Adam either does not understand how a computer works or he does not understand that inventions are combinations of elements and connection. Connections matter as my NAND and NOR example shows, but another example is stereoisomers. Stereoismers have exactly the same elements and almost the same connections (symmetrical connections), but if you are using a stereoisomer for its optical properties or its pharmacological properties if makes a profound difference if you have the left handed or right handed version. Things that appear similar to those not skilled in the art, such as Justice Steven and Adam, can be profoundly different and clearly deserving a patent.
.

However, rearrange what Legos connect to each other, and you can create a functional device.

Software is what transforms the computer (i.e., the bucket full of Legos) into a functional device.

As well-known observation is that “virtually all [inventions] are combinations of old elements.” In re Rouffet, 149 F.3d 1350, 47 USPQ2d 1453 (Fed. Cir. 1998). The difference between what is old and what is new is how the items are combined. A computer program simply changes how the switches in a processor connect to each other.

…a foot in each world.

You know as well as I do that software neither removes transistors from nor adds transistors to a computer, so I’m not sure why you even brought it up. Software merely changes which transistors the electricity flows through, like flipping a light switch. If you think flipping a light switch on your car makes it into a new machine, then we’re operating under such drastically different definitions of the word new that it will be very difficult for us to communicate.

Sure the form matters – do you actually read what you write. You are so wedded to your point of view you are unable to be rational about the subject. A NAND gate and a NOR gate are different machine – period. It does not matter whether this is done in hardware or software.

I suppose that depends partly on what mechanism you use to rearrange them. Consider if the designer of the circuit included a button that would use physical arms to rearrange the transistors when you press the button. Then I don’t think you’ve created a new machine by pressing the button and reconfiguring as the designer intended, since that was always how the machine was supposed to work. However, if the circuit was intended to stay the same and you had to pull off the transistors and resolder them, you could probably convince me that you’ve created a new machine.

I’d say it’s kind of like the wiring in a house. If you flip a light switch, you’ve closed a circuit and created new functionality in the light bulb (emitting light), but you haven’t created a “new machine.” Conversely, if you have to rip out the drywall and lay new cable to the lightbulb, then you probably have created a “new machine.”

1) the method/process is novel and not abstract (where to draw the line?),
2) or the physical machine with novel/enhanced attributes within context of the machine(if it’s a printer, how does it do printing better. if it’s a computer, how does it load/execute instructions faster).
3) or the act of deploying a known method to a known machine is non-obvious

The bilski patent can only be considered under 3).