Quote from: kingcrimson on October 25, 2014, 10:41:58 AM

If G3 failed why do we expect G4 to succeed? Both are just a lot of talk. It is just empty promises, lies, failings, then new promises, and new, different failings. Same cycle over and over again.

Honestly. It is very common in businesses and start-ups. People knock bitcoin stocks but the percentage is probably the same failure in real businesses and start-ups. Point being. AM's early success was an insane fluke that will never be repeated.

Gen 3 didn't fail, it just wasn't as good as expected and there was packaging problems initially too. Gen 2 was an actual failure. So, why expect Gen 4 to succeed? Because FC just told us that they had solved the problems with Gen 3. Sure, new problems might arise.

Quote from: Puppet on October 25, 2014, 11:20:33 AM

Gen 3 tape-out happened early February. Real world availability is only happening around now; ~8 months later.

Nonsense. Tubes went on sale on August 12 and chips were available months earlier. If AM would have just built miners instead of selling chips, Tubes would have been available a lot earlier.

Quote from: Puppet on October 25, 2014, 11:20:33 AM

Both bitfury and spondoolies expect their next gen 28nm miners this year, one with 0.2J/Gh (0.1J next year) the other 0.1J/GH at the wall. AM is aiming for 0.22-0.34J/GH (and that appears to be for the chip alone) at least one quarter later. Even if AM is on time and on spec for a change, I wouldnt get too excited if the network is pushing through 1EH by that time at todays exchange rate.

It's not just about power efficiency. Sure, competitors may have better power efficiency but what hash rate will those chips have? Look at Bitfury or Spondoolies 40nm ASICs and compare them to AMs gen 3. If those chips were overclocked to match the hash rate of AMs, what would happen to their power efficiency? What would AM gen 3 chips power efficiency be like if the chips were underclocked to match the hash rate of Bitfury's 55nm or Spondoolies 40nm?

Here's what FC said on the issue:

Quote from: friedcat on October 24, 2014, 04:19:32 PM

For BE300 in theory it is possible to get lower than 0.225W/GH by stressing the voltage down below 0.55V, but we do not have
solid simulation data yet, so let's see after the test chips are out.

December 16 is to be expected for us to get the chips. Testing time varies at 3-10 days since we had much more preparation work already done this time.