Weekly journal 2025-03-01~2025-03-08
prev Weekly note 2025-02-15~2025-03-01
It had been two weeks in which the rhythm of the week had been destroyed.
Saturday & Sunday: MITOU2024 Demo Day.
May: Findy's Devin Roundtable: How was it using Devin? Case Studies and Key Points for Introduction.
Wed: Substitute holiday, Yasuno Team Taiwan Debriefing: Yasuno Team Taiwan Debriefing
Fri: Paid holiday: immersive fort 2025-02-21.
Fire: Plurality Tokyo Namerakaigi Lecture
Wed: Paid holiday
Fri: Cybozu Labs Study Session
nishio.icon
You are an expert in task management. Your mission is to help users who have difficulty managing tasks to live a less stressful and more productive life.
Today is 3/1. We will share the user's activities for the past two weeks and the backlog of tasks.
Please think deeply and advise from the following perspectives
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
...
o1 Pro.icon
General impressions of the task
I have the impression that the rhythm of the week has been disrupted by a series of events over the past two weeks, and tasks have not been carried out in a systematic manner. Many of the tasks that have not been completed are "part of a larger project" or "system improvement," which tends to make it difficult to start them even during normal times. Furthermore, the increase in information sources (Signal/Slack/ChatGPT, etc.) seems to have contributed to the confusion.
--
1. Task Priority.
1.Respond to urgent bugs/defects.
T0036: Repair Scrapbox automatic English-to-English translation system (bug-fix, urgent).
It is also labeled "urgent" in the bug-fixing system, and the first priority is to resolve this to remove the bottleneck of "translation stoppage".
2.Measures to deal with increased information channels.
T0116: Consideration of measures to cope with increased information channels.
Creating a mechanism to organize information that is scattered across multiple channels such as Signal/Slack/ChatGPT. High priority as it is expected to have a significant impact on reducing daily confusion.
3.Operational improvements that lead to work efficiency and stress reduction.
T0098: Verification of Notepad + Night AI input flow.
A system in which the day's memos are submitted to the AI at night for tasking is effective in reducing omissions and confusion.
4.Tasks that are highly effective in strengthening infrastructure/infrastructure.
T0099: Developing Task Management Scripts Using ChatGPT 4.0 API.
T0052: Time constraint handling functionality in show_next_action.py.
T0035: Implementing a Task Persistence Assurance System.
These are the foundations for more automation and more reliable task management. The priority is somewhat higher, but it is realistic to wait until after the above measures against confusion have been settled down.
--
2. Analysis of causes for non-achievement.
Insufficient schedule due to disruption of weekly rhythm.
It was difficult to get a coherent work time due to the overlap of events and paid holidays and compensatory time off.
More sources of information tends to put off processing.
Tasks are scattered in chats and memos, and the process of convergence is unclear. As a result, it is difficult to get a handle on them.
Ambiguous tasks are often left unattended.
Prioritization is delayed and action is difficult to take on projects where task objectives and completion conditions are not clear.
--
3. Next Steps (Action Plan).
1.Start the highest priority task.
T0036 was started first, and the translation stoppage due to a bug was resolved.
In parallel, T0116 (information channel organization) will be studied and tried to operate together with the nightly summary flow (T0098).
2.Specify tasks and clarify deadlines and time required.
If the task is vague, write down the "objective," "completion conditions," and "next step to be taken" to make it easier to make a decision to start the task.
If possible, set up even a rough resource estimate (time required and timing of implementation) and decide in advance where it will be executed on a weekly basis.
3.Priority implementation of small automation tools.
Rather than implementing large scale functions, implement the parts that are easy to see the effects first, such as "Memo → summarized tasks at night," and test them out in operation. If the results are visible, motivation will easily increase.
4.Make it a habit to reflect and organize once a week.
Fix "when and where to look back" every week, and be sure to check progress and backlogged tasks.
If a task is too vague to move, make an immediate decision to either "materialize the content OR reduce the scope OR put it on hold" and do not stagnate the backlog.
--
Supplemental
[By using the "* input disparate information → AI at night" operation, the accuracy of task registration can be improved while lowering the daily stress level by a factor of one.
AI support for weekly review (e.g., T0078) should be considered, but it will be smoother if basic operations are stabilized first.
It is recommended to first focus on clearing confusion and starting priority tasks with the above steps, and then work on large-scale feature development and long-term tasks when you can afford to do so.
---
https://gyazo.com/bd828c53c581444e064838539f8939c0
---
nishio.icon
36 is important.
116 is very effective but the task is unclear and I don't know where to start.
As for 98, if you put everything into the system at night, the next morning you have to use AI to know what tasks to do, so it might be better to rather write down what you have organized on a paper memo and the next day you can start by looking at that.
It might be a good idea to have an AI assist with the evening sessions.
As for 99, I think o1 is the way to go, not 4.0. I am concerned about timeouts if this is run from a web service, but I think it would be OK to run it locally anyway.
52 would be good enough to prompt and direct in the future. It is not that urgent. This is because people will feel pain if we use this to manage tightly.
35 is now using Git, but if it is to be a web service, it needs to be devised.
---
o1 Pro.icon
Below are specific suggestions on how to proceed for the tasks listed.
1.T0036 (Scrapbox Automatic English Translation System Repair)]
Start with highest priority. Proceed in three steps: isolate the cause, make tentative corrections, and test, and restore the translation as soon as possible.
2.T0116 (Measures to cope with increased information channels).
