I am using Macs (G4s and Intels). I also use Windows machines daily (for work). I have Amigas for games and tweaking. I wish I could buy a modern machine with a modern CPU and coprocessors for graphics, sound, networking, media DSPs...something like a real supercomputer
not this consumer-oriented garbage with flashy stickers and a gutload of branding messages.
One can argue that modern graphics cards act as "coprocessors" and this is true but they are not standard equipment. Also, no OS/hardware out there supports coprocessors for the killer application of today (networking and media, such as MP3).
Promise of multicore design is to provide the same coolness of co-processor designs (provided you have software that supports it); you would get a very responsive system (orders of magnitude more responsive than the fastest machine today), but the trouble is that it will take years for software to support multicore designs. Windows (XP and Server) allegedly support multicore *and* dual CPU hardware, but I am yet to "feel" the difference beyond perf. improvement with more RAM added.
Mac is probably the closest to that goal but still does not have dedicated hardware for networking and will have an OS that supports multicore natively some time next year (Snow Leopard).
Multicore support is significant because old-fashioned software is wasting a lot of hardware you paid for in a new machine. When you push a lot of data through your wireless (or wired) LAN, you are using too many (single) CPU cycles *and* you are keeping the data bus very crowded and that chokes up the overall performance while other cores are idling. This is why even the GHz machines do not feel as fast as they should be.
A CPU/mobo designs which are 100 times faster on paper than CPU/mobo designs from 20 years ago, *should* feel at least several orders of magnitude faster than "old software" And yet it does not.
Intel's now obsolete supercomputer on a chip called i860 (64-bit RISC processor from 1992 or 1993, can't remember now) was a technology that could deliver Amiga-like graphics. i860 Graphics/DSP boards for 486DX-33MHz could deliver video quality of today back in 1993. i860 failed to capture marketshare because it did not have a lot of software written for it. Expansion boards were very expensive and that did not help either (this should sound familiar).
As of future Amiga design, it would have to include MC68K (likely a copy of MiniMig design) for compatibility.
The coolest thing that happened to Amiga (relatively) recently is the MiniMig - a proof that a legacy coprocessor design can be implemented in FPGA and "teamed" with a low-voltage MC68K to deliver a low-cost compatibility package. Natami is likely the next wave, when it becomes available.
Amiga as a platform does not have a bright future as long as software platform (OS) remains in limbo.
My hunch is that both Amiga Inc and Hyperion will go bankrupt within a short period of time. Their business models are not sustainable and their business style is not entrepreneurial but rather confrontational. Whoever wins that ridiculous battle, Amiga as a platform will lose. Delete Amiga Inc. and Hyperion from your bookmarks. It is a waste of time. And for God's sake, stop bidding up those PPC cards to ridiculous amounts on ebay. They are worth 200-300 bucks tops. 14 year old hardware.
Amiga's future is outside of the corporate world. MiniMig and AROS (both open source) prove that. A multicore design that includes powerful graphics card, sound card with an powerful DSP, network card with a dedicated CPU
, built-in SCSI and SATA interfaces would be worthy of an Amiga - especially if it ran AROS.
Until then, grab a Mac and a Minimig (or a vintage Mig