I guess there are bots here making all these political comments to diverge the discussion and not talk about what really matters: is the model good to do actual work?

Looks interesting, but I feel that as models get better and better, the need to have 'personas' for each agent is actually a drawback compared to having a well-defined starting plan and objectives, because more 'intelligent' models will just figure out a more efficient way to do what they need to do.

And in Brazilian Portuguese it is read the same way as a slang for penis

That's what you learn playing factorio

And still my 4g and WiFi won't work that well

You probably get more than 100 bits per second.

That's a conundrum that most developers who climb the chain face and the best article I've read so far that comes close to answer it is this one:

http://www.paulgraham.com/makersschedule.html

Probably posted a billion times here since y combinator is just one of the things that derived form the work of the master Paul Graham.

This is a cool project, but I guess it could be a phone app that your entire family could use collaboratively to scan, tag as used and add to a shopping list.

Thanks, these are some pretty nice ideas right there!

What about using this for bit/alt coin mining? Do you guys think it is a good ratio of performance / power consumption?

Notice they use the GigaFLOPs metric, which is a (maximum) floating point measure. Bitcoin hashing (it's been a while for me) is compute intensive on integer math. For a good ratio of performance/power you'd want to look at architectures like the GreenArray devices. But the world of bitcoin hashing is now firmly positioned in ASICs. You can't cut architecture fat beyond that so you'd have to look at (general purpose) architectures built on a faster silicon technology. Wouldn't count on it.

http://www.greenarraychips.com

Kepler is fairly useless for mining.