Last major update: July 5, 2024
I use this as a living document for my standing plan for the next months. I expect to update the document somewhere around once a month; though possibly less frequently (once in 2 or 3 months is also okay). My daily updates with summaries at https://updates.vipulnaik.com/ is a more accurate record of what I actually did, and I am more committed to keeping that up-to-date.
Bucket-based and cadence-based classification of activities, and how I try to balance buckets and cadences
-
Day job: Here, I set both minimum and maximum time targets; I need to hit some minimum targets to continue to be and appear useful at the day job, but also want to make sure I'm not spending too much time at the expense of other buckets.
-
Personal projects: Here, I only set minimum time targets; in my ideal world, I would spend unlimited time on this.
-
Personal finances, income and taxes: I only set maximum time targets; to the extent that I can streamline things enough to spend zero time here, that is good with me!
-
Personal chores: I only set maximum time targets; to the extent that I can streamline things enough to spend zero time here, that is good with me!
-
House chores: I set neither minimum nor maximum time targets, but may start setting maximum time targets at some point if it ends up being a time drain.
-
Family and career planning: So far, I haven't set minimum or maximum time targets, but I might start setting them temporarily when working on focused projects in this area.
-
Other activities: I don't set either minimum or maximum time targets.
Activities can be classified as follows based on their cadence and reminder system:
-
Activities embedded in my daily routine: Note that for personal chores, things that are strictly part of daily routine (such as eating, jogging, and brushing) are not even time-tracked. For other buckets, they're generally time-tracked. No significant planning is needed for these activities.
-
Activities embedded in my multi-day (less than a week) or weekly routine: Examples include buying groceries, cooking, and updating my daily update summaries. While these are generally time-tracked, they're sufficiently embedded in my routine that I don't generally need to make deliberate effort to do or remember them. Only in exceptional circumstances, such as ill health, time crunch, or inclement weather, do I need to devote planning effort to them.
-
Small activities that arise due to "incoming" signals: This includes responding to email messages, alerts (whether at work or for my personal servers). As I have reasonable routines for keeping an eye on incoming signals, and the actions to take (once I do receive the signals) are generally straightforward, I generally don't need to put in deliberate effort to do these.
-
Activities that are part of monthly checklists, such as contract work payment updates, monthly balance updates, etc. Some portion of the month change checklist is publicly available here.
-
Periodic activities that are less frequent than once in 1-2 months, that are generally managed through calendar reminders. These may include things like refilling balances, expiration dates for long-lasting foods and medicines, checking the condition of some equipment, etc.
-
One-time projects: Examples here include things like server migration, automating something, making development improvements to a repository, etc. If these are one-shot tasks, that can be done in one or two sittings, I may use calendar reminders to nudge myself to do them around some time. For more complex tasks, I generally do them in smaller increments. For personal projects, I'm starting to manage these increments by using weekly issues that record what I intend to do in personal projects this week.
-
Ongoing tasks with no set required cadence and a lot of time flexibility: Examples include DLW data entry, Wikipedia work, etc. Generally, there's no minimum or maximum amount that has to be done in a given week, and it makes sense to do more or less depending on my mood and time availability.
As of April 2024, I am reasonably on top of all the cadence categories other than the last two (and not coincidentally, the last two categories mainly fall under personal projects). My struggle with the last two has been that they get effectively deprioritized in favor of the other cadence categories. There are two broad solutions I can think of:
-
Some things can be moved to the other cadence categories. For instance, ongoing tasks with no set cadence can be given an artificial set cadence and then moved to a month change checklist.
However, I don't want to overdo this, as a lot of the joy and flow with these tasks comes from the variability and spontaneity in the task construction, as well as the open-endedness of how much to work on it. Adding it to yet another TODO item on the already-fairly-long month change checklist won't work.
-
Weekly issues where I outline roughly what I plan to do in these buckets, particularly for personal projects: The goal with the weekly issue isn't to set a very strict set of requirements for what to do in the week, but rather to orient and prepare myself to do the tasks and set an expectation that doing them is important. It's fine to end up doing slightly different personal projects tasks than what I outlined in the weekly issue, but if I'm systematically doing way less than planned, that's a warning sign. If I'm ending up doing slightly less, or doing about the same but slightly different things, that's totally fine.
