I was going to make a joke about mimicking that behaviour using a perceptron with certain weights but then I realized there's no side effect so they can all be zero. Perfect.
I think it's fair to say the problem space AI operates in is one of case statements and not general rationalization. The magic of "AI" is in working with subjective data to identify the best case to use.
AI/ML is so abused these days, we stopped using them entirely while pitching our Startup even though we wrote two home-grown algorithms already. It has become an addendum, "We also wrote our own Machine Learning Algorithm and we train them against 1-million acres of high resolution satellite data."
It’s kind of absurd that a number of the algorithms being used are remarkably simple and are being compared to sentient consciousness by people with minimal academic background and datasets that don’t contain the value for the proposed solution.
Shameless plug but if you’re looking for ideas which are drastically less ambitious ones along which have a startling lack of AI/ML, I have a few ideas[0][1]
Half of these projects are so general that are laughable, a quarter is so ambitious that a small prize would play 0 role in the potential development of the solution.
Less:
Solve world peace.
Create "dynamic" organizations.
Establish a colony in Titan.
More:
- Increase by Z the yield of X staple in Y country within the next 5 years.
- Integrate 1 million people to the Internet for less than 1 USD/month/person.
The whole idea is for you to come up with a project, that's whythey are general. Not sure why people need detailed instructions to innovate and get funded.
Your "more" section lacks a purpose. You mention price of "connection" without additional details. One could argue that the main benefit of the internet, both ways communication of information worldwide, can be fulfilled with very low bandwidth and today can already be priced at 1 dollar per month per person.
What happens when we allocate resources to enable 1 dollar per month per person internet with megabits of bandwidth and this group of people use it to consume social media and streaming (aka "old days television") old day long?
Yeah, Wikipedia loads at low speeds[0] but I think they specifically design to load on low bandwidth connections.
RANT
Most sites are built for high-end dev machines running OSX and/or chrome. I’m basically the only person running windows and Linux regularly at work and test things in windows for people when they can’t reproduce it anywhere else.
The number of times I’ve informed our editorial team that they are about to publish something full of unprintable characters (that are hidden in OSX but show up as a question mark on Windows) is non-zero. The dev mono-culture is real.
Worth highlighting the last item, since guessing most won’t make it that far:
>> Critiquing our approach
>> Research That Can Help Us Improve
>> We’d love to fund research that changes our worldview—for example, by highlighting a billion-dollar cause area we are missing—or significantly narrows down our range of uncertainty. We’d also be excited to fund research that tries to identify mistakes in our reasoning or approach, or in the reasoning or approach of effective altruism or longtermism more generally.
In the same way dropping foreign aid on a country to ‘solve’ hunger can make the problem worse, EA money could distort market forces if it became big enough. I suspect it’s very difficult to find investments that return more net good than standard businesses.
#5 Biological Weapons Shelters: I wonder how many of these already exist that are just classified. Or if an alternative is just to invest in SpaceX or some other team with a vision of making humanity a multi-planetary species (or long term self-sustaining space habitats)
our = US? I think FTX's concern is humanity broadly so I hope some country somewhere has a decent strategy. I know a few EU countries had lots of bunkers, no idea if any of them have been made into bioweapon bunkers
Looking at their grants page¹ this organisation has committed only a single grant under $50k and various grants over $1m, where does 'tiny amounts of money' come from?
For some reason I get an almost misanthropic vibe from a lot of these. Maybe I’m missing some context where they support eliminating global poverty and such in parallel, but it feels too much like the subtext is that they don’t feel that all the suffering and injustice in the world matter long term.
If humanity plays its cards right, there will be trillions and trillions of humans living in the future, all over the galaxy. Everyone who lives right now, or has ever lived, is barely noticeable compared to that.
I don't see any flaws in this argument, we all are on a fast-track to getting killed by AI, and the smartest people on the planet don't have a plan on how to fix it. "Build an AI that burns all GPUs to prevent all the future AIs, and hopefully doesn't kill us all in the process" is the best idea they've got.
If you believe that we're about to kill trillions and trillions of potential future human beings, these priorities make sense.
> Everyone who lives right now, or has ever lived, is barely noticeable compared to that.
> If you believe that we're about to kill trillions and trillions of potential future human beings, these priorities make sense.
Yeah, sorry. This justifies basically any human atrocity. As far as I’m concerned that’s a case of “Proving too much” and therefore the associated logic must be wrong.