3 ms·

100 times more energy would result in heat dissipation problems. The stock human body runs at roughly 100W, the brain runs at around 20W or 25W. So it'd proba

100 times more energy would result in heat dissipation problems. The stock human body runs at roughly 100W, the brain runs at around 20W or 25W. So it'd probably be worth it to upgrade a brain at your ratio, but I'd need confidence that the engineering problems are solvable. (2kW of power for the 100x brain cost, times 24 hours, would cost me about $5 a day if I bought it as electricity... $2000/year... pretty affordable). I'd glaaaadly take a 10% brainpower upgrade that required 100% extra energy, though. I can afford the extra 500 kCal per day of food intake. My understanding is that most first-worlders like their food enough that they'd take a 0% improvement in trade for being able to eat another 500kCal per day of food without weight gain. I mean, all that assumes that a "10x stronger brain" actually results in a useful performance upgrade. That part is pure fantasy speculation.

seanmcdirmid 9y ago
If we move away from optimal, then energy needs are likely not gone g to SoCal linearly. And ya, some of that extra energy wil,be used for cooling and such. I would take the 10% upgrade also, but if everyone did...well...it could become a problem. As food isn't transformed into energy uniformly, obesity could still be a problem.
wruza 9y ago
As all futurists, you seem to solve unexisting problem here. Only limbs are required to be local. If you can make artifical brain or parts without erasing your original, then you can easily offload new parts to datacenter, leaving only TCP/UDP interface inside your skull. BrainHub. Additional brain processing and storage for $25/mo for your team of 5. Free for public memories.
- tboughen 9y ago
  I suspect latency issues may preclude this.
  - wruza 9y ago
    Otoh, I already have various memory latency and processing issues, having just 1x conventional brainpower. If not complete military-grade win, it would be nice tradeoff anyway. Alternatives are: placing new heavy brain in abdomen (because stock digestion can't produce so much power by design), or replacing back, coxal and leg bones with carbon variants with embedded batteries, which should be far more expensive than subscription to low-latency BH channel for additional $1000/mo SLA per team. Also consider servicing and failure costs for businesses depending on 5*10x brainpower teams, and reduced labour market.
ohyes 9y ago
For 10% increase one would guess you could just overclock the brain, if we knew enough about the brain to connect hardware to it. I think you'd be more likely to get an arithmetic co-processor to start, wire neurons in such a way up to a small circuit that you get precise mathematical intuitions based on it. Being incredible at math from birth would likely change your life indescribably.

seanmcdirmid 9y ago

If we move away from optimal, then energy needs are likely not gone g to SoCal linearly. And ya, some of that extra energy wil,be used for cooling and such. I would take the 10% upgrade also, but if everyone did...well...it could become a problem. As food isn't transformed into energy uniformly, obesity could still be a problem.

wruza 9y ago

As all futurists, you seem to solve unexisting problem here. Only limbs are required to be local. If you can make artifical brain or parts without erasing your original, then you can easily offload new parts to datacenter, leaving only TCP/UDP interface inside your skull. BrainHub. Additional brain processing and storage for $25/mo for your team of 5. Free for public memories.

tboughen 9y ago
I suspect latency issues may preclude this.
- wruza 9y ago
  Otoh, I already have various memory latency and processing issues, having just 1x conventional brainpower. If not complete military-grade win, it would be nice tradeoff anyway. Alternatives are: placing new heavy brain in abdomen (because stock digestion can't produce so much power by design), or replacing back, coxal and leg bones with carbon variants with embedded batteries, which should be far more expensive than subscription to low-latency BH channel for additional $1000/mo SLA per team. Also consider servicing and failure costs for businesses depending on 5*10x brainpower teams, and reduced labour market.

ohyes 9y ago

For 10% increase one would guess you could just overclock the brain, if we knew enough about the brain to connect hardware to it. I think you'd be more likely to get an arithmetic co-processor to start, wire neurons in such a way up to a small circuit that you get precise mathematical intuitions based on it. Being incredible at math from birth would likely change your life indescribably.