For instance, I didn't do any of the listed things in the week of 2024-03-18 to 2024-03-24 and the week of 2024-03-25 to 2024-03-31, and so far in the week of 2024-04-01 to 2024-04-07, though I do expect to do at least a few things in this third week (that is not yet over as of the time of writing the document). These weeks have been unusual in terms of having a huge amount of personal chores work, but nonetheless seeing very clearly that I had plans to do work and haven't done any of it helps motivate me to pay more attention to doing this in the future.
| Quarter | Major one-time tasks (personal projects as well as other buckets) | Major ongoing activities (excluding stuff covered by regular checklists or daily routines) | Other ongoing activities (excluding stuff covered by regular checklists) |
|---|---|---|---|
| Q3 2024 | finish MediaWiki upgrades and server migration, catch up with backlog in DLW data entry; (personal finances, income and taxes) one-time work related to taxes | calculus subwiki expansion, learning subwiki expansion, continued DLW data entry | Updates to my timelines, timelines wiki meta and complementary work, Wikipedia page creation |
| Q4 2024 | wrap up any left-over pieces of server migration, destroy old server, AI safety subwiki setup; (personal chores) rethinking exercise, food portal | calculus subwiki expansion, learning subwiki expansion, continued DLW data entry | Updates to my timelines, timelines wiki meta and complementary work, Wikipedia page creation |
| Q1 2025 | (mix of personal projects and personal chores) digital will | calculus subwiki expansion, learning subwiki expansion, continued DLW data entry | updates to my timelines, timelines wiki meta and complementary work, Wikipedia page creation |
As of the end of August / start of September 2023, my work situation is pretty good; I am now aiming to stick to my half-time goal, with my median week being a little less than half-time, giving me some room to surge during periods when there are high-priority items. Based on feedback received from co-workers about my effective output over the past few months, this seems like a reasonably safe strategy in terms of producing enough output. With that said, I should revisit the strategy every few months and adjust upward or downward as needed.
To be more concrete, I'm planning to target the following with work:
-
I will be work-focused on Monday, Tuesday, and Thursday, though I might still take out time for time-sensitive personal chores. I expect to average at least 3 focused hours per day for these three days.
-
I will try to be responsive on Slack and work email on Wednesday and Friday. I expect that de facto I'll end up spending 30-60 minutes per day on these days, but I will not be trying to spend time at work on these days.
On weeks where I have to take time off on Monday, Tuesday, and Thursday for personal chores or time-sensitive stuff in personal projects, I'll usually compensate by spending a little more time for work on Wednesday and Friday.
-
On the weekends (Saturday and Sunday), my focus will be on things other than the day job.
By default, my day job work during weekends will only be on tasks like these:
-
Attending to incoming stuff such as emails, Slack messages, on-call alerts, and build failures, and making small adjustments and fixes based on these (where it's easier to just make the adjustment rather than add it to a list to do later).
-
Continuing to monitor or push forward any threads started during the week, particularly for cases where small durations of effort can help keep pushing things forward. An example is running a series of jobs where I have to trigger the next job after the previous job finished.
-
Working on a task where there is a lot of time-sensitivity to the completion of the task.
-
Working on a task where there are significant continuity benefits to working on the task now (for instance, there's a lot of context built up from working on it during the week).
Any work I do above and beyond the above should be only for one of these reasons:
-
I need to work on the weekend to meet my absolute minimum weekly target (see below).
-
I need to work on the weekend to meet my conditional minimum weekly target (see below).
-
I have already exceeded my minimum target on personal projects for the week by at least 2 hours, and spending time on day job over the weekend will not put me significantly behind on any non-work tasks (whether personal projects, personal finances, personal chores, or anything else).
