Collaborative Discussion and Presentation of Research Project Designs

90 min

Overview

The week ends the way research careers run: you present your project to a room of peers and defend it. Each team presents its research direction and paper skeleton — eight minutes, seven slides — followed by five minutes of discussion opened by a designated discussant team that has read the skeleton in advance. This is run as a real workshop session, not a show-and-tell: the discussion probes whether the claim, the measures, and the evidence actually hold together. We close with the synthesis the whole week has been building toward — the hunt for what your very different domains have in common.

The discussant's duty

You are not the audience; you are the first reviewer this project will ever have — and the friendliest one it will ever get. Open with the strongest thing in the skeleton, then ask the two questions whose answers would most improve the paper. Softballs waste the five minutes; cruelty wastes the week. Aim for the question the team will be glad someone asked NOW rather than in peer review.

The session at a glance

1
Presentations: eight minutes per team, hard-stopped.
2
Discussion: five minutes each, opened by the designated discussants.
3
The convergence synthesis, all teams together.
4
Exit: skeleton v2 + the first-90-days plan.

The discussion — what to probe

Discussants and the room, in rough priority order:

The reviewer's checklist — one last time

the readbackcan you state in one sentence what they are doing and why it matters? If your sentence and theirs differ, start there

the jointdoes the introduction promise only what the measures can establish and the proof-of-concept tally can support?

the fair fightis the baseline the strongest reasonable version of today's practice — including the best prompt it could be given?

the architectureis every mechanism — planning, memory, second agent — traceable to a failure it prevents? What did they reject, and does the rejection convince you?

the safety linewhich decisions stay with a human, and does the design enforce it?

the invent columnis the claimed contribution really missing from the field — or just missing from the team's reading so far?

The convergence synthesis

After the last presentation, we step back — the same move as Monday's Day After discussion, now with a week of evidence behind it.

The questions for the room

1
Across every domain presented today: which scarce things, missing methodologies, and bottlenecks appeared more than once?
2
When memory became cheap, organization and search were the inventions that mattered. When cognitive work becomes cheap — judging by THIS room's projects — what is the equivalent invention?
3
Which two projects in this room could share an experiment, a benchmark, a method, or a paper? Say it out loud; this is how collaborations start.

The exit deliverable

Before you leave, each team finalizes:

The take-home package

user prompt

Skeleton v2: the paper skeleton, updated with anything today's discussion changed The discussion log: the questions you were asked, and your answer or your "we don't know yet" for each The work plan: what you will go work on to complete the paper — the literature to read against your invent column, the data or access to secure, the benchmark to finish, the proposed study to run. You will complete the paper on the schedule the instructor announces — think weeks, not months

And a standing offer, said plainly before we part: we are happy to keep collaborating after this week — reading drafts, advising on the study you designed, and coauthoring where the work grows into something shared. Reach out.

Before we close

💾 Capture

Write down:

1The take-home package — all three parts.
2One line: the question from today's discussion that most changed your view of your own project.
3One line: the other team's project you most want to see written — and the one thing you'd tell them to keep.

Be ready to discuss:

You arrived with a domain and left with a research direction, an approach, a proposed study, proof-of-concept evidence, and a skeleton that traces every claim to evidence you made yourselves. The paper is started. Finish it.