barumrho

Is that a long term prediction? Are you saying that physical infra will always lag demand from here on?

Some do exactly that?

Setting aside politics and personalities, are Palantir's products any good? At least on the surface, their products seem to have many competitors, so is the product quality actually driving their growth?

Would you mind sharing some of your setup? Which model and which params do you tweak (e.g. temperature)?

Is this true? Hardware costs have only gone up during this time. Are you referring to electricity cost to serve these models? (i.e. compute got more efficient?)

Hooks are not checked in to the repo, so they used a zip archive.

If everyone took your position, then frontier labs can keep raising their price.

This looks pretty promising to me. It will likely replace the need to set up OpenClaw for average personal users. The work of getting email, messages, and all the personal data on the phone as context seamlessly is not as straightforward as one might think.

I'm curious how the pricing will work. Would it be free up to some limit and then some subscription pricing? I can't imagine it can be free unlimited usage given the price of serving these models.

You mean you aren't bullish on volume of illegal transactions? /s

well, you couldn't justify the cost if you still employed all 3.7B

Is it too high or too low? Honestly cannot tell

If we all 10x our productivity, then we also 10x the cost of taking a day off and theoretically progress is compounding, so it's even costlier...oh no.

That's a bit of a stretch, but the number is bigger than I expected!

Siri doesn't need to have conversations with you. ChatGPT can do that. But, it should be able to do actions you'd do on your phone.

Is there a reason why this has to be done at training time? Could the system prompt tell the model to convert the output to a different format?