-
-
I aim to follow these absolute minimum weekly targets:
-
At least 4 hours 30 minutes on weeks that I do not have any day off of my three working days (either marked as PTO by me or a company holiday)
-
At least 3 hours on weeks that I have one day off of my three working days (either marked as PTO by me or a company holiday)
-
At least 1 hour 30 minutes on weeks that I have marked two days off of my three working days (either marked as PTO by me or a company holiday)
-
-
I aim to follow these conditional minimum weekly targets: at least one of the following must be true:
- At least 7 hours in the week
- At least 7.25 hours average (21.75 hours total) across that week and the two preceding weeks
- At least 7.5 hours average (37.5 hours total) across that week and the four preceding weeks
-
I aim to follow these conditional maximum weekly targets: at least one of the following must be true:
- At most 14 hours in the week
- At most 13.75 hours average (41.25 hours total) across that week and the two preceding weeks
- At most 13.5 hours average (67.5 hours total) across that week and the four preceding weeks
-
Temporary provision until end of September 2024, or when this round of upgrades finishes, whichever is earlier: I can count up to 2 hours per week of time spent on personal devops stuff (not necessarily directly related to ugprades, since general devops stuff is partly a reflection of my not having completed upgrades) toward meeting the hours targets for my day job, including the conditional minimum and absolute minimum, but I am not obliged to count it toward the maximum. This temporary provision is to allow me to make use of lulls in my day job to focus on personal devops work that is time-sensitive and uses similar parts of my brain. Since this work has a definite end, and is similar to my day job, I don't see issues redirecting that time and energy towards day job once it is done.
-
I aim to follow this conditional minimum target each week for personal projects; at least one of the following must be true:
- At least 4 hours on personal projects in that week
- At least 4.25 hours average (12.75 hours total) on personal projects across that week and the two preceding weeks
- At least 4.5 hours average (22.5 hours total) on personal projects across that week and the four preceding weeks
These conditional minimum targets were suspended during my India trip and for the first few weeks after returning from India. They take effect again starting the week of 2024-04-08.
The following are some situations where these targets may be suspended:
-
The week of a work trip (for instance, my trip to Portland 2024-06-03 to 2024-06-07)
-
The week of intercity or international travel, as well as the week prior to it if the travel starts early in the week (Wednesday or sooner)
-
The week after intercity or international travel
-
Up to two weeks prior to travel that was not planned more than three weeks in advance (for instance my Portland trip)
The following is a situation where the target may be relaxed or substituted: if I'm doing significant one-off "building" work that falls under personal finances or personal chores, I can count the one-off "building" work towards the target. Some examples include:
-
Setting up data structures and verification queries related to taxes, that I expect to be useful over more than just one year
-
Refining my travel-related checklists and doing related research
-
Doing setup work related to nutrition recording (this does not include ongoing updates to the data -- just the setup portions of the work)
Another situation where the target may be relaxed or substituted: if I spend over three hours in a week on one-off catchups (excluding weekly catchup with parents), then the excess time I spend over three hours can be substituted for personal projects time as far as meeting the thresholds is concerned (officially, unless directly tied to personal projects, this time would still be counted in the other activities bucket). The idea here is that sometimes these catchups can be lengthy as it makes sense to spend a lot of contiguous time, but they are also relatively rare.
However, I cannot avail of these exceptions too often -- I haven't thought of a limit but, for instance, more than four weeks in a quarter would be excessive, and ideally these exceptions should happen about once a quarter or less. Ideally even in the weeks where I'm doing some work of the above sorts, my personal projects time should still exceed the threshold for that week or using a three-week or five-week average.
-
I aim to follow this conditional maximum target each week for personal chores: at least one of the following must be true:
- At most 13 hours on personal chores in that week
- At most 12.5 hours average (37.5 hours total) on personal chores across that week and the two preceding weeks
- At most 12 hours average (60 hours total) on personal chores across that week and the four preceding weeks
-
I aim to follow this conditional maximum target each week for personal finances, income and taxes: at least one of the following must be true:
- At most 3 hours on personal finances, income and taxes in that week
- At most 2.5 hours average (7.5 hours total) on personal finances, income and taxes across that week and the two preceding weeks
- At most 2 hours average (10 hours total) on personal finances, income and taxes across that week and the four preceding weeks
- Targeting for Q3 2024: Dispose of old cellphone, its swollen original battery, and its probably-not-well-functioning replacement battery
- Targeting for Q3 2024: Get next round of blood tests done
I did a bunch of server migration work in Q2 2023, then took a break to deal with wisdom tooth removal and taxes. I low-key resumed the work in November 2023, but then had to take a break in order to focus on my India trip. I resumed focus on the server migration work in Q2 2024, but some health issues and then the Portland trip distracted me, so I got somewhat less done than I had hoped. Nonetheless, I did make enough progress that it looks like I can do the bulk of the remaining work in Q3 2024, and wrap up the work in Q4 2024 so that I should be able to shut down the old server by the end of 2024.