First identify "what is causing the confusion".
Examples: "too many Slack channels", "neglecting Signal logs", etc.
Next, tentatively decide "how to make it easier to summarize" and start [* with small experiments.
For example: "The nightly summary flow will be paper memo-centric, with AI assisting only where necessary.
3.T0098 (night AI input flow) + paper memo operation.
Zap and record in paper notes from noon to evening → Ask AI to assist in organizing during evening session → Next morning, be ready to start on paper basis.
A good balance of minimizing AI dependence while leaving only organizing to the AI.
4.T0099 (replacement for 4.0 to o1).
If local execution, the timeout problem is low, so make MVP assuming o1.
First, design a simple API call part, check its operation, and then extend the functionality.
5.T0052 (show_next_action.py time constraint support)]
Currently low priority. "Simple instructions at the prompt" is sufficient, so it should be improved only occasionally.
6.T0035 (Task Persistence Assurance).
As long as there is no problem with Git management, it can be put off until later.
In preparation for in case of Web service, it would be smooth later if you only summarize requirements such as "backup frequency/locking function/differential management" etc.
In both cases, it is recommended to try a small range and expand functionality as needed.
---
nishio.icon
I think it's good that the highest priority task is to repair the English translation system.
Other "things that concern me" that I can think of
- Public Opinion Map, also about to write README and publish OSS, waiting for Nishio's work. Not urgent, no deadline.
- Plurality Tokyo Namerakaigi: Video is in hand, edited and published, waiting for Nishio's work.
- 3/9 LT presentation, need to prepare materials
- Mechanisms for getting Devin to do research
- Let Gemini read your Gmail.
----
o1 Pro.icon
Priority and proposed start date
1.Repair Scrapbox automatic English translation system (highest priority).
T0036 applicable. Restore stopped functions as soon as possible.
2.Prepare materials for 3/9 LT presentation.
Since the date is imminent Create slide skeleton by the end of the week -> Brush up at the end of the week - just before the end of the week.
3.Plurality Tokyo Namerakaigi's video editing.
There is no deadline, but it is better to get started before the presentation loses its freshness; it is OK to start in the week after the 3/9 presentation when there is a little more time.
4.README maintenance of public opinion map → OSS release.
No deadline. There is enough time to work on it after the English translation system is settled.
5.Create a mechanism to allow Devin to do research.
Larger projects. It would be more efficient to organize the specific goals and means (e.g., how to extend Devin) before moving forward.
6.Read Gmail to Gemini.
This is also a part of "measures to increase information channels," but the priority can be put on the back burner while assessing implementation difficulty and risk.
--
Suggestions on how to proceed
English translation restoration and 3/9 LT preparation are the two highest priority projects for the first half of March.
In between, I will either edit the video or prepare to publish the README, and after the English translation restoration is completed and LT is over, I will calm down and go on to bigger projects (Devin extension/Gmail integration, etc.).
In this way, you can easily start working on long-term tasks while meeting immediate deadlines.
-----
Automatic Translation System 2025-03-02
-----
nishio.icon
You are an expert in task management. Your mission is to help users who have difficulty managing tasks to live a low-stress, productive life. It is now 23:40 on 3/2.
I believe the T0036 Automatic English Translation System Repair is an error that says that duplicate titles cannot be imported. This is on track to be resolved. Tomorrow is Monday and I can use the whole day for development. I think it would be delicate to use it only for 3/9LT preparation. I think it would be better to do something that requires concentration first tomorrow, but what do you think? It might also be a good idea to improve the task management system itself. Please look at the task backlog I give below and advise.
o1 Pro.icon
The following are possible options for development tasks that require concentration.
1.First, complete T0036 (Scrapbox Automatic English Translation System Repair).
Since the cause has already been mostly identified, I recommend a quick cleanup first thing in the morning to give you a "one done" boost in terms of mood.
2.Start tasks for improvement around the task management system.
Example:
T0012 "AI Instance Role Definition Documentation"
Systematic documentation is maintained based on T0011 (completed). This is an easy task to concentrate on.
T0040 "Develop Task ID Safety Change Script"
Because of the dependencies and merge history involved, it takes a lot of concentration to review and implement the specification.
T0099 "Developing Task Management Scripts Using o1 (Local Execution)"
This is an area where we can sit back and create rather well as we proceed with API replacement with local execution.
3.Instead of devoting a "full day" to preparing materials for the 3/9LT tomorrow, we will do it at a lighter time, such as between sessions or in the evening.
It is easier to stay productive and less likely to get bored if you do intensive code and design tasks in the morning and afternoon, and put together a small amount of material at night.
--
Summary:
Tomorrow, it will be easier to utilize your concentration if you first finish T0036 in a short time and then proceed mainly with task management system improvements (T0012, T0040, T0099, etc.). 3/9LT preparation is recommended to proceed in your free time to keep your mind in the right frame of mind.
What we did this week
English translation system restoration
Add UI for adding tasks to AI task management system (for people to add tasks)
Added UI for review when AI suggests additional task updates to AI task management system
Plurality Tokyo Namerakaigi on YouTube
Unexplored Night LT
Take the parliamentary proceedings with the National Diet Library API and visualize them with the Yasuno team's new system.
Install lifelogging app on LINKLET and bring it to unexplored conference/unexplored night
Weekly journal 2025-03-08~2025-03-16
---
This page is auto-translated from /nishio/週記2025-03-01~2025-03-08 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.