Weekly note 2025-02-15~2025-03-01

2025-02-14 What we haven't been able to do

local deep research experiments and

Experiments with broadlistening in Devin and

Still experimenting with new algorithm [originality filter

TODO

backlog, last week's story, put what you did or didn't do this week and let o3 think about it

What should I make them think about?

Task Prioritization: Organize the backlog and tasks performed and not yet accomplished and identify which are important.

Cause Analysis: Identify causes of unaccomplished tasks and problems in the last week and consider areas for improvement.

Next Steps: Derive a future action plan and proposed resource allocation and scheduling adjustments

These axes allow for efficient task management and planning.

---

I get the impression that o3-mini-high says the right things without thinking it through.

---

o1 Pro.icon

2/25 Lecture

Highest priority due to impending dates (see T0057). Preparation, including materials and rehearsals, is needed.

Lack of task concretization

Experimental tasks are left vague as to "where to start. Clarifying the setup and scope will make it easier to proceed.

tightening schedule

Time was allocated for events such as lecture preparation and roundtable discussions, and experimental systems were put on the back burner.

Confusion due to information dispersion

Management of multiple channels (see T0113/ T0116) such as Signal, Slack, ChatGPT, etc. is not keeping up.

---

2025-02-28

Work in progress from last week 2/14

Mechanism for experimenting with clustering algorithms in Devin

Prototype of Deep Research system for data not available on the Internet (not yet tested)

Tried Computer-use, looked up how to use it from a script (untested).

I haven't been able to do all of it because of the busyness and preparation for the lecture afterwards.

---