To: davesd who wrote (9566 ) 10/26/1997 12:48:00 PM From: Alan Siegal Respond to of 70976
Dave: << There are more chips out there than people really need >>. That's probably true, but it doesn't account for new software. All of us long termers have to be a little futuristic here, and we all have to consider certain facts. 1) We have been hearing about language processing software for over ten years now. It's my own opinion that such technology is probably 50-75 years in the future, and that to begin on anything but it's most rudimentary aspects at this time is like Babbages's attempt to make a computer with cogs, wheels and connecting rods. A big PC has 32meg of memory. How much does a person have? 32 GIG? 320? And they fit in a head no larger than a laptop--and that includes stereo scanners, stereo sound receivers, a speaker, thermostat and olfaction device! To get an inorganic instrument that sophisticated in such a small space will require chips as cheap and plentiful as sand. Don't think it isn't gonna happen, and regardless of where they're made, these chips will have to be made by machines, and the startup cost of those machine-making firms is already unattainable by any new entrant. 2) In case language emulation is just too far out, let's take it a bit closer to home. I don't know if there's any generally available commercial security system that recognizes family members by their faces (including with a band-aid on the nose) and voice-print (even with laryngitis). But I'll betcha the technology is there, just unfeasable today owing to the high cost of DRAM. As soon we get to some market-determined number, you can be sure such systems wil be de rigeur. I don't think this scenario is way kooky, do you? Ditto for HDTV. Ditto for last-mile fiber optic switches. Ditto for everything medical. I'm not an imaginative guy, so I can't fathom what uses folks will drum up when DRAM costs less than drinking water, (which isn't getting cheaper in my neighborhood.)--but I sure know that they'll be utterly profligate in their use of the stuff. (I ran my business for over a year on cp/m using 64k memory.) So all we have to do is choose a few good companies--some large, some small--and buy them whenever we have some money. Who'll know about this week's bump when my grandson goes to sell 1000 shares of AMAT 40 years from now? Just MHO, natch. -Alan (I think the average investment in these industries will beat the S&P 500 handily over the next few decades. I ask for little more.)