LessWrong 2.0 Reader

Yes, by default.

I'd drop the inevitable word, but barring massive change like a technology regression or laws being passed that are very strong (and it surviving the AI era), near-perfect surveillance is very likely to happen, and the safe assumption is that nothing you do will be hidable/hidden by default.

I'm not addressing whether surveillance is good or bad.

I'd like to hear the arguments why you think perfect surveillance would be more likely in the future. I definitely think we will reach a state where surveillance is very high, high enough to massively increase policing of crimes, as well as empower authoritarian governments and the like, but I'm not sure why it would be perfect. 

It seems to me that the implications of "perfect" surveillance are similar enough to the implications of very high levels of surveillance that number 2 is still the more interesting area of research. 

A better way is to download it. See also Preserving and continuing alignment research through a severe global catastrophe [LW · GW]

Great stuff!

To expand on a few points:

1. The brain is less complex than people make it out to be. It is complicated, but there is clearly a logic to it, which we only need to discover. Often cited statistics about the numbers of cells and synapses include the cerebellum, which isnt really necessary for intelligence.
2. Evolution produces some beautiful designs, but if left to its own devices it takes longer than G.R.R. Martin takes to write a novel. According to the theory of punctuated equilibria: evolution spends long periods of time stuck in a stasis, when evolution essentially gets stuck in a local optimum.

I think I mostly understand what you're saying. Essentially, MONA training is:

1. strong model comes up with a step
2. weak model outputs a score
3. the step gets executed
4. repeat, starting from the new state

Your idea adds "1.5. weak model optionally asks the strong model follow-up questions" and "2.5. weak model optionally changes the step."

If you don't update the weak model at all, this seems fine. However, the strong model will eventually learn to come up with a step which explains itself enough that the weak model's questions are all preemptively answered, and which anticipates the weak model's preferences enough that the weak model never feels the need to change the step. So this is essentially equivalent to MONA.

However, if you do update the weak model, this seems problematic. The equilibrium is that the strong model suggests exactly the actions that maximize reward, and the weak model does exactly the actions that the strong model suggests. (Technically the weak model isn't even incentivized to score the strong model any particular way, but I'm just assuming it gives high scores for suggestions that it decides to carry out.)

I guess if you only train the weak model a small amount, it will do something in the middle, where it's kind of motivated by long-term reward and kind of motivated by its initial values. There's no guarantee on exactly what it does. I think the stipulation that the weak model remains safe "due to its architecture or whatever reason we trust it more" is doing a lot of work here; I'm not sure exactly what this would mean.

"How to do more things" - some basic tips 

Why do more things? Tl;dr It feels good and is highly correlated with other metrics of success. 

• According to Maslow's hierarchy, self-actualization is the ultimate need. And this comes through the feeling of accomplishment—achieving difficult things and seeing tangible results. I.e doing more things
• I claim that success in life can be measured by the quality and quantity of things we accomplish. This is invariant to your specific goals / values. C.f. "the world belongs to high-energy people". 

The tips

Keep track of things you want to do. Good ideas are rare and easily forgotten. 

• Have some way to capture ideas as they come. Sticky notes. Notebooks. Voice memos. Docs. Emails. Google calendar events. Anything goes.
• Make sure you will see these things later. Personally I have defaulted to starting all my notes in Todoist now; it's then very easy to set a date where I'll review them later. 

80/20 things aggressively to make them easier to do. 

• Perfection is difficult and takes a long time. Simple versions can deliver most of the value. This advice has been rehashed many times elsewhere so I won't elaborate.
• Almost all things worth doing have a shitty equivalent you can do in a day. or in a few hours. 

Focus on one thing at a time. Avoid 'spreading attention' too thin. 

• Attention / energy is a valuable and limited resource. It should be conserved for high value things.
• "What is the most important thing you could be doing?" ... "Why aren't you doing that?"
• I personally find that my motivation for things comes in waves. When motivation strikes I try to 'ride the wave' for as long as possible until I reach a natural stopping point.
• When you do have to stop, make sure you can delete the context from your mind and pick it up later. Write very detailed notes for yourself. Have a "mind like water".
• I've found voice recordings or AI voice-to-text very helpful along these lines.

Get feedback early and often. 

• Common pitfall: hoard idea and work in isolation.
• The more you 80/20 things the easier this will be. 

Why not just save them to an offline hard drive?

all at once in 2029.

I am almost totally positive that the plan is not that. 

If planning for 2029 is cheap, then it probably makes sense under a very broad class of timelines expectations.
If it is expensive, then the following applies to the hypothetical presented by the tweet:

The timeline evoked in the tweet seems extremely fast and multipolar. I'd expect planning for 2029 compute scaling to make sense only if the current paradigm gets stuck at ~AGI capabilities level (ie a very good scaffolding for a model similar to but a bit smarter than o3). This is because if it scales further than that it will do so fast (requiring little compute, as the tweet suggests). If capabilities arbitrarily better than o4-with-good-scaffolding are compute-cheap to develop, then things almost certainly get very unpredictable before 2029.

Having reflected on this decision more, I have decided I no longer endorse those feelings in point B of my second-to-last paragraph. In fact, I've decided that "I donated roughly 1k to a website that provided way more expected value than that to me over my lifetime, and also if it shut down I think that would be a major blow to one of the most important causes in the world" is something to be proud of, not embarrassed by, and something worthy of being occasionally reminded of.

So if you're still sending them out I'd gladly take one after all :)

I think both, by a long shot.  I estimate I spend over half my time outside of easy video surveillance (room without a webcam or phone pointed in a useful direction, or outdoors not in easy LOS of a traffic or business cam), and a slightly different half for audio.  For neither of these is high-fidelity POV data available at all, as described in the post.

For those times when I AM under surveillance, the quality is low and the integration is minimal.  There are legal and technical challenges for anyone to use it against me.  And it's VERY easy to find times/place where I'm not being recorded when I choose to.