Here is what I took from the article by Noortje Marres.
I think that the idea that stuck with me the most was the idea (similar to Latour's upstream/downstream science in the making argument) that Marres' argument can be shaped similarly, in that:
(being upstream) STS is about research (about values AND information), which then needs to be framed and mobilized to a public (whom generally will not act on information unless an issue affects them directly, per Marres--which is why values targeting is important)--the public will then frame and mobilize the issues toward the institution, whom then frames and mobilizes it to legislators... where we end up "downstream" with policy.
Through this process, a certain cohesiveness between particular groups can occur, and assemblages can be identified.
She also goes on to talk about the process of framing, and its important role in consensus building in a public. We have to look carefully about how we "select some aspects of a perceived reality and make them more salient in a communicating context" (Entman, 1993: 53)
She says that institutions can't do things on their own, or unilatterally, just as democratic governments cannot work without the votes of their constituents. She argues that the public needs to mobilize the issues to the institutions because their resources are better allocated. This idea reverberates through a few different parts of the assemblage.
She sees value in public discourse, and indicates that this is where the most/best action is. (Which is similar to Latour's arguement of being a part of science in the making)--She also thinks that we aren't doing enough to involve the public in sciences and technology research, and that steps must be taken to involve them, but exactly how, is debatable. (See also the Lippman/Dewey debate on "advocate of expertocracy vs. Proponent of participatory democracy")
I personally think that a balance has to be stuck between the two ideologies, and that experts can do a better job of involving the public. Presenting their work is a part of what they do... they should be good at it--and it should be relevant.
I can see the parallels between what Marres is saying, and what some of the other authors we've read are saying.