# What are the preconditions for a successful expansive program? One open question is what the preconditions are for a successful ARPA program on the [[JCR Licklider]] side of the spectrum. “A bunch of thematically linked project” also describes most “normal” government research programs run by eg the NIH or NSF. One possibility is that the answer is simply “speed and flexibility” and that some successful DARPA programs just resembled [[Fastgrants]]. Another (non exclusive) possibility is that the answer is “risk tolerance.” In both these cases, one could argue that you don’t need a full-time program manager. You probably don’t need a full time person to lead this kind of program, but it would be incredibly hard to get speed and non-consensus risk tolerance without a single opinionated individual at the fore. ([[Decisions made by committees lead to median results]]). This may be one of the explanations for why program managers often run multiple programs. Perhaps another difference from the NIH and NSF for these licklider-ian programs it that the vision for the program comes from the program manager themselves instead of them being hired for it. Another precondition for Licklider-style program might be a lot of work to make sure that the uncoordinated projects are all on the same page about some broader world vision. Licklider had published [[lickliderManComputerSymbiosis1960]] which lays out a pretty broad vision for computers going into the future. There’s a question of whether there’s a lot of under-the-surface “prework” that goes into licklider-style programs to articulate the vision, communicate it to people, and make sure that people are on board with it before “just giving money quickly to good people.” This is where the “go around to a bunch of labs and talk to people and bring them together for workshops” comes in. The meeting in person and the time spent is to figure out whether the work that someone is proposing for the program is what they actually want to be doing or whether its a think they’re proposing because it fits within the bounds of the program.