After the more cleaned-up state I create with the server migration, I'd like to revisit my digital will and restructure it in a way that is more robust.
I plan to do off-and-on editing of the following subwikis: learning, calculus, market, and cellbio. These wikis are all strategic:
-
The learning subwiki has various meta benefits -- better understanding of learning can influence my strategy around subwikis and other things in life.
-
The calculus subwiki has the benefit of being in an area where I am very knowledgeable that also has a sufficiently wide audience that I can get feedback from others (unlike groupprops where it's hard to get feedback). So, I can use it to experiment with various formats for data presentation and organization that can then be propagated to other areas.
-
The cellbio subwiki is a subwiki in an area I am somewhat knowledgeable about but not too knowledgeable about, so it is an example where I can combine learning and knowledge organization, and judge how good the subwiki format would be in a domain such as biology. I'm also excited to learn more about cellbio, so it serves a dual function of addressing my learning need.
-
The market subwiki is exciting because of my historical interest in economics and thinking about the world from a market lens.
I might also pick up more higher-math-wiki editing, covering wikis such as groupprops, topospaces, commalg, and diffgeom, but this is likely to come a little later.
Once I have streamlined by subwiki setup process and moved to a new server, I'll streamline the process for setting up brand new subwikis, and then set up additional subwikis. One such possibility is an AI safety subwiki.
While there may be more subwikis to add, I expect that the process will get streamlined more and more and the incremental effort of adding new subwikis will be low.
I added a bunch of food purchases verification queries in Q1 2024; see here for details. I continued working on these in Q2 and then also started recording food preparations and openings, and started building out associated verification queries, all as part of issue 7.
By the end of Q3, I'll have a full quarter of data for preparations and openings, as well as more experience with the system, so that in Q4 I can work on the food portal in earnest.
I resumed DLW data entry work to some extent in Q4 2023 and then paused it in Q1 2024 to focus on my India trip. I did a little bit of work in Q1 2024, then paused to focus on the server migration. My guess is that DLW work will remain in an almost-fully paused state until the bulk of the server migration is done, after which point I'll be picking it up more.
While I have a lot to catch up with on DLW to begin with, I don't yet have a clear sense of what the steady-state ongoing work level will be once I have caught up.
See here for details.
I did a little bit of this work in Q1 2024, including updating the timeline of BART and starting updates for the timeline of AI safety. However, on the whole, this work took a backseat to other tasks related to my India trip. I continued working on timeline of AI safety in Q2 2024, bringing it a bit closer to current but not fully current. My plan is to be current on my existing list of timelines by the end of Q3 2024 and then stay current, possibly making the list of timelines I monitor regularly longer over time.
I want to work on a bunch of timelines in the timelines wiki that have synergies with the work I'm doing on DLW. While doing this, the synergistic work will also give rise to ideas that I might develop into EA Forum posts.
Another side-effect of the above-mentioned work on timelines wiki will be more updates on general practices with timelines wiki, that I plan to then share with Sebastian and help inform his work as well. I also plan to make active efforts to articulate meta policies in collaboration with Issa and Sebastian.
I want to think systematically about plans for family and career. In Q4 2023 I started writing (private) notes documenting my thoughts on a variety of aspects of this. I'm planning to start picking this up more in Q3 2024, though it is lower in priority than my devops work.
I'm probably not going to say a lot about this publicly except that I'm thinking about the topic.
I will be discussing AI Watch further with Issa once Org Watch is caught up with.
This overlaps to some extent with timelines wiki work.
I'll decide the details regarding this after the outcome of the AI Watch and timelines work.
