Fireflite

Do you have a link to the electricity restoring sense of touch research? I've noticed the same thing by accident after some nerve damage and would love to be able to improve further.

The enormously wasteful environmental cost of PoW is certainly a turn off to those of us on the left as well.

The problem isn't at all soil or nutrients: those are cheap and make up a very small fraction of the final mass of food. Instead, it's energy required to power photosynthesis, turning them into delicious calories.

AI Dungeon uses it to excellent effect, but with extensive fine tuning and tweaks.

The way that you can turn normal Rust functions into systems, grabbing the relevant components through their type signature is wonderful sugar, and feels magic to use.

When scaling this up, I would be concerned about climate control, disease management, infrastructure limitations and labor / distribution details.

Light, nutrient and water distribution can be replicated, but making sure things scale properly and you don't get cascading failures is key.

Yes, you'll often get better results by transplanting earlier. You're missing the inverse control though.

In the absence of a specific mechanism it seems very implausible that the full moon has a measurable positive effect on the plants.

Whale corpses tend to reach the bottom of the ocean largely intact actually! It's hard for most creatures to eat them quickly enough, and they're typically largely eaten by sharks and scavengers on the ocean floor.

Injecting entropy into your data in order to generate differentially private access actually typically improves model performance, rather than degrades it.

This behaviour is counterintuitive, but effectively you're enforcing a degree of regularization in a way similar to how data augmentation works. You'll get reduced training set performance and improved test set performance, as with other forms of regularization until you make it too strong.

Most other word embeddings have hundreds of dimensions, not thousands. Are you able to hint at what causes this difference? Do you see better downstream task performance?

Money alone can't guarantee good outcomes in cases of severe mental illness sadly.

Our medications are not good enough, nor are they well-tolerated enough, and sometimes even with medication and money and strong supports and a willingness to try to improve patients still don't get better.

The whole point is that you can run your own algorithms; that's why it's on Google Compute Engine. As far as I can tell you should be able to avoid TensorFlow even, although I suspect that'll be a pain.

Why is asymmetric intellectual property protection a form of friction here?