sometimes connected to a central server (this is doable even today) - it'll revolutionize industries: human-less mining, processing, manufacturing, services, and transportation. These factories would eventually produce and install enough solar power or build sufficient nuclear plants and energy infrastructure, making energy clean and free.

With abundant electric cars (at this future point in time) and clean electricity powering heating, transportation, and manufacturing, some AIs could be repurposed for CO2 capture.

It sounds deceptively easy, but from an engineering standpoint, it likely holds up. With free energy and AGI handling labor and thinking, we can achieve what a civilization could do and more (cause no individual incentives come into play).

However, human factors could be a problem: protests (luddites), wireheading, misuse of AI, and AI-induced catastrophes (alignment).

Having more energy is intrinsically dangerous, though, because it's indiscriminate: more energy cannot enable bigger solutions without also enabling bigger problems. Energy is the limiting factor to how much damage we can do. If we have way more of it, all bets are off. For instance, the current issue may be that we are indirectly cooking the planet through CO2 emissions, so capturing that sounds like a good idea. But even with clean energy, there is a point where we would cook the planet directly via waste heat of AI and gizmos and factories and whatever unforeseen crap we'll conjure just because we can. And given our track record I'm far from confident that we wouldn't do precisely that.

This exactly. Every self replicating organism will eventually use all the energy available to it, there will never be an abundance. From the dawn of time, mankind has similarly used every bit of energy it generates. From the perspective of a subsistence farmer in the 1600s, if you told them how much energy would be available in 400s year they would think we surely must live in paradise with no labor. Here we are, still metaphorically tilling the land.

The incentives aren't structured properly for these things to happen, it has always been a sci-fi fairy tale that AGI would achieve these things.

Do you believe the average human has general intelligence, and do you believe the average human can intellectually achieve these things in ways existing technology cannot?

Yes, considering that AI operates differently from human minds, there are several advantages:

  AI does not experience fatigue or distractions => consistent performance.
  AI can scale its processing power significantly, despite the challenges associated with it (I understand the challenges)
  AI can ingest and process new information at an extraordinary speed.
  AIs can rewrite themselves
  AIs can be multiplicated (solving scarcity of intelligence in manufacturing)
  Once achieving AGI, progress could compound rapidly, for better or worse, due to the above points.

The first AGI will probably take way too much compute to have a significant effect, unless there is a revolution in architecture that gets us fast and cheap AGI at once the AGI revolution will be very slow and gradual.

A model that is as good as an average human but costs $10 000 per effective manhour to run is not very useful, but it is still an AGI.

> A model that is as good as an average human but costs $10 000 per effective manhour to run is not very useful, but it is still an AGI.

Geohot (https://geohot.github.io/blog/) estimates that a human brain equivalent requires 20 PFLOPS. Current top-of-the-line GPUs are around 2 PFLOPS and consume up to 500W. Scaling that linearly results in 5kW, which translates to approximately 3 EUR per hour if I calculate correctly.

That is if the first model we make is as efficient as a human.

now do magical flying unicorn ponies, which I understand there is also considerable demand for

Reply on news.ycombinator.com