I have already published https://www.lesswrong.com/posts/Efj8NCCv3TqDL5mbC/aligning-my-web-server-with-devops-practices-part-1-backups and https://www.lesswrong.com/posts/Emmvzz3xXjfH4342f/aligning-my-web-server-with-devops-practices-part-2-security
I have some private notes around spam and performance that I can convert to a public draft at some point (probably after publishing the security post).
Publishing these posts is secondary to actually doing the devops work, but it might still be complementary in a good way, so I might end up doing it even while the devops work is in progres.
I was previously writing stuff on wikiHow, but the level of specialization and recent changes to wikiHow make it not a great forum any more for the posts. So I'm thinking of writing these on LessWrong now.
I have a public draft https://github.com/vipulnaik/working-drafts/blob/master/wikihow/troubleshoot-regressions-in-production-systems.mediawiki that needs to be converted from wikiHow-style to LessWrong-style and published.
I have some private notes related to debugging that identifies three steps:
- Identify characteristics and try to replicate
- Identify proximal cause
- Identify root cause
I need to move this out of its original context (that can't be publicly disclosed) and frame it as general problem-solving tips.
I have one draft post https://github.com/vipulnaik/working-drafts/blob/master/blog/how-my-life-has-improved.md that is worth polishing and publishing. I might get around to this in Q4 2022 (the post is mostly ready so additional work needed to publish it is minimal) but I might also delay it further.
I also wrote a draft post https://github.com/vipulnaik/working-drafts/blob/master/blog/managing-personal-projects-with-a-demanding-day-job.md back in 2020. I might want to revive it and update it, but probably that won't happen till 2023.
I wrote another draft post https://github.com/vipulnaik/working-drafts/blob/master/lesswrong/steady-states-and-crunch-times.md in early March, in the context of the crunch time I was experiencing switching jobs. However, I got some negative feedback on the post so I decided to shelve it and revisit later. I might pick it up later.
I have thoughts on other posts but no ready-to-publish drafts that I can remember offhand.
Idea + mental notes: personal finances, credit scores and credit cards (possibly two different posts)
I have an idea for a post on managing personal finances, that includes a range of stuff including balancing money between accounts, setting up alerts and notifications, etc.
I also have a post about best practices for physical caution, based on my private notes about the subject.
I might write a few posts in 2023 or later about these topics:
- Reasons I switched to Equator/biotech, and reasons I had to leave
- Comparison of my life in academia and industry
I have one draft post at https://github.com/vipulnaik/working-drafts/blob/master/blog/comparative-retrospective-of-my-life-in-academia-and-tech.md with a comparative retrospective of academia and the tech sector.
There's no blocker really to working on this post; however, it's somewhat lower in priority than the other personal projects work, and I also think the post can be even more accurate after a greater passage of time because I'll have a better sense of how things shaped up longer-term.
I would like to write a retrospective of my EA Hotel donation. I might do this any time in the coming months, possibly during or after various DLW updates when I feel in the mood.
I think this retrospective will be useful as an effective review of the EA Hotel, but its bigger utility to me (and others) is in the evaluation of donations to small organizations with potential. It would be cool to do this before my next donation, but it's not a blocker.
I plan to write a retrospective of my COVID-19 forecasts post. The resurgence of the delta strain led me to decide to defer work on this retrospective. I think mid-2022 will be a good time, but I might choose an earlier or later time depending on how COVID-19 evolves.
Idea: Daniel Pink MasterClass review (either before April 2022, or delayed till I renew my MasterClass subscription)
I started writing a draft post https://github.com/vipulnaik/working-drafts/blob/master/lesswrong/daniel-pink-sales-and-persuasion-masterclass-review.md reviewing Daniel Pink's MasterClass. My MasterClass subscription expired in early April 2022, so I probably won't finish it unless and until I decide to get a MasterClass subscription again.
See https://github.com/vipulnaik/working-drafts/blob/master/wikipedia/pages-to-create.md for my current list (some of the articles here might already be present or might no longer make sense to create when I actually get around to them).
I'd like to do some Wikipedia editing occasionally to "stay in touch" with that side of things, and to get some gratification for contributing to something public and widely known (in contrast with timelines wiki where the traffic to pages tends to be low even when their quality is great).