GavinB

This is a compelling theory, especially the implication that humans are early. I do wonder whether we should see the evidence of spheres of growing alien influence out in the stars, but instead we see a highly uniform universe in all directions.

This would indicate a few possibilities:

1. Expanding alien civilizations are relatively low impact and don't collect all of the energy of stars in ways that are visible to our current telescopes.

2. We are a very early civilization, civs are fairly rare, and we're relatively alone in the parts of the universe that we can see. Civs that are expanding in a grabby fashion started less recently in years than their distance in light years.

3. Aliens expand at close to the speed of light, so there are a lot out there but we won't see them until they're almost here.

4. Something that we have already noticed is actually evidence of grabby aliens, but it is happening in every direction so we assume that it is a natural phenomenon, because it is so uniform.

At the very least, it seems likely that we either we are alone in the galaxy, or expansion is very slow. The idea of "expanding in a bubble of influence close to the speed of light" seems implausible to me, just because of the vast amounts of energy required to accelerate and decelerate to relativistic speeds, not to mention protecting the cargo in transit--when you're flying at .9c, almost every other piece of matter in the universe is flying towards you at you at .9c. Accelerating tiny nanomachine von Neumann probes might be a solution, but how would they decelerate enough to not be destroyed on arrival?

It's all fascinating to think about, at least.

Is this in comparison to some other evaluation function which is perfect? I agree that all positions should have a certainty of win, draw, or lose with perfect play, but no engine is close to that level of evaluation function.

I do suspect that this pathological behavior could be trained out with additional fine tuning, but likely not without slightly diminishing the model's overall ability.

I got downvotes for not showing the calculation, fair enough, my mistake. I understand the frustration there. And yeah, you've got to check ChatGPT's work and ask for clarification to understand what it did and why.

That said, I disagree with the assertions made in your post. ChatGPT is a great tool for saving time in doing simple research and calculations, as long as you double check it. It wasn't a year ago, but it is now.

I know, I'm scared by it, too!

If you're going to yell at someone, at least do your own calculation so you can prove they're wrong first, please?

Total Sky Area=4π×(2π360)^2×60^2 = 148,510,660 square arcminutes (you can also look this up easily)

NIRCam has 9.7 square arcminutes FOV (Yes, I double checked this)

NIRCam resolution is 2048 × 2048 x 2 modules for 2.4–5.0 µm (there are lots of different cameras and modules you could look at to make this estimate, but for a sense of magnitude this will work. If you looked at 0.6–2.3 µm that would be roughly 4x as many pixels for the 8 modules, I think? Again, napkin math for fun.)

Total FoVs Needed = 148,510,660.8 / 9.7 ≈ 15,310,377 FoVs

Total Pixels = (2048×2048×2)×15,310,377 ≈ 128,432,754,096,706 pixels

LLMs used to be quite awful at math. Now they're solving quite complicated problems with sophisticated plugins. This is a simple and straightforward calculation, just getting the FOV and its resolution and then projecting that onto a sphere. And, yes, I probably should have posted the full calculation, it's definitely more interesting to see it all written out.

I'd suspect it will actually accelerate moving everything into the cloud.

If your entire business is in the cloud, you can give an AI access to everything with a single sign or some passwords. If half is on the cloud and half is local, that's very annoying to have all in-context for your AI assistant. And there's no way we're getting everything locally stored again at this point!

Alternatively, figure out how to use GPT-3 in a market that involves some schleps. I'm working on one in Education Tech, building something for a need that teachers have been practically begging for. There are particular regulatory challenges, unique sales paths, and a big first to market advantage because educators aren't online constantly researching their options. Once you're embedded in the education consciousness, you're there for years.

But yes, another random thing GUI for generating marketing copy from GPT3 isn't a good long term play.

Having this be a universal policy has a big clear benefit--when someone at the company dies, their family's immediate needs and expenses are taken care of. It avoids the awkward "passing the hat" to the company or fellow employees for support. There's no need to make special exceptions, determine how much to help out in a crisis, etc--it's taken care of automatically, to a pre-determined extent. No one can complain that an employee died and big rich Google did nothing for their family. It saves a lot of potential handwringing, bad PR, and employee dissatisfaction, etc.

I agree in principle. On the other hand, global crises have always called for extraordinary measures.

Even in this pandemic, we've (correctly, I would say) basically thrown out important principles like freedom of movement and freedom of association during lockdowns. I think that allowing volunteers to put themselves at a relatively low risk to speed up vaccine trials would have been a relatively small transgression, compared to the lockdowns and travel bans (which, again, I think were justified).

But I do take your point.

Yes, all of them. Or possibly expanding capacity for each of the major vaccine types, to be ready when we know which one is the winner.

And then, yes, a lot of those factories would have ended up sold for scrap (or mothballed for future pandemics or mutations). But the few that worked then save trillions of dollars of value and hundreds of thousands of lives.

It's a simple matter of calculating expected values and investing accordingly. But thats not how our civilization works.