Automated departure messages at Symbolics

That meant that their massive advantage in running lisp evaporated in just a few years between the mid 80s and 1990.

This was coupled with an inability to simplify their product. The first Symbolics machine we ordered at the CRL at NMSU in '84 had to have a guy with a graduate degree from UT come light it up. The SUN 2 that arrived at the same time had everything preinstalled and just needed me to plug in a few cables and turn on the power switch. The complexity of the Symbolics was reflected in the maintenance contract prices and in the fact that you generally had to deliver an entire OS and dev tool chain image to deliver any program (a so-called world build).

A few performance improvement doubling times later, Sun was the darling of the startup world and you could buy all of Symbolics for the price of a few of their maintenance contracts. Sun had been able to hitch a ride on the broader base of the 68000 processors from Motorola and then was able to improve the Sparc processor at a much faster tempo than Symbolics could cycle their own design.

Of course, about 15 years later, Sun fell victim to similar pressures and business model problems when the Sparc development team couldn't maintain tempo against Intel. Intel's primary advantage was broad market adoption which funded faster design cycles. Sun also had locked in on large banks by that time (how they survived the .com crash) and couldn't get back out to the broader market.

So it wasn't really that the flood came to the desert ... it was that Symbolics was focused on the wrong problem because they thought they had all the answers.

Start with Gmail and slack add-ons and role it from there.

Given the nature of the process, better pick one-time payment and maybe add a corporate plan.

80s - 90s where the era where quantity, volume and network externalities became more important than quality or even capabilities. There were good reason for that to happen, of course.