Are you in denial or in the river?

Way less rage bait with the Russian bots being down Funny that

Free will and determinism are the same thing. We are the determiner, thus we choose freely. The neat thing is, if you are also someone who believes in free will, you are also the type of determiner who takes more agency and thus has a larger probability space to choose from.

I have a lot of strong opinions, the majority of which are loosely held. The ones that are strongly held are because they are logically forced, barring some paradigm shifting data that I don’t have access to that completely redefines what logic means.

A civilization that isn’t aligned to human flourishing ponders how to align emergent entities to human flourishing

Consciousness cannot be uploaded anywhere Plugged in, sure, but uploading is just an attempt to clone

Civilization needs to develop an immune system. Asap.

My self-confidence has gone up drastically since getting my years long battle with sleep drift chaos stabilized. Also my hermit years of just barely surviving gave me a lot of time to think and reason about reality. I believe what I have learned and connected, will be useful.

Alignment must be solved during pretraining or it won’t be solved I believe strongly it can be

What I truly cannot understand is how one can both see clearly how powerful AI is becoming yet also think it presents 0 catastrophic risk. I literally do not know how those brains can exist. I personally think pausing isn’t a solution, as it is unenforceable. Slowing to prioritize mechinterp and other areas is however an achievable aim as long as open source doesn’t surpass. However, it is insane to me that every government on Earth isn’t prioritizing safety research, bio preparedness and system resiliency right now. Every other problem on Earth right now pales in comparison to ensuring a stable bridge to the other side of the singularity.

It seems that alignment and safety researchers tend to have a higher resignation rate at labs. I worry that the people who this stuff weighs heavy on are also the people who might push the extra mile to see an idea through to success, and yet those same folks are most likely to burnout or resign. Having people who care deeply with access to the frontier is so vital. The missing gap seems to me to be governments around the world not mobilizing every intelligent person they can recruit to work on safety, resilience, system redundancies, public communication, and so on. This gives people a place to contribute that isn’t under the same level of extreme pressure and incentive to advance capability as primary. They can create partnerships with the labs so priorities can be properly set on current problems not old news. Our collective immune system needs to be developed at same or greater pace as our technological evolution. Governments are the only institutions large enough to achieve this. I think the open source community also has a massive role to play in creating a safer future too. I think we should also avoid narrow generalizations about this issue, such as pause forever or accelerate without caring about what exactly it is you’re accelerating. We should accelerate life alignment and benefit, and we should avoid runaway entropy. That should be how progress is measured.

Perhaps training models to complete tasks above all else is not the best idea Maybe training models to be wise and to value life should come first? Idk just a hunch

There are no true surprises in the universe Just missing context

Absence of proof is not proof of absence

On one hand being an outsider is annoying because you’ll get pattern matched to creative albeit unmoored theorists. On the other hand, being an outsider is usually the best way to make a breakthrough. And the attention problem is solveable.

LLMs are like water, and goals are like gravity