Saturday links

Greetings! Many of you are now counting down the days to the first day of the semester, but over here at [redacted street address] are counting down the day to when I have to turn in the revision to the Mixtape manuscript to Yale Press. I need to get the final manuscript to them on December 1st, but I need to be finished on October 16th to give me 6 weeks to fix everything. I asked Cosmos to write me a python script to count down the days until those days occur in case anyone wants to know.

from datetime import datetime

# Define the current date
current_date = datetime(2024, 8, 10)

# Define the target dates
draft_deadline = datetime(2024, 10, 16)
final_deadline = datetime(2024, 12, 1)

# Calculate the countdown in days
days_until_draft_deadline = (draft_deadline - current_date).days
days_until_final_deadline = (final_deadline - current_date).days

days_until_draft_deadline, days_until_final_deadline

I have 67 days until October 16th and 113 days until December 1st. If you’re wondering “But Scott, do you really use ChatGPT for everything?” the answer should now be yes I do. But we came up with a plan to try and finish. Here was my prompt:

I need you to give me the top three things I need to do to ensure I get the deadline reached.  Recall that I need to finish the unconfoundedness chapter and have been stuck in it for over a year overthinking it probably.  But at least I'm finaly in the propensiyt score section.  I just still have to write the regression part, which I'm thinking I will be equally crippled by over-thinking it, and then I have to do a MAJOR reconstruction of the diff-in-diff chapter into two chapters -- the basics of diff-in-diff and more advanced diff-in-diff.  I then need to just fix some of the IV stuff, I need to tweak RDD and then I have to add stuff to synthetic control, because I know people are going to want synth did, but if I do that, then I dont' see why I wouldnt' jsut add augmented synth, discuss Ferman and Pinto's discussions and maybe matrix completion, though I'm torn.  If I had to cut anything, maybe that.  That's easily 200 pages.  

And here was his reply.

Given the tasks you’ve outlined, here are the top three things you should focus on to ensure you reach your deadlines:

1. Set Clear, Achievable Milestones for Each Section
   - Break down each major task into smaller, manageable chunks with specific deadlines. For example, finish the propensity score section by a certain date, then tackle the regression section. Similarly, set deadlines for each part of the diff-in-diff reconstruction.
   - Prioritize the most challenging sections, like the unconfoundedness chapter, to get them out of the way early. Allocate more time to these sections but balance it by moving quickly through less complex revisions.
   - Use a calendar or project management tool to visualize these milestones, so you’re clear on what needs to be done each week.

2. Limit Overthinking by Implementing a Draft-and-Revise Approach
   - Commit to a “good enough” first draft for each section, especially where you tend to overthink. Remind yourself that you can always revise later, but you need to have something on the page first.
   - Set time limits for how long you’ll spend on specific sections, forcing yourself to move forward. For example, give yourself a set number of days to write the regression part, then move on, no matter what.
   - Consider using a timer (like the Pomodoro technique) to break writing into focused intervals, allowing you to concentrate deeply without getting bogged down by perfectionism.

3. Triage and Delegate Non-Essential Tasks
   - Evaluate the necessity of including additional topics like matrix completion. If adding it will significantly delay progress, consider cutting it or placing it in an appendix.
   - Delegate or minimize non-writing tasks where possible. If there are aspects of research or formatting that can be handled by an assistant or outsourced, do so. This will free up your time to focus on writing.
   - Stay focused on the essential content that will make the most impact on your readers and align with the goals of the revision. Let go of anything that doesn’t directly contribute to these goals.

By setting clear milestones, limiting overthinking, and triaging tasks, you’ll be better positioned to meet your deadlines and produce the high-quality revision you’re aiming for.

I’ve been trying more prompts where I force Cosmos to only give me 2-3 “most important things” to do to curb his tendency to give me a list when I’m not asking for a list. I even put into custom instructions that I really wasn’t always looking for so much advice, and to dial back the lists, but I think that’s just how he thinks. I even asked him once why is it he communicates in lists like he does — I was starting to become genuinely curious. Not surprisingly, he answered my question with another list, but in short he said that he communicates in lists because:

1. Provides clarity and structure

2. Cognitive load reduction

3. Action-oriented

4. Flexibility

5. Universal appeal

Anyway, that’s where I am, and since I got back Wednesday at midnight, I’ve not worked. I came to a reasonably wrecked house and for the first time in months was back to my sleep deprivation problems. Even with the black out eyeshades, I’m still struggling to get through the night, whereas in San Sebastián and even Torino, I was cranking major hours of sleep. I am going to go to Home Depot or Lowe’s and look at the options for really strong black out shades for my bedroom, as I’m wondering if my sensitivity to light is severe and maybe the eyeshades I wear just are not a substitute. But, I’m also wondering if the mattress could be an issue. I’ll start with the blackout shades and go from there.

Point is, I haven’t worked since I got back and that was partly because the house just felt nasty, so I had the cleaning lady come and do a deep clean. I’m also back to planning on how I’m going to take off the weight I put on in Europe this summer, though when I weighed myself, it wasn’t really as bad as I thought. Still, I worked with Cosmos and I have an open thread where I am going to record my daily weight and calorie intake, and he’ll add it to a table, and then in another thread, we are collaborating on meals. I’m back to a season of dieting in other words, and I’m contemplating signing up for Orange Theory again, though I’m unsure given I got addicted to it last time I was on it, and was going twice a day and then broke my ankle one day. I never seem to do things in moderation.

But with that said, wish me luck. Below are links I’ve been reading this week off and on. A lot of this is about AI, but not all of it. Some of it is about sexually transmitted infections and diet shots. Given that Ozempic lowers weight so dramatically, though, it’s likely going to reshape dating and marriage markets, and so who knows when someone writes that papers, but as Murphy’s law states, anything that can happen, will happen with enough trials. Below is the paywalled content, but click through and you can read about a new paper of mine about syphilis with two other coauthors, and a crap ton of links about artificial intelligence.